Showing 1–50 of 96 results for author: Pineda, L

Geometric Retargeting: A Principled, Ultrafast Neural Hand Retargeting Algorithm

Authors: Zhao-Heng Yin, Changhao Wang, Luis Pineda, Krishna Bodduluri, Tingfan Wu, Pieter Abbeel, Mustafa Mukadam

Abstract: We introduce Geometric Retargeting (GeoRT), an ultrafast, and principled neural hand retargeting algorithm for teleoperation, developed as part of our recent Dexterity Gen (DexGen) system. GeoRT converts human finger keypoints to robot hand keypoints at 1KHz, achieving state-of-the-art speed and accuracy with significantly fewer hyperparameters. This high-speed capability enables flexible postproc… ▽ More We introduce Geometric Retargeting (GeoRT), an ultrafast, and principled neural hand retargeting algorithm for teleoperation, developed as part of our recent Dexterity Gen (DexGen) system. GeoRT converts human finger keypoints to robot hand keypoints at 1KHz, achieving state-of-the-art speed and accuracy with significantly fewer hyperparameters. This high-speed capability enables flexible postprocessing, such as leveraging a foundational controller for action correction like DexGen. GeoRT is trained in an unsupervised manner, eliminating the need for manual annotation of hand pairs. The core of GeoRT lies in novel geometric objective functions that capture the essence of retargeting: preserving motion fidelity, ensuring configuration space (C-space) coverage, maintaining uniform response through high flatness, pinch correspondence and preventing self-collisions. This approach is free from intensive test-time optimization, offering a more scalable and practical solution for real-time hand retargeting. △ Less

Submitted 10 March, 2025; originally announced March 2025.

Comments: Project Website: https://zhaohengyin.github.io/geort

An 18-25 GHz spectroscopic survey of dense cores in the Chamaeleon I molecular cloud

Authors: Dariusz C. Lis, William D. Langer, Jorge L. Pineda, Kahaan Gandhi, Karen Willacy, Paul F. Goldsmith, Susanna Widicus Weaver, Liton Majumdar, Youngmin Seo, Shinji Horiuchi, Cheikh Bop, François Lique

Abstract: We extend the survey for organics in the southern hemisphere by observing two cores in the Chamaeleon complex using NASA's Deep Space Network 70-m antenna in Canberra, Australia, over the frequency range of 18 to 25 GHz. We surveyed the class 0 protostar Cha-MMS1 and the prestellar core Cha-C2, which represent two stages in the evolution of dense cores. We detect several molecules including HC… ▽ More We extend the survey for organics in the southern hemisphere by observing two cores in the Chamaeleon complex using NASA's Deep Space Network 70-m antenna in Canberra, Australia, over the frequency range of 18 to 25 GHz. We surveyed the class 0 protostar Cha-MMS1 and the prestellar core Cha-C2, which represent two stages in the evolution of dense cores. We detect several molecules including HC$_3$N, HC$_5$N, C$_4$H, CCS, C$_3$S, NH$_3$, and c-C$_3$H$_2$. A longer cyanopolyyne, HC$_7$N, is detected with high confidence via spectral stacking analysis. While molecular column densities in the two Chamaeleon cores are typically an order of magnitude lower compared to the cynaopolyyne peak in TMC-1, the molecular abundance ratios are in general agreement with the TMC-1 values. The two exceptions are c-C$_3$H$_2$, which is enhanced by a factor of \about 25 with respect to cyanopolyynes in the Chamaeleon cores, and ammonia, which is enhanced by a factor of ~ 125. The deuterated species c-C$_3$HD is detected in both cores, with a high D/H ratio of ~0.23 in c-C$_3$H$_2$. A rare isotopologue of ammonia, $^{15}$NH$_3$, is also detected in Cha-MMS1 suggesting a high $^{14}$N/$^{15}$N ratio of ~ 690 in ammonia. However, this ratio may be artificially enhanced due to the high optical depth of the $^{14}$NH$_3$ (1,1) line, which increases the effective source size. We use the detections of ammonia, cyanopolyynes, and far-infrared dust continuum to characterize the density and temperature in the Chamaeleon cores and calculate the molecular column densities and their relative ratios. The ring molecule benzonitrile is not detected in either Chamaeleon core. The $3 σ$ upper limits for its column density are a factor of 2 higher than the value derived for TMC-1 and the upper limits for its relative abundance with respect to HC$_5$N are a factor of 3 higher than the TMC-1 value. △ Less

Submitted 3 March, 2025; originally announced March 2025.

Comments: Accepted for publication in Astronomy & Astrophysics, 15 pages, 10 figures

Journal ref: A&A 696, A61 (2025)

Representation and Interpretation in Artificial and Natural Computing

Authors: Luis A. Pineda

Abstract: Artificial computing machinery transforms representations through an objective process, to be interpreted subjectively by humans, so the machine and the interpreter are different entities, but in the putative natural computing both processes are performed by the same agent. The method or process that transforms a representation is called here \emph{the mode of computing}. The mode used by digital… ▽ More Artificial computing machinery transforms representations through an objective process, to be interpreted subjectively by humans, so the machine and the interpreter are different entities, but in the putative natural computing both processes are performed by the same agent. The method or process that transforms a representation is called here \emph{the mode of computing}. The mode used by digital computers is the algorithmic one, but there are others, such as quantum computers and diverse forms of non-conventional computing, and there is an open-ended set of representational formats and modes that could be used in artificial and natural computing. A mode based on a notion of computing different from Turing's may perform feats beyond what the Turing Machine does but the modes would not be of the same kind and could not be compared. For a mode of computing to be more powerful than the algorithmic one, it ought to compute functions lacking an effective algorithm, and Church Thesis would not hold. Here, a thought experiment including a computational demon using a hypothetical mode for such an effect is presented. If there is natural computing, there is a mode of natural computing whose properties may be causal to the phenomenological experience. Discovering it would come with solving the hard problem of consciousness; but if it turns out that such a mode does not exist, there is no such thing as natural computing, and the mind is not a computational process. △ Less

Submitted 14 February, 2025; originally announced February 2025.

ACM Class: F.0

Reconstruction of frequency-localized functions from pointwise samples via least squares and deep learning

Authors: A. Martina Neuman, Andres Felipe Lerma Pineda, Jason J. Bramburger, Simone Brugiapaglia

Abstract: Recovering frequency-localized functions from pointwise data is a fundamental task in signal processing. We examine this problem from an approximation-theoretic perspective, focusing on least squares and deep learning-based methods. First, we establish a novel recovery theorem for least squares approximations using the Slepian basis from uniform random samples in low dimensions, explicitly trackin… ▽ More Recovering frequency-localized functions from pointwise data is a fundamental task in signal processing. We examine this problem from an approximation-theoretic perspective, focusing on least squares and deep learning-based methods. First, we establish a novel recovery theorem for least squares approximations using the Slepian basis from uniform random samples in low dimensions, explicitly tracking the dependence of the bandwidth on the sampling complexity. Building on these results, we then present a recovery guarantee for approximating bandlimited functions via deep learning from pointwise data. This result, framed as a practical existence theorem, provides conditions on the network architecture, training procedure, and data acquisition sufficient for accurate approximation. To complement our theoretical findings, we perform numerical comparisons between least squares and deep learning for approximating one- and two-dimensional functions. We conclude with a discussion of the theoretical limitations and the practical gaps between theory and implementation. △ Less

Submitted 13 February, 2025; originally announced February 2025.

DexterityGen: Foundation Controller for Unprecedented Dexterity

Authors: Zhao-Heng Yin, Changhao Wang, Luis Pineda, Francois Hogan, Krishna Bodduluri, Akash Sharma, Patrick Lancaster, Ishita Prasad, Mrinal Kalakrishnan, Jitendra Malik, Mike Lambeta, Tingfan Wu, Pieter Abbeel, Mustafa Mukadam

Abstract: Teaching robots dexterous manipulation skills, such as tool use, presents a significant challenge. Current approaches can be broadly categorized into two strategies: human teleoperation (for imitation learning) and sim-to-real reinforcement learning. The first approach is difficult as it is hard for humans to produce safe and dexterous motions on a different embodiment without touch feedback. The… ▽ More Teaching robots dexterous manipulation skills, such as tool use, presents a significant challenge. Current approaches can be broadly categorized into two strategies: human teleoperation (for imitation learning) and sim-to-real reinforcement learning. The first approach is difficult as it is hard for humans to produce safe and dexterous motions on a different embodiment without touch feedback. The second RL-based approach struggles with the domain gap and involves highly task-specific reward engineering on complex tasks. Our key insight is that RL is effective at learning low-level motion primitives, while humans excel at providing coarse motion commands for complex, long-horizon tasks. Therefore, the optimal solution might be a combination of both approaches. In this paper, we introduce DexterityGen (DexGen), which uses RL to pretrain large-scale dexterous motion primitives, such as in-hand rotation or translation. We then leverage this learned dataset to train a dexterous foundational controller. In the real world, we use human teleoperation as a prompt to the controller to produce highly dexterous behavior. We evaluate the effectiveness of DexGen in both simulation and real world, demonstrating that it is a general-purpose controller that can realize input dexterous manipulation commands and significantly improves stability by 10-100x measured as duration of holding objects across diverse tasks. Notably, with DexGen we demonstrate unprecedented dexterous skills including diverse object reorientation and dexterous tool use such as pen, syringe, and screwdriver for the first time. △ Less

Submitted 6 February, 2025; originally announced February 2025.

Comments: Project: https://zhaohengyin.github.io/dexteritygen

Entropic Hetero-Associative Memory

Authors: Rafael Morales, Luis A. Pineda

Abstract: The Entropic Associative Memory holds objects in a 2D relation or ``memory plane'' using a finite table as the medium. Memory objects are stored by reinforcing simultaneously the cells used by the cue, implementing a form of Hebb's learning rule. Stored objects are ``overlapped'' on the medium, hence the memory is indeterminate and has an entropy value at each state. The retrieval operation constr… ▽ More The Entropic Associative Memory holds objects in a 2D relation or ``memory plane'' using a finite table as the medium. Memory objects are stored by reinforcing simultaneously the cells used by the cue, implementing a form of Hebb's learning rule. Stored objects are ``overlapped'' on the medium, hence the memory is indeterminate and has an entropy value at each state. The retrieval operation constructs an object from the cue and such indeterminate content. In this paper we present the extension to the hetero-associative case in which these properties are preserved. Pairs of hetero-associated objects, possibly of different domain and/or modalities, are held in a 4D relation. The memory retrieval operation selects a largely indeterminate 2D memory plane that is specific to the input cue; however, there is no cue left to retrieve an object from such latter plane. We propose three incremental methods to address such missing cue problem, which we call random, sample and test, and search and test. The model is assessed with composite recollections consisting of manuscripts digits and letters selected from the MNIST and the EMNIST corpora, respectively, such that cue digits retrieve their associated letters and vice versa. We show the memory performance and illustrate the memory retrieval operation using all three methods. The system shows promise for storing, recognizing and retrieving very large sets of object with very limited computing resources. △ Less

Submitted 2 November, 2024; originally announced November 2024.

Comments: 17 pages, 5 figures

Nitrogen Abundance Distribution in the inner Milky Way

Authors: Jorge L. Pineda, Shinji Horiuchi, L. D. Anderson, Matteo Luisi, William D. Langer, Paul F. Goldsmith, Thomas B. H. Kuiper, Christian Fischer, Yan Gong, Andreas Brunthaler, Michael Rugel, Karl M. Menten

Abstract: We combine a new Galactic plane survey of Hydrogen Radio Recombination Lines (RRLs) with far-infrared (FIR) surveys of ionized Nitrogen, N+, to determine Nitrogen abundance across Galactic radius. RRLs were observed with NASA DSS-43 70m antenna and the Green Bank Telescope in 108 lines-of-sight spanning -135 degrees < l < 60 degrees, at b=0 degrees. These positions were also observed in [N II] 122… ▽ More We combine a new Galactic plane survey of Hydrogen Radio Recombination Lines (RRLs) with far-infrared (FIR) surveys of ionized Nitrogen, N+, to determine Nitrogen abundance across Galactic radius. RRLs were observed with NASA DSS-43 70m antenna and the Green Bank Telescope in 108 lines-of-sight spanning -135 degrees < l < 60 degrees, at b=0 degrees. These positions were also observed in [N II] 122 um and 205 um lines with the Herschel Space Observatory. Combining RRL and [N II] 122 um and 205 um observations in 41 of 108 samples with high signal-to-noise ratio, we studied ionized Nitrogen abundance distribution across Galactocentric distances of 0-8 kpc. Combined with existing Solar neighborhood and Outer galaxy N/H abundance determinations, we studied this quantity's distribution within the Milky Way's inner 17 kpc for the first time. We found a Nitrogen abundance gradient extending from Galactocentric radii of 4-17 kpc in the Galactic plane, while within 0-4 kpc, the N/H distribution remained flat. The gradient observed at large Galactocentric distances supports inside-out galaxy growth with the additional steepening resulting from variable star formation efficiency and/or radial flows in the Galactic disk, while the inner 4 kpc flattening, coinciding with the Galactic bar's onset, may be linked to radial flows induced by the bar potential. Using SOFIA/FIFI-LS and Herschel/PACS, we observed the [N III] 57 um line to trace doubly ionized gas contribution in a sub-sample of sightlines. We found negligible N++ contributions along these sightlines, suggesting mostly singly ionized Nitrogen originating from low ionization H II region outskirts. △ Less

Submitted 19 July, 2024; v1 submitted 17 July, 2024; originally announced July 2024.

Comments: Accepted for publication at the Astrophysical Journal. 25 pages, 13 figures

Entropic associative memory for real world images

Authors: Noé Hernández, Rafael Morales, Luis A. Pineda

Abstract: The entropic associative memory (EAM) is a computational model of natural memory incorporating some of its putative properties of being associative, distributed, declarative, abstractive and constructive. Previous experiments satisfactorily tested the model on structured, homogeneous and conventional data: images of manuscripts digits and letters, images of clothing, and phone representations. In… ▽ More The entropic associative memory (EAM) is a computational model of natural memory incorporating some of its putative properties of being associative, distributed, declarative, abstractive and constructive. Previous experiments satisfactorily tested the model on structured, homogeneous and conventional data: images of manuscripts digits and letters, images of clothing, and phone representations. In this work we show that EAM appropriately stores, recognizes and retrieves complex and unconventional images of animals and vehicles. Additionally, the memory system generates meaningful retrieval association chains for such complex images. The retrieved objects can be seen as proper memories, associated recollections or products of imagination. △ Less

Submitted 21 May, 2024; originally announced May 2024.

Neural feels with neural fields: Visuo-tactile perception for in-hand manipulation

Authors: Sudharshan Suresh, Haozhi Qi, Tingfan Wu, Taosha Fan, Luis Pineda, Mike Lambeta, Jitendra Malik, Mrinal Kalakrishnan, Roberto Calandra, Michael Kaess, Joseph Ortiz, Mustafa Mukadam

Abstract: To achieve human-level dexterity, robots must infer spatial awareness from multimodal sensing to reason over contact interactions. During in-hand manipulation of novel objects, such spatial awareness involves estimating the object's pose and shape. The status quo for in-hand perception primarily employs vision, and restricts to tracking a priori known objects. Moreover, visual occlusion of objects… ▽ More To achieve human-level dexterity, robots must infer spatial awareness from multimodal sensing to reason over contact interactions. During in-hand manipulation of novel objects, such spatial awareness involves estimating the object's pose and shape. The status quo for in-hand perception primarily employs vision, and restricts to tracking a priori known objects. Moreover, visual occlusion of objects in-hand is imminent during manipulation, preventing current systems to push beyond tasks without occlusion. We combine vision and touch sensing on a multi-fingered hand to estimate an object's pose and shape during in-hand manipulation. Our method, NeuralFeels, encodes object geometry by learning a neural field online and jointly tracks it by optimizing a pose graph problem. We study multimodal in-hand perception in simulation and the real-world, interacting with different objects via a proprioception-driven policy. Our experiments show final reconstruction F-scores of $81$% and average pose drifts of $4.7\,\text{mm}$, further reduced to $2.3\,\text{mm}$ with known CAD models. Additionally, we observe that under heavy visual occlusion we can achieve up to $94$% improvements in tracking compared to vision-only methods. Our results demonstrate that touch, at the very least, refines and, at the very best, disambiguates visual estimates during in-hand manipulation. We release our evaluation dataset of 70 experiments, FeelSight, as a step towards benchmarking in this domain. Our neural representation driven by multimodal sensing can serve as a perception backbone towards advancing robot dexterity. Videos can be found on our project website https://suddhu.github.io/neural-feels/ △ Less

Submitted 20 December, 2023; originally announced December 2023.

Comments: 43 pages, 20 figures, 1 table; https://suddhu.github.io/neural-feels/

Perceiving Extrinsic Contacts from Touch Improves Learning Insertion Policies

Authors: Carolina Higuera, Joseph Ortiz, Haozhi Qi, Luis Pineda, Byron Boots, Mustafa Mukadam

Abstract: Robotic manipulation tasks such as object insertion typically involve interactions between object and environment, namely extrinsic contacts. Prior work on Neural Contact Fields (NCF) use intrinsic tactile sensing between gripper and object to estimate extrinsic contacts in simulation. However, its effectiveness and utility in real-world tasks remains unknown. In this work, we improve NCF to ena… ▽ More Robotic manipulation tasks such as object insertion typically involve interactions between object and environment, namely extrinsic contacts. Prior work on Neural Contact Fields (NCF) use intrinsic tactile sensing between gripper and object to estimate extrinsic contacts in simulation. However, its effectiveness and utility in real-world tasks remains unknown. In this work, we improve NCF to enable sim-to-real transfer and use it to train policies for mug-in-cupholder and bowl-in-dishrack insertion tasks. We find our model NCF-v2, is capable of estimating extrinsic contacts in the real-world. Furthermore, our insertion policy with NCF-v2 outperforms policies without it, achieving 33% higher success and 1.36x faster execution on mug-in-cupholder, and 13% higher success and 1.27x faster execution on bowl-in-dishrack. △ Less

Submitted 28 September, 2023; originally announced September 2023.

Comments: Under review

Diverse Molecular Structures Across The Whole Star-Forming Disk of M83: High fidelity Imaging at 40pc Resolution

Authors: Jin Koda, Akihiko Hirota, Fumi Egusa, Kazushi Sakamoto, Tsuyoshi Sawada, Mark Heyer, Junichi Baba, Samuel Boissier, Daniela Calzetti, Jennifer Donovan Meyer, Bruce G. Elmegreen, Armando Gil de Paz, Nanase Harada, Luis C. Ho, Masato I. N. Kobayashi, Nario Kuno, Amanda M Lee, Barry F. Madore, Fumiya Maeda, Sergio Martin, Kazuyuki Muraoka, Kouichiro Nakanishi, Sachiko Onodera, Jorge L. Pineda, Nick Scoville , et al. (1 additional authors not shown)

Abstract: We present high-fidelity CO(1-0) imaging of molecular gas across the full star-forming disk of M83, using ALMA's 12m, 7m, and TP arrays and the MIRIAD package. The data have a mass sensitivity and resolution of 10^4Msun and 40 pc. The full disk coverage shows that the characteristics of molecular gas change radially from the center to outer disk. The molecular gas distribution shows coherent large… ▽ More We present high-fidelity CO(1-0) imaging of molecular gas across the full star-forming disk of M83, using ALMA's 12m, 7m, and TP arrays and the MIRIAD package. The data have a mass sensitivity and resolution of 10^4Msun and 40 pc. The full disk coverage shows that the characteristics of molecular gas change radially from the center to outer disk. The molecular gas distribution shows coherent large-scale structures in the inner part, including the central concentration, bar offset ridges, and prominent molecular spiral arms. In the outer disk, the spiral arms appear less spatially coherent, and even flocculent. Massive filamentary gas concentrations are abundant even in the interarm regions. Building up these structures in the interarm regions would require a very long time (~>100Myr). Instead, they must have formed within stellar spiral arms and been released into the interarm regions. For such structures to survive through the dynamical processes, the lifetimes of these structures and their constituent molecules and molecular clouds must be long (~>100Myr). These interarm structures host little or no star formation traced by Halpha. The new map also shows extended CO emission, which likely represents an ensemble of unresolved molecular clouds. △ Less

Submitted 21 March, 2023; originally announced March 2023.

Comments: Accepted for publication in ApJ

Spiral Arms are Metal Freeways: Azimuthal Gas-Phase Metallicity Variations in Simulated Cosmological Zoom-in Flocculent Disks

Authors: Matthew E. Orr, Blakesley Burkhart, Andrew Wetzel, Philip F. Hopkins, Ivanna A. Escala, Allison L. Strom, Paul F. Goldsmith, Jorge L. Pineda, Christopher C. Hayward, Sarah R. Loebman

Abstract: We examine the azimuthal variations in gas-phase metallicity profiles in simulated Milky Way mass disk galaxies from the Feedback in Realistic Environments (FIRE-2) cosmological zoom-in simulation suite, which includes a sub-grid turbulent metal mixing model. We produce spatially resolved maps of the disks at $z \approx 0$ with pixel sizes ranging from 250 to 750~pc, analogous to modern integral f… ▽ More We examine the azimuthal variations in gas-phase metallicity profiles in simulated Milky Way mass disk galaxies from the Feedback in Realistic Environments (FIRE-2) cosmological zoom-in simulation suite, which includes a sub-grid turbulent metal mixing model. We produce spatially resolved maps of the disks at $z \approx 0$ with pixel sizes ranging from 250 to 750~pc, analogous to modern integral field unit (IFU) galaxy surveys, mapping the gas-phase metallicities in both the cold & dense gas and the ionized gas correlated with HII regions. We report that the spiral arms alternate in a pattern of metal rich and metal poor relative to the median metallicity on the order of $\lesssim 0.1$~dex, appearing generally in this sample of flocculent spirals. The pattern persists even in a simulation with different strengths of metal mixing, indicating that the pattern emerges from physics above the sub-grid scale. Local enrichment does not appear to be the dominant source of the azimuthal metallicity variations at $z \approx 0$: there is no correlation with local star formation on these spatial scales. Rather, the arms are moving inwards and outwards relative to each other, carrying their local metallicity gradients with them radially before mixing into the larger-scale interstellar medium. We propose that the arms act as freeways channeling relatively metal poor gas radially inwards, and relatively enriched gas radially outwards. △ Less

Submitted 28 September, 2022; originally announced September 2022.

Comments: 16 Pages, 14 Figures, submitted to MNRAS

Theseus: A Library for Differentiable Nonlinear Optimization

Authors: Luis Pineda, Taosha Fan, Maurizio Monge, Shobha Venkataraman, Paloma Sodhi, Ricky T. Q. Chen, Joseph Ortiz, Daniel DeTone, Austin Wang, Stuart Anderson, Jing Dong, Brandon Amos, Mustafa Mukadam

Abstract: We present Theseus, an efficient application-agnostic open source library for differentiable nonlinear least squares (DNLS) optimization built on PyTorch, providing a common framework for end-to-end structured learning in robotics and vision. Existing DNLS implementations are application specific and do not always incorporate many ingredients important for efficiency. Theseus is application-agnost… ▽ More We present Theseus, an efficient application-agnostic open source library for differentiable nonlinear least squares (DNLS) optimization built on PyTorch, providing a common framework for end-to-end structured learning in robotics and vision. Existing DNLS implementations are application specific and do not always incorporate many ingredients important for efficiency. Theseus is application-agnostic, as we illustrate with several example applications that are built using the same underlying differentiable components, such as second-order optimizers, standard costs functions, and Lie groups. For efficiency, Theseus incorporates support for sparse solvers, automatic vectorization, batching, GPU acceleration, and gradient computation with implicit differentiation and direct loss minimization. We do extensive performance evaluation in a set of applications, demonstrating significant efficiency gains and better scalability when these features are incorporated. Project page: https://sites.google.com/view/theseus-ai △ Less

Submitted 18 January, 2023; v1 submitted 19 July, 2022; originally announced July 2022.

Comments: Advances in Neural Information Processing Systems (NeurIPS), 2022

K-level Reasoning for Zero-Shot Coordination in Hanabi

Authors: Brandon Cui, Hengyuan Hu, Luis Pineda, Jakob N. Foerster

Abstract: The standard problem setting in cooperative multi-agent settings is self-play (SP), where the goal is to train a team of agents that works well together. However, optimal SP policies commonly contain arbitrary conventions ("handshakes") and are not compatible with other, independently trained agents or humans. This latter desiderata was recently formalized by Hu et al. 2020 as the zero-shot coordi… ▽ More The standard problem setting in cooperative multi-agent settings is self-play (SP), where the goal is to train a team of agents that works well together. However, optimal SP policies commonly contain arbitrary conventions ("handshakes") and are not compatible with other, independently trained agents or humans. This latter desiderata was recently formalized by Hu et al. 2020 as the zero-shot coordination (ZSC) setting and partially addressed with their Other-Play (OP) algorithm, which showed improved ZSC and human-AI performance in the card game Hanabi. OP assumes access to the symmetries of the environment and prevents agents from breaking these in a mutually incompatible way during training. However, as the authors point out, discovering symmetries for a given environment is a computationally hard problem. Instead, we show that through a simple adaption of k-level reasoning (KLR) Costa Gomes et al. 2006, synchronously training all levels, we can obtain competitive ZSC and ad-hoc teamplay performance in Hanabi, including when paired with a human-like proxy bot. We also introduce a new method, synchronous-k-level reasoning with a best response (SyKLRBR), which further improves performance on our synchronous KLR by co-training a best response. △ Less

Submitted 14 July, 2022; originally announced July 2022.

Comments: Neurips 2021. 15 pages. 2 figures

Journal ref: Advances in Neural Information Processing Systems 2021. Vol 34. 8215--8228

Deep neural networks can stably solve high-dimensional, noisy, non-linear inverse problems

Authors: Andrés Felipe Lerma Pineda, Philipp Christian Petersen

Abstract: We study the problem of reconstructing solutions of inverse problems when only noisy measurements are available. We assume that the problem can be modeled with an infinite-dimensional forward operator that is not continuously invertible. Then, we restrict this forward operator to finite-dimensional spaces so that the inverse is Lipschitz continuous. For the inverse operator, we demonstrate that th… ▽ More We study the problem of reconstructing solutions of inverse problems when only noisy measurements are available. We assume that the problem can be modeled with an infinite-dimensional forward operator that is not continuously invertible. Then, we restrict this forward operator to finite-dimensional spaces so that the inverse is Lipschitz continuous. For the inverse operator, we demonstrate that there exists a neural network which is a robust-to-noise approximation of the operator. In addition, we show that these neural networks can be learned from appropriately perturbed training data. We demonstrate the admissibility of this approach to a wide range of inverse problems of practical interest. Numerical examples are given that support the theoretical findings. △ Less

Submitted 20 October, 2023; v1 submitted 2 June, 2022; originally announced June 2022.

MSC Class: 35R30; 41A25; 68T05

On learning adaptive acquisition policies for undersampled multi-coil MRI reconstruction

Authors: Tim Bakker, Matthew Muckley, Adriana Romero-Soriano, Michal Drozdzal, Luis Pineda

Abstract: Most current approaches to undersampled multi-coil MRI reconstruction focus on learning the reconstruction model for a fixed, equidistant acquisition trajectory. In this paper, we study the problem of joint learning of the reconstruction model together with acquisition policies. To this end, we extend the End-to-End Variational Network with learnable acquisition policies that can adapt to differen… ▽ More Most current approaches to undersampled multi-coil MRI reconstruction focus on learning the reconstruction model for a fixed, equidistant acquisition trajectory. In this paper, we study the problem of joint learning of the reconstruction model together with acquisition policies. To this end, we extend the End-to-End Variational Network with learnable acquisition policies that can adapt to different data points. We validate our model on a coil-compressed version of the large scale undersampled multi-coil fastMRI dataset using two undersampling factors: $4\times$ and $8\times$. Our experiments show on-par performance with the learnable non-adaptive and handcrafted equidistant strategies at $4\times$, and an observed improvement of more than $2\%$ in SSIM at $8\times$ acceleration, suggesting that potentially-adaptive $k$-space acquisition trajectories can improve reconstructed image quality for larger acceleration factors. However, and perhaps surprisingly, our best performing policies learn to be explicitly non-adaptive. △ Less

Submitted 30 March, 2022; originally announced March 2022.

Comments: Accepted to MIDL 2022 as conference paper

Entropic Associative Memory for Manuscript Symbols

Authors: Rafael Morales, Noé Hernández, Ricardo Cruz, Victor D. Cruz, Luis A. Pineda

Abstract: Manuscript symbols can be stored, recognized and retrieved from an entropic digital memory that is associative and distributed but yet declarative; memory retrieval is a constructive operation, memory cues to objects not contained in the memory are rejected directly without search, and memory operations can be performed through parallel computations. Manuscript symbols, both letters and numerals,… ▽ More Manuscript symbols can be stored, recognized and retrieved from an entropic digital memory that is associative and distributed but yet declarative; memory retrieval is a constructive operation, memory cues to objects not contained in the memory are rejected directly without search, and memory operations can be performed through parallel computations. Manuscript symbols, both letters and numerals, are represented in Associative Memory Registers that have an associated entropy. The memory recognition operation obeys an entropy trade-off between precision and recall, and the entropy level impacts on the quality of the objects recovered through the memory retrieval operation. The present proposal is contrasted in several dimensions with neural networks models of associative memory. We discuss the operational characteristics of the entropic associative memory for retrieving objects with both complete and incomplete information, such as severe occlusions. The experiments reported in this paper add evidence on the potential of this framework for developing practical applications and computational models of natural memory. △ Less

Submitted 16 February, 2022; originally announced February 2022.

Comments: 24 pages, 13 figures

The APEX Large CO Heterodyne Orion Legacy Survey (ALCOHOLS). I. Survey overview

Authors: Thomas Stanke, H. G. Arce, J. Bally, P. Bergman, J. Carpenter, C. J. Davis, W. Dent, J. Di Francesco, J. Eislöffel, D. Froebrich, A. Ginsburg, M. Heyer, D. Johnstone, D. Mardones, M. J. McCaughrean, S. T. Megeath, F. Nakamura, M. D. Smith, A. Stutz, K. Tatematsu, C. Walker, J. P. Williams, H. Zinnecker, B. J. Swift, C. Kulesa , et al. (7 additional authors not shown)

Abstract: The Orion molecular cloud complex harbours the nearest GMCs and site of high-mass star formation. Its YSO populations are thoroughly characterized. The region is therefore a prime target for the study of star formation. Here, we verify the performance of the SuperCAM 64 pixel heterodyne array on APEX. We give a descriptive overview of a set of wide-field CO(3-2) spectral cubes obtained towards t… ▽ More The Orion molecular cloud complex harbours the nearest GMCs and site of high-mass star formation. Its YSO populations are thoroughly characterized. The region is therefore a prime target for the study of star formation. Here, we verify the performance of the SuperCAM 64 pixel heterodyne array on APEX. We give a descriptive overview of a set of wide-field CO(3-2) spectral cubes obtained towards the Orion GMC complex, aimed at characterizing the dynamics and structure of the extended molecular gas in diverse regions of the clouds, ranging from very active sites of clustered star formation in Orion B to comparatively quiet regions in southern Orion A. We present a 2.7 square degree (130pc$^2$) mapping survey in the CO(3-2) transition, obtained using SuperCAM on APEX at an angular resolution of 19'' (7600AU or 0.037pc at a distance of 400pc), covering L1622, NGC2071, NGC2068, OriB9, NGC2024, and NGC2023 in Orion B, and the southern part of the L1641 cloud in Orion A. We describe CO integrated emission and line moment maps and position-velocity diagrams and discuss a few sub-regions in some detail. Evidence for expanding bubbles is seen with lines splitting into double components, most prominently in NGC2024, where we argue that the bulk of the molecular gas is in the foreground of the HII region. High CO(3-2)/CO(1-0) line ratios reveal warm CO along the western edge of Orion B in the NGC2023/NGC2024 region facing the IC434 HII region. Multiple, well separated radial velocity components seen in L1641-S suggest that it consists of a sequence of clouds at increasingly larger distances. We find a small, spherical cloud - the 'Cow Nebula' globule - north of NGC2071. We trace high velocity line wings for the NGC2071-IR outflow and the NGC2024 CO jet. The protostellar dust core FIR4 (rather than FIR5) is the true driving source of the NGC2024 monopolar outflow. △ Less

Submitted 2 January, 2022; originally announced January 2022.

Comments: Accepted for publication in Astronomy and Astrophysics

Journal ref: A&A 658, A178 (2022)

HI-H$_2$ transition: exploring the role of the magnetic field

Authors: R. Skalidis, K. Tassis, G. V. Panopoulou, J. L. Pineda, Y. Gong, N. Mandarakas, D. Blinov, S. Kiehlmann, J. A. Kypriotakis

Abstract: Atomic gas in the diffuse interstellar medium (ISM) is organized in filamentary structures. These structures usually host cold and dense molecular clumps. The Galactic magnetic field is considered to play an important role in the formation of these clumps. Our goal is to explore the role of the magnetic field in the HI - H$_{2}$ transition process. We targeted a filamentary cloud where gas transit… ▽ More Atomic gas in the diffuse interstellar medium (ISM) is organized in filamentary structures. These structures usually host cold and dense molecular clumps. The Galactic magnetic field is considered to play an important role in the formation of these clumps. Our goal is to explore the role of the magnetic field in the HI - H$_{2}$ transition process. We targeted a filamentary cloud where gas transitions from atomic to molecular. This cloud is located at the edges of an expanding structure, known as the North Celestial Pole Loop (NCPL). We probed the magnetic field properties of the cloud with optical polarization observations. We performed multi-wavelength spectroscopic observations of different species in order to probe the gas phase properties of the cloud. We identified two distinct sub-regions within the cloud. One of the regions hosts purely atomic gas, while the other is dominated by molecular gas although most of it is CO-dark. The estimated plane-of-the-sky magnetic field strength between the two regions remains constant within uncertainties and lies in the range 20 ~ 30$~μ$G. The total magnetic field strength does not scale with density which implies that gas is compressed along the field lines. We also found that turbulence is sub-Alfvénic. The HI velocity gradients are in general perpendicular to the mean magnetic field orientation, except for the region close to the CO clump where they tend to become parallel. The latter is likely related to gas undergoing gravitational infall. The magnetic field morphology of the target cloud is parallel to the HI column density structure of the cloud in the atomic region, while it tends to become perpendicular to the HI structure in the molecular region. If this is verified in more cases it has important consequences for the ISM magnetic field modeling with HI data. △ Less

Submitted 27 May, 2022; v1 submitted 22 October, 2021; originally announced October 2021.

Comments: Accepted for publication in A&A

Journal ref: A&A 665, A77 (2022)

DIISC-I: The Discovery of Kinematically Anomalous HI Clouds in M 100

Authors: Hansung B. Gim, Sanchayeeta Borthakur, Emmanuel Momjian, Mansi Padave, Rolf A. Jansen, Dylan Nelson, Timothy M. Heckman, Robert C. Kennicutt Jr., Andrew J. Fox, Jorge L. Pineda, David Thilker, Guinevere Kauffmann, Jason Tumlinson

Abstract: We report the discovery of two kinematically anomalous atomic hydrogen (HI) clouds in M 100 (NGC 4321), which was observed as part of the Deciphering the Interplay between the Interstellar medium, Stars, and the Circumgalactic medium (DIISC) survey in HI 21 cm at 3.3 km s$^{-1}$ spectroscopic and 44 arcsec$\times$30 arcsec spatial resolution using the Karl G. Jansky Very Large Array. These clouds… ▽ More We report the discovery of two kinematically anomalous atomic hydrogen (HI) clouds in M 100 (NGC 4321), which was observed as part of the Deciphering the Interplay between the Interstellar medium, Stars, and the Circumgalactic medium (DIISC) survey in HI 21 cm at 3.3 km s$^{-1}$ spectroscopic and 44 arcsec$\times$30 arcsec spatial resolution using the Karl G. Jansky Very Large Array. These clouds were identified as structures that show significant kinematic offsets from the rotating disk of M100. The velocity offsets of 40 km s$^{-1}$ observed in these clouds are comparable to the offsets seen in intermediate-velocity clouds (IVCs) in the circumgalactic medium (CGM) of the Milky Way and nearby galaxies. We find that one anomalous cloud in M 100 is associated with star-forming regions detected in H$α$ and far-ultraviolet imaging. Our investigation shows that anomalous clouds in M 100 may originate from multiple mechanisms, such as star formation feedback-driven outflows, ram-pressure stripping, and tidal interactions with satellite galaxies. Moreover, we do not detect any cool CGM at 38.8 kpc from the center of M 100, giving an upper limit of N(HI) $\le$ $1.7\times10^{13}$ cm$^{-2}$ (3$σ$). Since M 100 is in the Virgo cluster, the non-existence of neutral/cool CGM is a likely pathway for turning it into a red galaxy. △ Less

Submitted 13 September, 2021; originally announced September 2021.

Comments: 19 pages, 3 tables, 7 figures, Accepted for publication at ApJ

Active 3D Shape Reconstruction from Vision and Touch

Authors: Edward J. Smith, David Meger, Luis Pineda, Roberto Calandra, Jitendra Malik, Adriana Romero, Michal Drozdzal

Abstract: Humans build 3D understandings of the world through active object exploration, using jointly their senses of vision and touch. However, in 3D shape reconstruction, most recent progress has relied on static datasets of limited sensory data such as RGB images, depth maps or haptic readings, leaving the active exploration of the shape largely unexplored. Inactive touch sensing for 3D reconstruction,… ▽ More Humans build 3D understandings of the world through active object exploration, using jointly their senses of vision and touch. However, in 3D shape reconstruction, most recent progress has relied on static datasets of limited sensory data such as RGB images, depth maps or haptic readings, leaving the active exploration of the shape largely unexplored. Inactive touch sensing for 3D reconstruction, the goal is to actively select the tactile readings that maximize the improvement in shape reconstruction accuracy. However, the development of deep learning-based active touch models is largely limited by the lack of frameworks for shape exploration. In this paper, we focus on this problem and introduce a system composed of: 1) a haptic simulator leveraging high spatial resolution vision-based tactile sensors for active touching of 3D objects; 2)a mesh-based 3D shape reconstruction model that relies on tactile or visuotactile signals; and 3) a set of data-driven solutions with either tactile or visuotactile priors to guide the shape exploration. Our framework enables the development of the first fully data-driven solutions to active touch on top of learned models for object understanding. Our experiments show the benefits of such solutions in the task of 3D shape understanding where our models consistently outperform natural baselines. We provide our framework as a tool to foster future research in this direction. △ Less

Submitted 26 October, 2021; v1 submitted 20 July, 2021; originally announced July 2021.

Journal ref: Published at Neurips 2021

Probing Polarization and the Role of Magnetic Fields in Cloud Destruction in the Keyhole Nebula

Authors: Young Min Seo, C. Darren Dowell, Paul F. Goldsmith, Jorge L. Pineda, Liton Majumdar

Abstract: We present polarimetric observations of the Keyhole Nebula in the Carina Nebula Complex carried out using the Stratospheric Observatory for Infrared Astronomy. The Keyhole Nebula located to the west of $η$ Carinae is believed to be disturbed by the stellar winds from the star. We observed the Keyhole Nebula at 89 $μ$m wavelength with the HAWC+ instrument. The observations cover the entire Keyhole… ▽ More We present polarimetric observations of the Keyhole Nebula in the Carina Nebula Complex carried out using the Stratospheric Observatory for Infrared Astronomy. The Keyhole Nebula located to the west of $η$ Carinae is believed to be disturbed by the stellar winds from the star. We observed the Keyhole Nebula at 89 $μ$m wavelength with the HAWC+ instrument. The observations cover the entire Keyhole Nebula spanning 8$'$ by 5$'$ with central position RA = 10:44:43 and Dec = -59:38:04. The typical uncertainty of polarization measurement is less than 0.5\% in the region with intensity above 5,500 MJy sr$^{-1}$. The polarization has a mean of 2.4\% with a standard deviation of 1.6\% in the region above this intensity, similar to values in other high--mass star--forming regions. The magnetic field orientation in the bar--shaped structure is similar to the large--scale magnetic field orientation. On the other hand, the magnetic field direction in the loop is not aligned with the large--scale magnetic fields but has tight alignment with the loop itself. Analysis of the magnetic field angles and the gas turbulence suggests that the field strength is $\sim$70 $μ$G in the loop. A simple comparison of the magnetic field tension to the ram pressure of $η$ Carinae's stellar wind suggests that the magnetic fields in the Keyhole Nebula are not strong enough to maintain the current structure against the impact of the stellar wind, and that the role of the magnetic field in resisting stellar feedback in the Keyhole Nebula is limited. △ Less

Submitted 21 June, 2021; originally announced June 2021.

Comments: ApJ accepted. This is the version of the article before peer review or editing, as submitted by an author to ApJ. IOP Publishing Ltd is not responsible for any errors or omissions in this version of the manuscript or any version derived from it

The Dense Warm Ionized Medium in the Inner Galaxy

Authors: W. D. Langer, J. L. Pineda, P. F. Goldsmith, E. T. Chambers, D. Riquelme, L. D. Anderson, M. Luisi, M. Justen, C. Buchbender

Abstract: Ionized interstellar gas is an important component of the interstellar medium and its lifecycle. The recent evidence for a widely distributed highly ionized warm interstellar gas with a density intermediate between the warm ionized medium (WIM) and compact HII regions suggests that there is a major gap in our understanding of the interstellar gas. Here we investigate the properties of the dense wa… ▽ More Ionized interstellar gas is an important component of the interstellar medium and its lifecycle. The recent evidence for a widely distributed highly ionized warm interstellar gas with a density intermediate between the warm ionized medium (WIM) and compact HII regions suggests that there is a major gap in our understanding of the interstellar gas. Here we investigate the properties of the dense warm ionized medium (D-WIM) in the Milky Way using spectrally resolved SOFIA GREAT [NII] 205 micron line emission and Green Bank Telescope hydrogen radio recombination lines (RRL) data, supplemented by Herschel PACS [NII] 122 micron data, and spectrally resolved 12CO. We observed eight lines of sight in the 20deg <l < 30deg region in the Galactic plane. We derived the kinetic temperature, and the thermal and turbulent velocity dispersions from the [NII] and RRL linewidths. The regions with [NII] 205 micron emission are characterized by electron densities, n(e) ~ 10 to 35 cm(-3), temperatures from 3400 to 8500 K, and column densities N(N+) ~ 7e16 to 3e17 cm(-2). The ionized hydrogen column densities range from 6e20 to 1.7e21 cm(-2) and the fractional nitrogen ion abundance x(N+) ~1 to 3e-4, implying an enhanced nitrogen abundance at ~ 4.3 kpc from the Galactic Center. The [NII] 205 micron emission coincides with CO emission, although often with an offset in velocity, which suggests that the D-WIM gas is located in, or near, star-forming regions, which themselves are associated with molecular gas. These dense ionized regions are found to contribute > 50% of the observed [CII] intensity along these LOS. The kinetic temperatures we derive are too low to explain the presence of N+ resulting from electron collisional ionization and/or proton charge transfer of atomic nitrogen. Rather, these regions most likely are ionized by extreme ultraviolet radiation. △ Less

Submitted 14 May, 2021; originally announced May 2021.

Comments: 16 pages, 10 figures

Journal ref: A&A 651, A59 (2021)

MBRL-Lib: A Modular Library for Model-based Reinforcement Learning

Authors: Luis Pineda, Brandon Amos, Amy Zhang, Nathan O. Lambert, Roberto Calandra

Abstract: Model-based reinforcement learning is a compelling framework for data-efficient learning of agents that interact with the world. This family of algorithms has many subcomponents that need to be carefully selected and tuned. As a result the entry-bar for researchers to approach the field and to deploy it in real-world tasks can be daunting. In this paper, we present MBRL-Lib -- a machine learning l… ▽ More Model-based reinforcement learning is a compelling framework for data-efficient learning of agents that interact with the world. This family of algorithms has many subcomponents that need to be carefully selected and tuned. As a result the entry-bar for researchers to approach the field and to deploy it in real-world tasks can be daunting. In this paper, we present MBRL-Lib -- a machine learning library for model-based reinforcement learning in continuous state-action spaces based on PyTorch. MBRL-Lib is designed as a platform for both researchers, to easily develop, debug and compare new algorithms, and non-expert user, to lower the entry-bar of deploying state-of-the-art algorithms. MBRL-Lib is open-source at https://github.com/facebookresearch/mbrl-lib. △ Less

Submitted 20 April, 2021; originally announced April 2021.

Off-Belief Learning

Authors: Hengyuan Hu, Adam Lerer, Brandon Cui, David Wu, Luis Pineda, Noam Brown, Jakob Foerster

Abstract: The standard problem setting in Dec-POMDPs is self-play, where the goal is to find a set of policies that play optimally together. Policies learned through self-play may adopt arbitrary conventions and implicitly rely on multi-step reasoning based on fragile assumptions about other agents' actions and thus fail when paired with humans or independently trained agents at test time. To address this,… ▽ More The standard problem setting in Dec-POMDPs is self-play, where the goal is to find a set of policies that play optimally together. Policies learned through self-play may adopt arbitrary conventions and implicitly rely on multi-step reasoning based on fragile assumptions about other agents' actions and thus fail when paired with humans or independently trained agents at test time. To address this, we present off-belief learning (OBL). At each timestep OBL agents follow a policy $π_1$ that is optimized assuming past actions were taken by a given, fixed policy ($π_0$), but assuming that future actions will be taken by $π_1$. When $π_0$ is uniform random, OBL converges to an optimal policy that does not rely on inferences based on other agents' behavior (an optimal grounded policy). OBL can be iterated in a hierarchy, where the optimal policy from one level becomes the input to the next, thereby introducing multi-level cognitive reasoning in a controlled manner. Unlike existing approaches, which may converge to any equilibrium policy, OBL converges to a unique policy, making it suitable for zero-shot coordination (ZSC). OBL can be scaled to high-dimensional settings with a fictitious transition mechanism and shows strong performance in both a toy-setting and the benchmark human-AI & ZSC problem Hanabi. △ Less

Submitted 17 August, 2021; v1 submitted 5 March, 2021; originally announced March 2021.

On the Importance of Hyperparameter Optimization for Model-based Reinforcement Learning

Authors: Baohe Zhang, Raghu Rajan, Luis Pineda, Nathan Lambert, André Biedenkapp, Kurtland Chua, Frank Hutter, Roberto Calandra

Abstract: Model-based Reinforcement Learning (MBRL) is a promising framework for learning control in a data-efficient manner. MBRL algorithms can be fairly complex due to the separate dynamics modeling and the subsequent planning algorithm, and as a result, they often possess tens of hyperparameters and architectural choices. For this reason, MBRL typically requires significant human expertise before it can… ▽ More Model-based Reinforcement Learning (MBRL) is a promising framework for learning control in a data-efficient manner. MBRL algorithms can be fairly complex due to the separate dynamics modeling and the subsequent planning algorithm, and as a result, they often possess tens of hyperparameters and architectural choices. For this reason, MBRL typically requires significant human expertise before it can be applied to new problems and domains. To alleviate this problem, we propose to use automatic hyperparameter optimization (HPO). We demonstrate that this problem can be tackled effectively with automated HPO, which we demonstrate to yield significantly improved performance compared to human experts. In addition, we show that tuning of several MBRL hyperparameters dynamically, i.e. during the training itself, further improves the performance compared to using static hyperparameters which are kept fixed for the whole training. Finally, our experiments provide valuable insights into the effects of several hyperparameters, such as plan horizon or learning rate and their influence on the stability of training and resulting rewards. △ Less

Submitted 26 February, 2021; originally announced February 2021.

Comments: 19 pages, accepted by AISTATS 2021

Deliberative and Conceptual Inference in Service Robots

Authors: Luis A. Pineda, Noé Hernández, Arturo Rodríguez, Ricardo Cruz, Gibrán Fuentes

Abstract: Service robots need to reason to support people in daily life situations. Reasoning is an expensive resource that should be used on demand whenever the expectations of the robot do not match the situation of the world and the execution of the task is broken down; in such scenarios the robot must perform the common sense daily life inference cycle consisting on diagnosing what happened, deciding wh… ▽ More Service robots need to reason to support people in daily life situations. Reasoning is an expensive resource that should be used on demand whenever the expectations of the robot do not match the situation of the world and the execution of the task is broken down; in such scenarios the robot must perform the common sense daily life inference cycle consisting on diagnosing what happened, deciding what to do about it, and inducing and executing a plan, recurring in such behavior until the service task can be resumed. Here we examine two strategies to implement this cycle: (1) a pipe-line strategy involving abduction, decision-making and planning, which we call deliberative inference and (2) the use of the knowledge and preferences stored in the robot's knowledge-base, which we call conceptual inference. The former involves an explicit definition of a problem space that is explored through heuristic search, and the latter is based on conceptual knowledge including the human user preferences, and its representation requires a non-monotonic knowledge-based system. We compare the strengths and limitations of both approaches. We also describe a service robot conceptual model and architecture capable of supporting the daily life inference cycle during the execution of a robotics service task. The model is centered in the declarative specification and interpretation of robot's communication and task structure. We also show the implementation of this framework in the fully autonomous robot Golem-III. The framework is illustrated with two demonstration scenarios. △ Less

Submitted 13 December, 2020; originally announced December 2020.

Comments: 31 pages, 7 figures and 2 appendices

A Broadband Digital Spectrometer for the Deep Space Network

Authors: Kristen Virkler, Jonathon Kocz, Melissa Soriano, Shinji Horiuchi, Jorge L. Pineda, Tyrone McNichols

Abstract: The Deep Space Network (DSN) enables NASA to communicate with its spacecraft in deep space. By virtue of its large antennas, the DSN can also be used as a powerful instrument for radio astronomy. Specifically, Deep Space Station (DSS) 43, the 70 m antenna at the Canberra Deep Space Communications Complex (CDSCC) has a K-band radio astronomy system covering a 10 GHz bandwidth at 17 GHz to 27 GHz. T… ▽ More The Deep Space Network (DSN) enables NASA to communicate with its spacecraft in deep space. By virtue of its large antennas, the DSN can also be used as a powerful instrument for radio astronomy. Specifically, Deep Space Station (DSS) 43, the 70 m antenna at the Canberra Deep Space Communications Complex (CDSCC) has a K-band radio astronomy system covering a 10 GHz bandwidth at 17 GHz to 27 GHz. This spectral range covers a number of atomic and molecular lines, produced in a rich variety of interstellar gas conditions. Lines include hydrogen radio recombination lines (RRLs), cyclopropenylidene, water masers, and ammonia. A new high-resolution spectrometer was deployed at CDSCC in November 2019 and connected to the K-band downconverter. The spectrometer has a total bandwidth of 16 GHz. Such a large total bandwidth enables, for example, the simultaneous observations of a large number of RRLs, which can be combined together to significantly improve the sensitivity of these observations. The system has two firmware modes: 1) A 65k-pt FFT to provide 32768 spectral channels at 30.5 kHz and 2) A 16k-pt polyphase filterbank (PFB) to provide 8192 spectral channels with 122 kHz resolution. The observation process is designed to maximize autonomy, from the Principle Investigator's inputs to the output data in FITS file format. We present preliminary mapping observations of hydrogen RRLs in Orion KL mapping taken using the new spectrometer. △ Less

Submitted 5 October, 2020; originally announced October 2020.

Comments: 10 pages, 5 figures, Accepted for publication in ApJS

An Entropic Associative Memory

Authors: Luis A. Pineda, Gibrán Fuentes, Rafael Morales

Abstract: Natural memories are associative, declarative and distributed. Symbolic computing memories resemble natural memories in their declarative character, and information can be stored and recovered explicitly; however, they lack the associative and distributed properties of natural memories. Sub-symbolic memories developed within the connectionist or artificial neural networks paradigm are associative… ▽ More Natural memories are associative, declarative and distributed. Symbolic computing memories resemble natural memories in their declarative character, and information can be stored and recovered explicitly; however, they lack the associative and distributed properties of natural memories. Sub-symbolic memories developed within the connectionist or artificial neural networks paradigm are associative and distributed, but are unable to express symbolic structure and information cannot be stored and retrieved explicitly; hence, they lack the declarative property. To address this dilemma, we use Relational-Indeterminate Computing to model associative memory registers that hold distributed representations of individual objects. This mode of computing has an intrinsic computing entropy which measures the indeterminacy of representations. This parameter determines the operational characteristics of the memory. Associative registers are embedded in an architecture that maps concrete images expressed in modality-specific buffers into abstract representations, and vice versa, and the memory system as a whole fulfills the three properties of natural memories. The system has been used to model a visual memory holding the representations of hand-written digits, and recognition and recall experiments show that there is a range of entropy values, not too low and not too high, in which associative memory registers have a satisfactory performance. The similarity between the cue and the object recovered in memory retrieve operations depends on the entropy of the memory register holding the representation of the corresponding object. The experiments were implemented in a simulation using a standard computer, but a parallel architecture may be built where the memory operations would take a very reduced number of computing steps. △ Less

Submitted 28 September, 2020; originally announced September 2020.

Comments: 25 pages, 6 figures, 17 references

Analysis of the main factors for the configuration of green ports in Colombia

Authors: Abraham Londono Pineda, Tatiana Arias Naranjo, Jose Alejandro Cano Arenas

Abstract: This study analyzes the factors affecting the configuration and consolidation of green ports in Colombia. For this purpose a case stady of maritime cargo ports of Cartagena, Barranquilla and Santa Marta is performed addressing semiestructured interviews to identify the factors contributing to the consolidation of green ports and the factors guiding the sustainability management in the ports that h… ▽ More This study analyzes the factors affecting the configuration and consolidation of green ports in Colombia. For this purpose a case stady of maritime cargo ports of Cartagena, Barranquilla and Santa Marta is performed addressing semiestructured interviews to identify the factors contributing to the consolidation of green ports and the factors guiding the sustainability management in the ports that have not yet been certified as green ports. The results show that environmental regulations are atarting point not the key factor to consolidate asgreen ports. As a conclusions, the conversion of Colombian to green ports should not be limited to the attaiment of certifications, such as Ecoport certification, but should ensure the contribution to sustainable development through economic, social and environmental dimensions and the achievement of the SDGs △ Less

Submitted 22 September, 2020; originally announced September 2020.

Entropy, Computing and Rationality

Authors: Luis A. Pineda

Abstract: Making decisions freely presupposes that there is some indeterminacy in the environment and in the decision making engine. The former is reflected on the behavioral changes due to communicating: few changes indicate rigid environments; productive changes manifest a moderate indeterminacy, but a large communicating effort with few productive changes characterize a chaotic environment. Hence, commun… ▽ More Making decisions freely presupposes that there is some indeterminacy in the environment and in the decision making engine. The former is reflected on the behavioral changes due to communicating: few changes indicate rigid environments; productive changes manifest a moderate indeterminacy, but a large communicating effort with few productive changes characterize a chaotic environment. Hence, communicating, effective decision making and productive behavioral changes are related. The entropy measures the indeterminacy of the environment, and there is an entropy range in which communicating supports effective decision making. This conjecture is referred to here as the The Potential Productivity of Decisions. The computing engine that is causal to decision making should also have some indeterminacy. However, computations performed by standard Turing Machines are predetermined. To overcome this limitation an entropic mode of computing that is called here Relational-Indeterminate is presented. Its implementation in a table format has been used to model an associative memory. The present theory and experiment suggest the Entropy Trade-off: There is an entropy range in which computing is effective but if the entropy is too low computations are too rigid and if it is too high computations are unfeasible. The entropy trade-off of computing engines corresponds to the potential productivity of decisions of the environment. The theory is referred to an Interaction-Oriented Cognitive Architecture. Memory, perception, action and thought involve a level of indeterminacy and decision making may be free in such degree. The overall theory supports an ecological view of rationality. The entropy of the brain has been measured in neuroscience studies and the present theory supports that the brain is an entropic machine. The paper is concluded with a number of predictions that may be tested empirically. △ Less

Submitted 21 September, 2020; originally announced September 2020.

Comments: 43 pages, 4 figures, 44 references

Sulfur emission reduction in cargo ship manufacturers and shipping companies based on MARPOL Annex VI

Authors: Abraham Londono Pineda, Jose Alejandro Cano, Lissett Pulgarin

Abstract: This article explores the challenges for the adoption of scrubbers and low sulfur fuels on ship manufacturers and shipping companies. Results show that ship manufacturers, must finance their working capital and operating costs, which implies an increase in the prices of the ships employing these new technologies. On the other hand, shipping companies must adopt the most appropriate technology acco… ▽ More This article explores the challenges for the adoption of scrubbers and low sulfur fuels on ship manufacturers and shipping companies. Results show that ship manufacturers, must finance their working capital and operating costs, which implies an increase in the prices of the ships employing these new technologies. On the other hand, shipping companies must adopt the most appropriate technology according to the areas where ships navigate, the scale economies of trade routes, and the cost-benefit analysis of ship modernization. △ Less

Submitted 20 September, 2020; originally announced September 2020.

HAWC+ Far-Infrared Observations of the Magnetic Field Geometry in M51 and NGC 891

Authors: Terry Jay Jones, Jin-Ah Kim, C. Darren Dowell, Mark R. Morris, Jorge L. Pineda, Dominic J. Benford, Marc Berthoud, David T. Chuss, Daniel A. Dale, L. M. Fissel, Paul F. Goldsmith, Ryan T. Hamilton, Shaul Hanany, Doyal A. Harper, Thomas K. Henning, Alex Lazarian, Leslie W. Looney, Joseph M. Michail, Giles Novak, Fabio P. Santos, Kartik Sheth, Javad Siah, Gordon J. Stacey, Johannes Staguhn, Ian W. Stephens , et al. (7 additional authors not shown)

Abstract: SOFIA HAWC+ polarimetry at $154~\micron$ is reported for the face-on galaxy M51 and the edge-on galaxy NGC 891. For M51, the polarization vectors generally follow the spiral pattern defined by the molecular gas distribution, the far-infrared (FIR) intensity contours, and other tracers of star formation. The fractional polarization is much lower in the FIR-bright central regions than in the outer r… ▽ More SOFIA HAWC+ polarimetry at $154~\micron$ is reported for the face-on galaxy M51 and the edge-on galaxy NGC 891. For M51, the polarization vectors generally follow the spiral pattern defined by the molecular gas distribution, the far-infrared (FIR) intensity contours, and other tracers of star formation. The fractional polarization is much lower in the FIR-bright central regions than in the outer regions, and we rule out loss of grain alignment and variations in magnetic field strength as causes. When compared with existing synchrotron observations, which sample different regions with different weighting, we find the net position angles are strongly correlated, the fractional polarizations are moderately correlated, but the polarized intensities are uncorrelated. We argue that the low fractional polarization in the central regions must be due to significant numbers of highly turbulent segments across the beam and along lines of sight in the beam in the central 3 kpc of M51. For NGC 891, the FIR polarization vectors within an intensity contour of 1500 $\rm{MJy~sr^{-1}}$ are oriented very close to the plane of the galaxy. The FIR polarimetry is probably sampling the magnetic field geometry in NGC 891 much deeper into the disk than is possible with NIR polarimetry and radio synchrotron measurements. In some locations in NGC 891 the FIR polarization is very low, suggesting we are preferentially viewing the magnetic field mostly along the line of sight, down the length of embedded spiral arms. There is tentative evidence for a vertical field in the polarized emission off the plane of the disk. △ Less

Submitted 18 August, 2020; originally announced August 2020.

Comments: AJ, Accepted, In Press

A SOFIA Survey of [CII] in the galaxy M51 II. [CII] and CO kinematics across spiral arms

Authors: Jorge L. Pineda, Juergen Stutzki, Christof Buchbender, Jin Koda, Christian Fischer, Paul F. Goldsmith, Simon C. O. Glover, Ralf S. Klessen, Carsten Kramer, Bhaswati Mookerjea, Rowan Smith, Robin Tress, Monika Ziebart

Abstract: We present the first complete, velocity-resolved [CII] 158um image of the M51 grand-design spiral galaxy, observed with the upGREAT instrument on SOFIA. [CII] is an important tracer of various phases of the interstellar medium (ISM), including ionized gas, neutral atomic, and diffuse molecular regions. We combine the [CII] data with HI, CO, 24um dust continuum, FUV, and near-infrared K-band observ… ▽ More We present the first complete, velocity-resolved [CII] 158um image of the M51 grand-design spiral galaxy, observed with the upGREAT instrument on SOFIA. [CII] is an important tracer of various phases of the interstellar medium (ISM), including ionized gas, neutral atomic, and diffuse molecular regions. We combine the [CII] data with HI, CO, 24um dust continuum, FUV, and near-infrared K-band observations to study the evolution of the ISM across M51's spiral arms in both position-position, and position-velocity space. Our data show strong velocity gradients in HI, 12CO, and [CII] at the locations of stellar arms (traced by K--band data) with a clear offset in position-velocity space between upstream molecular gas (traced by 12CO) and downstream star formation (traced by [CII]). We compare the observed position--velocity maps across spiral arms with synthetic observations from numerical simulations of galaxies with both dynamical and quasi-stationary steady spiral arms that predict both tangential and radial velocities at the location of spiral arms. We find that our observations, based on the observed velocity gradients and associated offset between CO and [CII], are consistent with the presence of shocks in spiral arms in the inner parts of M51 and in the arm connecting the companion galaxy, M51b, in the outer parts of M51. △ Less

Submitted 21 August, 2020; v1 submitted 4 August, 2020; originally announced August 2020.

Comments: 20 pages, 14 figures, Accepted for publication in ApJ

Active MR k-space Sampling with Reinforcement Learning

Authors: Luis Pineda, Sumana Basu, Adriana Romero, Roberto Calandra, Michal Drozdzal

Abstract: Deep learning approaches have recently shown great promise in accelerating magnetic resonance image (MRI) acquisition. The majority of existing work have focused on designing better reconstruction models given a pre-determined acquisition trajectory, ignoring the question of trajectory optimization. In this paper, we focus on learning acquisition trajectories given a fixed image reconstruction mod… ▽ More Deep learning approaches have recently shown great promise in accelerating magnetic resonance image (MRI) acquisition. The majority of existing work have focused on designing better reconstruction models given a pre-determined acquisition trajectory, ignoring the question of trajectory optimization. In this paper, we focus on learning acquisition trajectories given a fixed image reconstruction model. We formulate the problem as a sequential decision process and propose the use of reinforcement learning to solve it. Experiments on a large scale public MRI dataset of knees show that our proposed models significantly outperform the state-of-the-art in active MRI acquisition, over a large range of acceleration factors. △ Less

Submitted 7 October, 2020; v1 submitted 20 July, 2020; originally announced July 2020.

Comments: Presented at the 23rd International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2020

Journal ref: LNCS vol. 12262 (2020) 23-33

Swirls of FIRE: Spatially Resolved Gas Velocity Dispersions and Star Formation Rates in FIRE-2 Disk Environments

Authors: Matthew E. Orr, Christopher C. Hayward, Anne M. Medling, Philip F. Hopkins, Norman Murray, Jorge L. Pineda, Claude-André Faucher-Giguère, Dušan Kereš, Kung-Yi Su

Abstract: We study the spatially resolved (sub-kpc) gas velocity dispersion ($σ$)--star formation rate (SFR) relation in the FIRE-2 (Feedback in Realistic Environments) cosmological simulations. We specifically focus on Milky Way mass disk galaxies at late times. In agreement with observations, we find a relatively flat relationship, with $σ\approx 15-30$ km/s in neutral gas across 3 dex in SFRs. We show th… ▽ More We study the spatially resolved (sub-kpc) gas velocity dispersion ($σ$)--star formation rate (SFR) relation in the FIRE-2 (Feedback in Realistic Environments) cosmological simulations. We specifically focus on Milky Way mass disk galaxies at late times. In agreement with observations, we find a relatively flat relationship, with $σ\approx 15-30$ km/s in neutral gas across 3 dex in SFRs. We show that higher dense gas fractions (ratios of dense gas to neutral gas) and SFRs are correlated at constant $σ$. Similarly, lower gas fractions (ratios of gas to stellar mass) are correlated with higher $σ$ at constant SFR. The limits of the $σ$-$Σ_{\rm SFR}$ relation correspond to the onset of strong outflows. We see evidence of "on-off" cycles of star formation in the simulations, corresponding to feedback injection timescales of 10-100 Myr, where SFRs oscillate about equilibrium SFR predictions. Finally, SFRs and velocity dispersions in the simulations agree well with feedback-regulated and marginally stable gas disk (Toomre's $Q =1$) model predictions, and the data effectively rule out models assuming that gas turns into stars at (low) constant efficiency (i.e., ${\rm 1\%}$ per free-fall time). And although the simulation data do not entirely exclude gas accretion/gravitationally powered turbulence as a driver of $σ$, it appears to be strongly subdominant to stellar feedback in the simulated galaxy disks. △ Less

Submitted 31 October, 2019; originally announced November 2019.

Comments: 18 pages, 12 figures, submitted to MNRAS

LitGen: Genetic Literature Recommendation Guided by Human Explanations

Authors: Allen Nie, Arturo L. Pineda, Matt W. Wright Hannah Wand, Bryan Wulf, Helio A. Costa, Ronak Y. Patel, Carlos D. Bustamante, James Zou

Abstract: As genetic sequencing costs decrease, the lack of clinical interpretation of variants has become the bottleneck in using genetics data. A major rate limiting step in clinical interpretation is the manual curation of evidence in the genetic literature by highly trained biocurators. What makes curation particularly time-consuming is that the curator needs to identify papers that study variant pathog… ▽ More As genetic sequencing costs decrease, the lack of clinical interpretation of variants has become the bottleneck in using genetics data. A major rate limiting step in clinical interpretation is the manual curation of evidence in the genetic literature by highly trained biocurators. What makes curation particularly time-consuming is that the curator needs to identify papers that study variant pathogenicity using different types of approaches and evidences---e.g. biochemical assays or case control analysis. In collaboration with the Clinical Genomic Resource (ClinGen)---the flagship NIH program for clinical curation---we propose the first machine learning system, LitGen, that can retrieve papers for a particular variant and filter them by specific evidence types used by curators to assess for pathogenicity. LitGen uses semi-supervised deep learning to predict the type of evidence provided by each paper. It is trained on papers annotated by ClinGen curators and systematically evaluated on new test data collected by ClinGen. LitGen further leverages rich human explanations and unlabeled data to gain 7.9%-12.6% relative performance improvement over models learned only on the annotated papers. It is a useful framework to improve clinical variant curation. △ Less

Submitted 23 September, 2019; originally announced September 2019.

Comments: 12 pages; 5 figures. Accepted by PSB 2020 (Pacific Symposium on Biocomputing) track: Artificial Intelligence for Enhancing Clinical Medicine

On the Evaluation of Conditional GANs

Authors: Terrance DeVries, Adriana Romero, Luis Pineda, Graham W. Taylor, Michal Drozdzal

Abstract: Conditional Generative Adversarial Networks (cGANs) are finding increasingly widespread use in many application domains. Despite outstanding progress, quantitative evaluation of such models often involves multiple distinct metrics to assess different desirable properties, such as image quality, conditional consistency, and intra-conditioning diversity. In this setting, model benchmarking becomes a… ▽ More Conditional Generative Adversarial Networks (cGANs) are finding increasingly widespread use in many application domains. Despite outstanding progress, quantitative evaluation of such models often involves multiple distinct metrics to assess different desirable properties, such as image quality, conditional consistency, and intra-conditioning diversity. In this setting, model benchmarking becomes a challenge, as each metric may indicate a different "best" model. In this paper, we propose the Frechet Joint Distance (FJD), which is defined as the Frechet distance between joint distributions of images and conditioning, allowing it to implicitly capture the aforementioned properties in a single metric. We conduct proof-of-concept experiments on a controllable synthetic dataset, which consistently highlight the benefits of FJD when compared to currently established metrics. Moreover, we use the newly introduced metric to compare existing cGAN-based models for a variety of conditioning modalities (e.g. class labels, object masks, bounding boxes, images, and text captions). We show that FJD can be used as a promising single metric for cGAN benchmarking and model selection. Code can be found at https://github.com/facebookresearch/fjd. △ Less

Submitted 23 December, 2019; v1 submitted 11 July, 2019; originally announced July 2019.

Learning Causal State Representations of Partially Observable Environments

Authors: Amy Zhang, Zachary C. Lipton, Luis Pineda, Kamyar Azizzadenesheli, Anima Anandkumar, Laurent Itti, Joelle Pineau, Tommaso Furlanello

Abstract: Intelligent agents can cope with sensory-rich environments by learning task-agnostic state abstractions. In this paper, we propose an algorithm to approximate causal states, which are the coarsest partition of the joint history of actions and observations in partially-observable Markov decision processes (POMDP). Our method learns approximate causal state representations from RNNs trained to predi… ▽ More Intelligent agents can cope with sensory-rich environments by learning task-agnostic state abstractions. In this paper, we propose an algorithm to approximate causal states, which are the coarsest partition of the joint history of actions and observations in partially-observable Markov decision processes (POMDP). Our method learns approximate causal state representations from RNNs trained to predict subsequent observations given the history. We demonstrate that these learned state representations are useful for learning policies efficiently in reinforcement learning problems with rich observation spaces. We connect causal states with causal feature sets from the causal inference literature, and also provide theoretical guarantees on the optimality of the continuous version of this causal state representation under Lipschitz assumptions by proving equivalence to bisimulation, a relation between behaviorally equivalent systems. This allows for lower bounds on the optimal value function of the learned representation, which is tight given certain assumptions. Finally, we empirically evaluate causal state representations using multiple partially observable tasks and compare with prior methods. △ Less

Submitted 8 February, 2021; v1 submitted 25 June, 2019; originally announced June 2019.

Comments: 35 pages, 8 figures

Relations Between Molecular Cloud Structure Sizes and Line Widths in the Large Magellanic Cloud

Authors: Tony Wong, Annie Hughes, Kazuki Tokuda, Rémy Indebetouw, Toshikazu Onishi, Jeffrey B. Bandurski, C. -H. Rosie Chen, Yasuo Fukui, Simon C. O. Glover, Ralf S. Klessen, Jorge L. Pineda, Julia Roman-Duval, Marta Sewiło, Evan Wojciechowski, Sarolta Zahorecz

Abstract: We present a comparative study of the size-line width relation for substructures within six molecular clouds in the Large Magellanic Cloud (LMC) mapped with the Atacama Large Millimeter/submillimeter Array (ALMA). Our sample extends our previous study, which compared a Planck detected cold cloud in the outskirts of the LMC with the 30 Doradus molecular cloud and found the typical line width for 1… ▽ More We present a comparative study of the size-line width relation for substructures within six molecular clouds in the Large Magellanic Cloud (LMC) mapped with the Atacama Large Millimeter/submillimeter Array (ALMA). Our sample extends our previous study, which compared a Planck detected cold cloud in the outskirts of the LMC with the 30 Doradus molecular cloud and found the typical line width for 1 pc radius structures to be 5 times larger in 30 Doradus. By observing clouds with intermediate levels of star formation activity, we find evidence that line width at a given size increases with increasing local and cloud-scale 8$μ$m intensity. At the same time, line width at a given size appears to independently correlate with measures of mass surface density. Our results suggest that both virial-like motions due to gravity and local energy injection by star formation feedback play important roles in determining intracloud dynamics. △ Less

Submitted 15 September, 2019; v1 submitted 28 May, 2019; originally announced May 2019.

Comments: 20 pages, to appear in ApJ. Data presented in this paper can be found at https://mmwave.astro.illinois.edu/almalmc/

Electron Densities and Nitrogen Abundances in Ionized Gas Derived Using [NII] Fine-structure and Hydrogen Recombination lines

Authors: Jorge L. Pineda, Shinji Horiuchi, L. D. Anderson, Matteo Luisi, Paul F. Goldsmith, William D. Langer, Thomas B. H. Kuiper, Geoff Bryden, Melissa Soriano, Joseph W. Lazio

Abstract: We present a method for deriving the electron density of ionized gas using the ratio of the intensity of the [NII] 205um line to that of Hydrogen radio recombination lines (RRL). We use this method to derive electron densities of 21 velocity components in 11 lines of sight through the Galaxy, including the Galactic center. We observed, at high--spectral resolution, the [NII] 205um line with the He… ▽ More We present a method for deriving the electron density of ionized gas using the ratio of the intensity of the [NII] 205um line to that of Hydrogen radio recombination lines (RRL). We use this method to derive electron densities of 21 velocity components in 11 lines of sight through the Galaxy, including the Galactic center. We observed, at high--spectral resolution, the [NII] 205um line with the Herschel/HIFI and SOFIA/GREAT instruments and the radio recombination lines with the Green Bank Telescope and the NASA Deep Space Network Deep Space Station 43 (DSS--43) telescope. We find typical electron densities between 6 to 170 cm^-3, which are consistent with those derived at low spectral resolution using the [NII] 205um/122um line ratio with Herschel/PACS on a larger sample of sight lines in the Galactic plane. By matching the electron densities derived from the [NII] 205um/RRL intensity ratio and the [NII] 122um/205um intensity ratio, we derive the nitrogen fractional abundance for most of the velocity components. We investigate the dependence of the N/H ratio with Galactocentric distance in the inner Galaxy (R_gal<6 kpc, which is inaccessible in optical studies due to dust extinction. We find that the distribution of nitrogen abundances in the inner galaxy derived from our data has a slope that is consistent to that found in the outer Galaxy in optical studies. This result is inconsistent with some suggestions of a flatter distribution of the nitrogen abundance in the inner galaxy. △ Less

Submitted 23 September, 2019; v1 submitted 16 May, 2019; originally announced May 2019.

Comments: 20 pages, 9 figures, ApJ in press

Elucidating image-to-set prediction: An analysis of models, losses and datasets

Authors: Luis Pineda, Amaia Salvador, Michal Drozdzal, Adriana Romero

Abstract: In this paper, we identify an important reproducibility challenge in the image-to-set prediction literature that impedes proper comparisons among published methods, namely, researchers use different evaluation protocols to assess their contributions. To alleviate this issue, we introduce an image-to-set prediction benchmark suite built on top of five public datasets of increasing task complexity t… ▽ More In this paper, we identify an important reproducibility challenge in the image-to-set prediction literature that impedes proper comparisons among published methods, namely, researchers use different evaluation protocols to assess their contributions. To alleviate this issue, we introduce an image-to-set prediction benchmark suite built on top of five public datasets of increasing task complexity that are suitable for multi-label classification (VOC, COCO, NUS-WIDE, ADE20k and Recipe1M). Using the benchmark, we provide an in-depth analysis where we study the key components of current models, namely the choice of the image representation backbone as well as the set predictor design. Our results show that (1) exploiting better image representation backbones leads to higher performance boosts than enhancing set predictors, and (2) modeling both the label co-occurrences and ordering has a slight positive impact in terms of performance, whereas explicit cardinality prediction only helps when training on complex datasets, such as Recipe1M. To facilitate future image-to-set prediction research, we make the code, best models and dataset splits publicly available at: https://github.com/facebookresearch/image-to-set. △ Less

Submitted 27 May, 2020; v1 submitted 11 April, 2019; originally announced April 2019.

The Mode of Computing

Authors: Luis A. Pineda

Abstract: The Turing Machine is the paradigmatic case of computing machines, but there are others such as analogical, connectionist, quantum and diverse forms of unconventional computing, each based on a particular intuition of the phenomenon of computing. This variety can be captured in terms of system levels, re-interpreting and generalizing Newell's hierarchy, which includes the knowledge level at the to… ▽ More The Turing Machine is the paradigmatic case of computing machines, but there are others such as analogical, connectionist, quantum and diverse forms of unconventional computing, each based on a particular intuition of the phenomenon of computing. This variety can be captured in terms of system levels, re-interpreting and generalizing Newell's hierarchy, which includes the knowledge level at the top and the symbol level immediately below it. In this re-interpretation the knowledge level consists of human knowledge and the symbol level is generalized into a new level that here is called The Mode of Computing. Mental processes performed by natural brains are often thought of informally as computing process and that the brain is alike to computing machinery. However, if natural computing does exist it should be characterized on its own. A proposal to such an effect is that natural computing appeared when interpretations were first made by biological entities, so natural computing and interpreting are two aspects of the same phenomenon, or that consciousness and experience are the manifestations of computing/interpreting. By analogy with computing machinery, there must be a system level at the top of the neural circuitry and directly below the knowledge level that is named here The mode of Natural Computing. If it turns out that such putative object does not exist the proposition that the mind is a computing process should be dropped; but characterizing it would come with solving the hard problem of consciousness. △ Less

Submitted 9 October, 2023; v1 submitted 25 March, 2019; originally announced March 2019.

Comments: 47 pages, 7 figures

Probing ISM Structure in Trumpler 14 & Carina I Using The Stratospheric Terahertz Observatory 2

Authors: Young Min Seo, Paul F. Goldsmith, Chris Walker, David J. Hollenbach, Mark G. Wolfire, Craig Kulesa, Volker Tolls, Pietro N. Bernasconi, Umit Kavak, Floris F. S. van der Tak, Russ Shipman, Jian Rong Gao, Alexander Tielens, Michael G. Burton, Harold Yorke, Erick Young, William L. Peters, Abram Young, Christopher Groppi, Kristina Davis, Jorge L. Pineda, William D. Langer, Jonathan H. Kawamura, Antony Stark, Gary Melnick , et al. (4 additional authors not shown)

Abstract: We present observations of the Trumpler 14/Carina I region carried out using the Stratospheric Terahertz Observatory 2 (STO2). The Trumpler 14/Carina I region is in the west part of the Carina Nebula Complex, which is one of the most extreme star-forming regions in the Milky Way. We observed Trumpler 14/Carina I in the 158 $μ$m transition of [C\,{\sc ii}] with a spatial resolution of 48$''$ and a… ▽ More We present observations of the Trumpler 14/Carina I region carried out using the Stratospheric Terahertz Observatory 2 (STO2). The Trumpler 14/Carina I region is in the west part of the Carina Nebula Complex, which is one of the most extreme star-forming regions in the Milky Way. We observed Trumpler 14/Carina I in the 158 $μ$m transition of [C\,{\sc ii}] with a spatial resolution of 48$''$ and a velocity resolution of 0.17 km s$^{-1}$. The observations cover a 0.25$^\circ$ by 0.28$^\circ$ area with central position {\it l} = 297.34$^\circ$, {\it b} = -0.60$^\circ$. The kinematics show that bright [C\,{\sc ii}] structures are spatially and spectrally correlated with the surfaces of CO clouds, tracing the photodissociation region and ionization front of each molecular cloud. Along 7 lines of sight that traverse Tr 14 into the dark ridge to the southwest, we find that the [C\,{\sc ii}] luminosity from the HII region is 3.7 times that from the PDR. In same los we find in the PDRs an average ratio of 1:4.1:5.6 for the mass in atomic gas:dark-CO gas: molecular gas traced by CO. Comparing multiple gas tracers including HI 21cm, [C\,{\sc ii}], CO, and radio recombination lines, we find that the HII regions of the Carina Nebula Complex are well-described as HII regions with one-side freely expanding towards us, consistent with the champagne model of ionized gas evolution. The dispersal of the GMC in this region is dominated by EUV photoevaporation; the dispersal timescale is 20-30 Myr. △ Less

Submitted 22 May, 2019; v1 submitted 22 March, 2019; originally announced March 2019.

Comments: ApJ accepted

New combinational therapies for cancer using modern statistical mechanics

Authors: Jorge A. González, M. Acanda, Z. Akhtar, D. Andrews, J. I. Azqueta, E. Bass, A. Bellorín, J. Couso, Mónica A. García-Ñustes, Y. Infante, S. Jiménez, L. Lester, L. Maldonado, Juan F. Marín, L. Pineda, I. Rodríguez, C. C. Tamayo, D. Valdes, L. Vázquez

Abstract: We investigate a new dynamical system that describes tumor-host interaction. The equation that describes the untreated tumor growth is based on non-extensive statistical mechanics. Recently, this model has been shown to fit successfully exponential, Gompertz, logistic, and power-law tumor growths. We have been able to include as many hallmarks of cancer as possible. We study also the dynamic respo… ▽ More We investigate a new dynamical system that describes tumor-host interaction. The equation that describes the untreated tumor growth is based on non-extensive statistical mechanics. Recently, this model has been shown to fit successfully exponential, Gompertz, logistic, and power-law tumor growths. We have been able to include as many hallmarks of cancer as possible. We study also the dynamic response of cancer under therapy. Using our model, we can make predictions about the different outcomes when we change the parameters, and/or the initial conditions. We can determine the importance of different factors to influence tumor growth. We discover synergistic therapeutic effects of different treatments and drugs. Cancer is generally untreatable using conventional monotherapy. We consider conventional therapies, oncogene-targeted therapies, tumor-suppressors gene-targeted therapies, immunotherapies, anti-angiogenesis therapies, virotherapy, among others. We need therapies with the potential to target both tumor cells and the tumors' microenvironment. Drugs that target oncogenes and tumor-suppressor genes can be effective in the treatment of some cancers. However, most tumors do reoccur. We have found that the success of the new therapeutic agents can be seen when used in combination with other cancer-cell-killing therapies. Our results have allowed us to design a combinational therapy that can lead to the complete eradication of cancer. △ Less

Submitted 2 February, 2019; originally announced February 2019.

Comments: 35 pages, 6 figures

A SOFIA Survey of [CII] in the galaxy M51 I. [CII] as a tracer of Star Formation

Authors: Jorge L. Pineda, Christian Fischer, Maria Kapala, Juergen Stutzki, Christof Buchbender, Paul F. Goldsmith, Monika Ziebart, Simon C. O. Glover, Ralf S. Klessen, Jin Koda, Carsten Kramer, Bhaswati Mookerjea, Karin Sandstrom, Nick Scoville, Rowan Smith

Abstract: We present a [CII] 158um map of the entire M51 (including M51b) grand--design spiral galaxy observed with the FIFI-LS instrument on SOFIA. We compare the [CII] emission with the total far--infrared (TIR) intensity and star formation rate(SFR) surface density maps (derived using H_alpha and 24um emission) to study the relationship between [CII] and the star formation activity in a variety of enviro… ▽ More We present a [CII] 158um map of the entire M51 (including M51b) grand--design spiral galaxy observed with the FIFI-LS instrument on SOFIA. We compare the [CII] emission with the total far--infrared (TIR) intensity and star formation rate(SFR) surface density maps (derived using H_alpha and 24um emission) to study the relationship between [CII] and the star formation activity in a variety of environments within M51 on scales of 16" corresponding to ~660 pc. We find that [CII] and the SFR surface density are well correlated in the central, spiral arm, and inter-arm regions. The correlation is in good agreement with that found for a larger sample of nearby galaxies at kpc scales. We find that the SFR, and [CII] and TIR luminosities in M51 are dominated by the extended emission in M51's disk. The companion galaxy M51b, however, shows a deficit of [CII] emission compared with the TIR emission and SFR surface density, with [CII] emission detected only in the S-W part of this galaxy. The [CII] deficit is associated with an enhanced dust temperature in this galaxy. We interpret the faint [CII] emission in M51b to be a result of suppressed star formation in this galaxy, while the bright mid- and far-infrared emission, which drive the TIR and SFR values, are powered by other mechanisms. A similar but less pronounced effect is seen at the location of the black hole in M51's center. The observed [CII] deficit in M51b suggests that this galaxy is a valuable laboratory to study the origin of the apparent [CII] deficit observed in ultra-luminous galaxies. △ Less

Submitted 28 November, 2018; originally announced November 2018.

Comments: 10 pages, 4 figures, Accepted for publication in Apj Letters

Planning in Stochastic Environments with Goal Uncertainty

Authors: Sandhya Saisubramanian, Kyle Hollins Wray, Luis Pineda, Shlomo Zilberstein

Abstract: We present the Goal Uncertain Stochastic Shortest Path (GUSSP) problem -- a general framework to model path planning and decision making in stochastic environments with goal uncertainty. The framework extends the stochastic shortest path (SSP) model to dynamic environments in which it is impossible to determine the exact goal states ahead of plan execution. GUSSPs introduce flexibility in goal spe… ▽ More We present the Goal Uncertain Stochastic Shortest Path (GUSSP) problem -- a general framework to model path planning and decision making in stochastic environments with goal uncertainty. The framework extends the stochastic shortest path (SSP) model to dynamic environments in which it is impossible to determine the exact goal states ahead of plan execution. GUSSPs introduce flexibility in goal specification by allowing a belief over possible goal configurations. The unique observations at potential goals helps the agent identify the true goal during plan execution. The partial observability is restricted to goals, facilitating the reduction to an SSP with a modified state space. We formally define a GUSSP and discuss its theoretical properties. We then propose an admissible heuristic that reduces the planning time using FLARES -- a start-of-the-art probabilistic planner. We also propose a determinization approach for solving this class of problems. Finally, we present empirical results on a search and rescue mobile robot and three other problem domains in simulation. △ Less

Submitted 3 April, 2020; v1 submitted 18 October, 2018; originally announced October 2018.

Comments: 6 pages, IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2019

Journal ref: 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Macau, China, 2019, pp. 1649-1654

DeepTag: inferring all-cause diagnoses from clinical notes in under-resourced medical domain

Authors: Allen Nie, Ashley Zehnder, Rodney L. Page, Arturo L. Pineda, Manuel A. Rivas, Carlos D. Bustamante, James Zou

Abstract: Large scale veterinary clinical records can become a powerful resource for patient care and research. However, clinicians lack the time and resource to annotate patient records with standard medical diagnostic codes and most veterinary visits are captured in free text notes. The lack of standard coding makes it challenging to use the clinical data to improve patient care. It is also a major impedi… ▽ More Large scale veterinary clinical records can become a powerful resource for patient care and research. However, clinicians lack the time and resource to annotate patient records with standard medical diagnostic codes and most veterinary visits are captured in free text notes. The lack of standard coding makes it challenging to use the clinical data to improve patient care. It is also a major impediment to cross-species translational research, which relies on the ability to accurately identify patient cohorts with specific diagnostic criteria in humans and animals. In order to reduce the coding burden for veterinary clinical practice and aid translational research, we have developed a deep learning algorithm, DeepTag, which automatically infers diagnostic codes from veterinary free text notes. DeepTag is trained on a newly curated dataset of 112,558 veterinary notes manually annotated by experts. DeepTag extends multi-task LSTM with an improved hierarchical objective that captures the semantic structures between diseases. To foster human-machine collaboration, DeepTag also learns to abstain in examples when it is uncertain and defers them to human experts, resulting in improved performance. DeepTag accurately infers disease codes from free text even in challenging cross-hospital settings where the text comes from different clinical settings than the ones used for training. It enables automated disease annotation across a broad range of clinical diagnoses with minimal pre-processing. The technical framework in this work can be applied in other medical domains that currently lack medical coding resources. △ Less

Submitted 3 September, 2018; v1 submitted 27 June, 2018; originally announced June 2018.

Comments: 17 pages, 6 figures. Updated the text for clarity

The Connection Between Different Tracers Of The Diffuse Interstellar Medium: Kinematics

Authors: Johnathan S. Rice, S. R. Federman, Nicolas Flagey, Paul F. Goldsmith, William D. Langer, Jorge L. Pineda, D. L. Lambert

Abstract: Using visible, radio, microwave, and sub-mm data, we study several lines of sight toward stars generally closer than 1 kpc on a component-by-component basis. We derive the component structure seen in absorption at visible wavelengths from Ca II, Ca I, K I, CH, CH$^{+}\!,$ and CN and compare it to emission from H I, CO and its isotopologues, and C$^{+}$ from the GOT C+ survey. The correspondence be… ▽ More Using visible, radio, microwave, and sub-mm data, we study several lines of sight toward stars generally closer than 1 kpc on a component-by-component basis. We derive the component structure seen in absorption at visible wavelengths from Ca II, Ca I, K I, CH, CH$^{+}\!,$ and CN and compare it to emission from H I, CO and its isotopologues, and C$^{+}$ from the GOT C+ survey. The correspondence between components in emission and absorption help create a more unified picture of diffuse atomic and molecular gas in the interstellar medium. We also discuss how these tracers are related to the CO-dark H$_{2}$ gas probed by C$^{+}$ emission and discuss the kinematic connections among the species observed. △ Less

Submitted 18 April, 2018; originally announced April 2018.

Comments: To be published in ApJ

The CARMA-NRO Orion Survey

Authors: Shuo Kong, Héctor G. Arce, Jesse R. Feddersen, John M. Carpenter, Fumitaka Nakamura, Yoshito Shimajiri, Andrea Isella, Volker Ossenkopf-Okada, Anneila I. Sargent, Álvaro Sánchez-Monge, Sümeyye T. Suri, Jens Kauffmann, Thushara Pillai, Jaime E. Pineda, Jin Koda, John Bally, Dariusz C. Lis, Paolo Padoan, Ralf Klessen, Steve Mairs, Alyssa Goodman, Paul Goldsmith, Peregrine McGehee, Peter Schilke, Peter J. Teuben , et al. (13 additional authors not shown)

Abstract: We present the first results from a new, high resolution, $^{12}$CO(1-0), $^{13}$CO(1-0), and C$^{18}$O(1-0) molecular line survey of the Orion A cloud, hereafter referred to as the CARMA-NRO Orion Survey. CARMA observations have been combined with single-dish data from the Nobeyama 45m telescope to provide extended images at about 0.01 pc resolution, with a dynamic range of approximately 1200 in… ▽ More We present the first results from a new, high resolution, $^{12}$CO(1-0), $^{13}$CO(1-0), and C$^{18}$O(1-0) molecular line survey of the Orion A cloud, hereafter referred to as the CARMA-NRO Orion Survey. CARMA observations have been combined with single-dish data from the Nobeyama 45m telescope to provide extended images at about 0.01 pc resolution, with a dynamic range of approximately 1200 in spatial scale. Here we describe the practical details of the data combination in uv space, including flux scale matching, the conversion of single dish data to visibilities, and joint deconvolution of single dish and interferometric data. A $Δ$-variance analysis indicates that no artifacts are caused by combining data from the two instruments. Initial analysis of the data cubes, including moment maps, average spectra, channel maps, position-velocity diagrams, excitation temperature, column density, and line ratio maps provides evidence of complex and interesting structures such as filaments, bipolar outflows, shells, bubbles, and photo-eroded pillars. The implications for star formation processes are profound and follow-up scientific studies by the CARMA-NRO Orion team are now underway. We plan to make all the data products described here generally accessible; some are already available at https://dataverse.harvard.edu/dataverse/CARMA-NRO-Orion △ Less

Submitted 30 March, 2018; originally announced March 2018.

Comments: 46 pages, 28 figures, 2 tables, accepted by ApJS