-
vailá: Versatile Anarcho Integrated Liberation Ánalysis in Multimodal Toolbox
Authors:
Paulo Roberto Pereira Santiago,
Abel Gonçalves Chinaglia,
Kira Flanagan,
Bruno L. S. Bedo,
Ligia Yumi Mochida,
Juan Aceros,
Aline Bononi,
Guilherme Manna Cesar
Abstract:
Human movement analysis is crucial in health and sports biomechanics for understanding physical performance, guiding rehabilitation, and preventing injuries. However, existing tools are often proprietary, expensive, and function as "black boxes", limiting user control and customization. This paper introduces vailá-Versatile Anarcho Integrated Liberation Ánalysis in Multimodal Toolbox-an open-sourc…
▽ More
Human movement analysis is crucial in health and sports biomechanics for understanding physical performance, guiding rehabilitation, and preventing injuries. However, existing tools are often proprietary, expensive, and function as "black boxes", limiting user control and customization. This paper introduces vailá-Versatile Anarcho Integrated Liberation Ánalysis in Multimodal Toolbox-an open-source, Python-based platform designed to enhance human movement analysis by integrating data from multiple biomechanical systems. vailá supports data from diverse sources, including retroreflective motion capture systems, inertial measurement units (IMUs), markerless video capture technology, electromyography (EMG), force plates, and GPS or GNSS systems, enabling comprehensive analysis of movement patterns. Developed entirely in Python 3.11.9, which offers improved efficiency and long-term support, and featuring a straightforward installation process, vailá is accessible to users without extensive programming experience. In this paper, we also present several workflow examples that demonstrate how vailá allows the rapid processing of large batches of data, independent of the type of collection method. This flexibility is especially valuable in research scenarios where unexpected data collection challenges arise, ensuring no valuable data point is lost. We demonstrate the application of vailá in analyzing sit-to-stand movements in pediatric disability, showcasing its capability to provide deeper insights even with unexpected movement patterns. By fostering a collaborative and open environment, vailá encourages users to innovate, customize, and freely explore their analysis needs, potentially contributing to the advancement of rehabilitation strategies and performance optimization.
△ Less
Submitted 7 October, 2024;
originally announced October 2024.
-
Unmasking Biases and Navigating Pitfalls in the Ophthalmic Artificial Intelligence Lifecycle: A Review
Authors:
Luis Filipe Nakayama,
João Matos,
Justin Quion,
Frederico Novaes,
William Greig Mitchell,
Rogers Mwavu,
Ju-Yi Ji Hung,
Alvina Pauline dy Santiago,
Warachaya Phanphruk,
Jaime S. Cardoso,
Leo Anthony Celi
Abstract:
Over the past two decades, exponential growth in data availability, computational power, and newly available modeling techniques has led to an expansion in interest, investment, and research in Artificial Intelligence (AI) applications. Ophthalmology is one of many fields that seek to benefit from AI given the advent of telemedicine screening programs and the use of ancillary imaging. However, bef…
▽ More
Over the past two decades, exponential growth in data availability, computational power, and newly available modeling techniques has led to an expansion in interest, investment, and research in Artificial Intelligence (AI) applications. Ophthalmology is one of many fields that seek to benefit from AI given the advent of telemedicine screening programs and the use of ancillary imaging. However, before AI can be widely deployed, further work must be done to avoid the pitfalls within the AI lifecycle. This review article breaks down the AI lifecycle into seven steps: data collection; defining the model task; data pre-processing and labeling; model development; model evaluation and validation; deployment; and finally, post-deployment evaluation, monitoring, and system recalibration and delves into the risks for harm at each step and strategies for mitigating them.
△ Less
Submitted 7 October, 2023;
originally announced October 2023.
-
Symbolic Regression via Neural-Guided Genetic Programming Population Seeding
Authors:
T. Nathan Mundhenk,
Mikel Landajuela,
Ruben Glatt,
Claudio P. Santiago,
Daniel M. Faissol,
Brenden K. Petersen
Abstract:
Symbolic regression is the process of identifying mathematical expressions that fit observed output from a black-box process. It is a discrete optimization problem generally believed to be NP-hard. Prior approaches to solving the problem include neural-guided search (e.g. using reinforcement learning) and genetic programming. In this work, we introduce a hybrid neural-guided/genetic programming ap…
▽ More
Symbolic regression is the process of identifying mathematical expressions that fit observed output from a black-box process. It is a discrete optimization problem generally believed to be NP-hard. Prior approaches to solving the problem include neural-guided search (e.g. using reinforcement learning) and genetic programming. In this work, we introduce a hybrid neural-guided/genetic programming approach to symbolic regression and other combinatorial optimization problems. We propose a neural-guided component used to seed the starting population of a random restart genetic programming component, gradually learning better starting populations. On a number of common benchmark tasks to recover underlying expressions from a dataset, our method recovers 65% more expressions than a recently published top-performing model using the same experimental setup. We demonstrate that running many genetic programming generations without interdependence on the neural-guided component performs better for symbolic regression than alternative formulations where the two are more strongly coupled. Finally, we introduce a new set of 22 symbolic regression benchmark problems with increased difficulty over existing benchmarks. Source code is provided at www.github.com/brendenpetersen/deep-symbolic-optimization.
△ Less
Submitted 17 November, 2021; v1 submitted 29 October, 2021;
originally announced November 2021.
-
Incorporating domain knowledge into neural-guided search
Authors:
Brenden K. Petersen,
Claudio P. Santiago,
Mikel Landajuela
Abstract:
Many AutoML problems involve optimizing discrete objects under a black-box reward. Neural-guided search provides a flexible means of searching these combinatorial spaces using an autoregressive recurrent neural network. A major benefit of this approach is that builds up objects sequentially--this provides an opportunity to incorporate domain knowledge into the search by directly modifying the logi…
▽ More
Many AutoML problems involve optimizing discrete objects under a black-box reward. Neural-guided search provides a flexible means of searching these combinatorial spaces using an autoregressive recurrent neural network. A major benefit of this approach is that builds up objects sequentially--this provides an opportunity to incorporate domain knowledge into the search by directly modifying the logits emitted during sampling. In this work, we formalize a framework for incorporating such in situ priors and constraints into neural-guided search, and provide sufficient conditions for enforcing constraints. We integrate several priors and constraints from existing works into this framework, propose several new ones, and demonstrate their efficacy in informing the task of symbolic regression.
△ Less
Submitted 19 July, 2021;
originally announced July 2021.
-
Improving exploration in policy gradient search: Application to symbolic optimization
Authors:
Mikel Landajuela,
Brenden K. Petersen,
Soo K. Kim,
Claudio P. Santiago,
Ruben Glatt,
T. Nathan Mundhenk,
Jacob F. Pettit,
Daniel M. Faissol
Abstract:
Many machine learning strategies designed to automate mathematical tasks leverage neural networks to search large combinatorial spaces of mathematical symbols. In contrast to traditional evolutionary approaches, using a neural network at the core of the search allows learning higher-level symbolic patterns, providing an informed direction to guide the search. When no labeled data is available, suc…
▽ More
Many machine learning strategies designed to automate mathematical tasks leverage neural networks to search large combinatorial spaces of mathematical symbols. In contrast to traditional evolutionary approaches, using a neural network at the core of the search allows learning higher-level symbolic patterns, providing an informed direction to guide the search. When no labeled data is available, such networks can still be trained using reinforcement learning. However, we demonstrate that this approach can suffer from an early commitment phenomenon and from initialization bias, both of which limit exploration. We present two exploration methods to tackle these issues, building upon ideas of entropy regularization and distribution initialization. We show that these techniques can improve the performance, increase sample efficiency, and lower the complexity of solutions for the task of symbolic regression.
△ Less
Submitted 19 July, 2021;
originally announced July 2021.
-
Competition between magnetic order and charge localization in Na$_2$IrO$_3$ thin crystal devices
Authors:
Josue Rodriguez,
Gilbert Lopez,
Samantha Crouch,
Nicholas P. Breznay,
Robert Kealhofer,
Vikram Nagarajan,
Drew Latzke,
Francisco Ramirez,
Naomy Marrufo,
Peter Santiago,
Jared Lara,
Amirari Diego,
Everardo Molina,
David Rosser,
Hadi Tavassol,
Alessandra Lanzara,
James G. Analytis,
Claudia Ojeda-Aristizabal
Abstract:
Spin orbit assisted Mott insulators such as sodium iridate (Na$_2$IrO$_3$) have been an important subject of study in the recent years. In these materials, the interplay of electronic correlations, spin-orbit coupling, crystal field effects and a honeycomb arrangement of ions bring exciting ground states, predicted in the frame of the Kitaev model. The insulating character of Na$_2$IrO$_3$ has ham…
▽ More
Spin orbit assisted Mott insulators such as sodium iridate (Na$_2$IrO$_3$) have been an important subject of study in the recent years. In these materials, the interplay of electronic correlations, spin-orbit coupling, crystal field effects and a honeycomb arrangement of ions bring exciting ground states, predicted in the frame of the Kitaev model. The insulating character of Na$_2$IrO$_3$ has hampered its integration to an electronic device, desirable for applications, such as the manipulation of quasiparticles interesting for topological quantum computing. Here we show through electronic transport measurements supported by Angle Resolved Photoemission Spectroscopy (ARPES) experiments, that electronic transport in Na$_2$IrO$_3$ is ruled by variable range hopping and it is strongly dependent on the magnetic ordering transition known for bulk Na$_2$IrO$_3$, as well as on external electric fields. Electronic transport measurements allow us to deduce a value for the localization length and the density of states in our Na$_2$IrO$_3$ thin crystals devices, offering an alternative approach to study insulating layered materials.
△ Less
Submitted 11 February, 2020;
originally announced February 2020.
-
Deep symbolic regression: Recovering mathematical expressions from data via risk-seeking policy gradients
Authors:
Brenden K. Petersen,
Mikel Landajuela,
T. Nathan Mundhenk,
Claudio P. Santiago,
Soo K. Kim,
Joanne T. Kim
Abstract:
Discovering the underlying mathematical expressions describing a dataset is a core challenge for artificial intelligence. This is the problem of $\textit{symbolic regression}$. Despite recent advances in training neural networks to solve complex tasks, deep learning approaches to symbolic regression are underexplored. We propose a framework that leverages deep learning for symbolic regression via…
▽ More
Discovering the underlying mathematical expressions describing a dataset is a core challenge for artificial intelligence. This is the problem of $\textit{symbolic regression}$. Despite recent advances in training neural networks to solve complex tasks, deep learning approaches to symbolic regression are underexplored. We propose a framework that leverages deep learning for symbolic regression via a simple idea: use a large model to search the space of small models. Specifically, we use a recurrent neural network to emit a distribution over tractable mathematical expressions and employ a novel risk-seeking policy gradient to train the network to generate better-fitting expressions. Our algorithm outperforms several baseline methods (including Eureqa, the gold standard for symbolic regression) in its ability to exactly recover symbolic expressions on a series of benchmark problems, both with and without added noise. More broadly, our contributions include a framework that can be applied to optimize hierarchical, variable-length objects under a black-box performance metric, with the ability to incorporate constraints in situ, and a risk-seeking policy gradient formulation that optimizes for best-case performance instead of expected performance.
△ Less
Submitted 5 April, 2021; v1 submitted 10 December, 2019;
originally announced December 2019.
-
Assisted-hydrothermal Synthesis and Characterization of Flower-like ZnO Nanostructures
Authors:
S. Lopez-Romero,
P. Santiago,
D. Mendoza
Abstract:
Flower-like nanostructures formed by ZnO nanorods were synthesized and deposited on seeded silicon and glass substrates by a hexamethylenetetramine (HMTA) - assisted hydrothermal method at low temperature (90 oC) with methenamine ((CH3)6N4), as surfactant and catalyst. The substrates were seeded with ZnO nanoparticles. The structure and morphology of the nanostructures were studied by means of x-r…
▽ More
Flower-like nanostructures formed by ZnO nanorods were synthesized and deposited on seeded silicon and glass substrates by a hexamethylenetetramine (HMTA) - assisted hydrothermal method at low temperature (90 oC) with methenamine ((CH3)6N4), as surfactant and catalyst. The substrates were seeded with ZnO nanoparticles. The structure and morphology of the nanostructures were studied by means of x-ray diffraction (XRD), high resolution transmission electron microscopy (HRTEM), and scanning electron microscopy (SEM) techniques. Influence of the seed nanoparticle on the formation of the flower-like ZnO nanostructures is demonstrated. The influence of the organic oxygenated chains on the crystalline habit during the growth process is also observed.
△ Less
Submitted 26 April, 2012;
originally announced April 2012.
