-
DriveRX: A Vision-Language Reasoning Model for Cross-Task Autonomous Driving
Authors:
Muxi Diao,
Lele Yang,
Hongbo Yin,
Zhexu Wang,
Yejie Wang,
Daxin Tian,
Kongming Liang,
Zhanyu Ma
Abstract:
Autonomous driving requires real-time, robust reasoning across perception, prediction, planning, and behavior. However, conventional end-to-end models fail to generalize in complex scenarios due to the lack of structured reasoning. Recent vision-language models (VLMs) have been applied to driving tasks, but they typically rely on isolated modules and static supervision, limiting their ability to s…
▽ More
Autonomous driving requires real-time, robust reasoning across perception, prediction, planning, and behavior. However, conventional end-to-end models fail to generalize in complex scenarios due to the lack of structured reasoning. Recent vision-language models (VLMs) have been applied to driving tasks, but they typically rely on isolated modules and static supervision, limiting their ability to support multi-stage decision-making. We present AutoDriveRL, a unified training framework that formulates autonomous driving as a structured reasoning process over four core tasks. Each task is independently modeled as a vision-language question-answering problem and optimized using task-specific reward models, enabling fine-grained reinforcement signals at different reasoning stages. Within this framework, we train DriveRX, a cross-task reasoning VLM designed for real-time decision-making. DriveRX achieves strong performance on a public benchmark, outperforming GPT-4o in behavior reasoning and demonstrating robustness under complex or corrupted driving conditions. Our analysis further highlights the impact of vision encoder design and reward-guided reasoning compression. We will release the AutoDriveRL framework and the DriveRX model to support future research.
△ Less
Submitted 26 May, 2025;
originally announced May 2025.
-
Harnessing Caption Detailness for Data-Efficient Text-to-Image Generation
Authors:
Xinran Wang,
Muxi Diao,
Yuanzhi Liu,
Chunyu Wang,
Kongming Liang,
Zhanyu Ma,
Jun Guo
Abstract:
Training text-to-image (T2I) models with detailed captions can significantly improve their generation quality. Existing methods often rely on simplistic metrics like caption length to represent the detailness of the caption in the T2I training set. In this paper, we propose a new metric to estimate caption detailness based on two aspects: image coverage rate (ICR), which evaluates whether the capt…
▽ More
Training text-to-image (T2I) models with detailed captions can significantly improve their generation quality. Existing methods often rely on simplistic metrics like caption length to represent the detailness of the caption in the T2I training set. In this paper, we propose a new metric to estimate caption detailness based on two aspects: image coverage rate (ICR), which evaluates whether the caption covers all regions/objects in the image, and average object detailness (AOD), which quantifies the detailness of each object's description. Through experiments on the COCO dataset using ShareGPT4V captions, we demonstrate that T2I models trained on high-ICR and -AOD captions achieve superior performance on DPG and other benchmarks. Notably, our metric enables more effective data selection-training on only 20% of full data surpasses both full-dataset training and length-based selection method, improving alignment and reconstruction ability. These findings highlight the critical role of detail-aware metrics over length-based heuristics in caption selection for T2I tasks.
△ Less
Submitted 21 May, 2025;
originally announced May 2025.
-
CineTechBench: A Benchmark for Cinematographic Technique Understanding and Generation
Authors:
Xinran Wang,
Songyu Xu,
Xiangxuan Shan,
Yuxuan Zhang,
Muxi Diao,
Xueyan Duan,
Yanhua Huang,
Kongming Liang,
Zhanyu Ma
Abstract:
Cinematography is a cornerstone of film production and appreciation, shaping mood, emotion, and narrative through visual elements such as camera movement, shot composition, and lighting. Despite recent progress in multimodal large language models (MLLMs) and video generation models, the capacity of current models to grasp and reproduce cinematographic techniques remains largely uncharted, hindered…
▽ More
Cinematography is a cornerstone of film production and appreciation, shaping mood, emotion, and narrative through visual elements such as camera movement, shot composition, and lighting. Despite recent progress in multimodal large language models (MLLMs) and video generation models, the capacity of current models to grasp and reproduce cinematographic techniques remains largely uncharted, hindered by the scarcity of expert-annotated data. To bridge this gap, we present CineTechBench, a pioneering benchmark founded on precise, manual annotation by seasoned cinematography experts across key cinematography dimensions. Our benchmark covers seven essential aspects-shot scale, shot angle, composition, camera movement, lighting, color, and focal length-and includes over 600 annotated movie images and 120 movie clips with clear cinematographic techniques. For the understanding task, we design question answer pairs and annotated descriptions to assess MLLMs' ability to interpret and explain cinematographic techniques. For the generation task, we assess advanced video generation models on their capacity to reconstruct cinema-quality camera movements given conditions such as textual prompts or keyframes. We conduct a large-scale evaluation on 15+ MLLMs and 5+ video generation models. Our results offer insights into the limitations of current models and future directions for cinematography understanding and generation in automatically film production and appreciation. The code and benchmark can be accessed at https://github.com/PRIS-CV/CineTechBench.
△ Less
Submitted 21 May, 2025;
originally announced May 2025.
-
SAND: One-Shot Feature Selection with Additive Noise Distortion
Authors:
Pedram Pad,
Hadi Hammoud,
Mohamad Dia,
Nadim Maamari,
L. Andrea Dunbar
Abstract:
Feature selection is a critical step in data-driven applications, reducing input dimensionality to enhance learning accuracy, computational efficiency, and interpretability. Existing state-of-the-art methods often require post-selection retraining and extensive hyperparameter tuning, complicating their adoption. We introduce a novel, non-intrusive feature selection layer that, given a target featu…
▽ More
Feature selection is a critical step in data-driven applications, reducing input dimensionality to enhance learning accuracy, computational efficiency, and interpretability. Existing state-of-the-art methods often require post-selection retraining and extensive hyperparameter tuning, complicating their adoption. We introduce a novel, non-intrusive feature selection layer that, given a target feature count $k$, automatically identifies and selects the $k$ most informative features during neural network training. Our method is uniquely simple, requiring no alterations to the loss function, network architecture, or post-selection retraining. The layer is mathematically elegant and can be fully described by: \begin{align} \nonumber \tilde{x}_i = a_i x_i + (1-a_i)z_i \end{align} where $x_i$ is the input feature, $\tilde{x}_i$ the output, $z_i$ a Gaussian noise, and $a_i$ trainable gain such that $\sum_i{a_i^2}=k$. This formulation induces an automatic clustering effect, driving $k$ of the $a_i$ gains to $1$ (selecting informative features) and the rest to $0$ (discarding redundant ones) via weighted noise distortion and gain normalization. Despite its extreme simplicity, our method delivers state-of-the-art performance on standard benchmark datasets and a novel real-world dataset, outperforming or matching existing approaches without requiring hyperparameter search for $k$ or retraining. Theoretical analysis in the context of linear regression further validates its efficacy. Our work demonstrates that simplicity and performance are not mutually exclusive, offering a powerful yet straightforward tool for feature selection in machine learning.
△ Less
Submitted 6 May, 2025;
originally announced May 2025.
-
From Simple to Professional: A Combinatorial Controllable Image Captioning Agent
Authors:
Xinran Wang,
Muxi Diao,
Baoteng Li,
Haiwen Zhang,
Kongming Liang,
Zhanyu Ma
Abstract:
The Controllable Image Captioning Agent (CapAgent) is an innovative system designed to bridge the gap between user simplicity and professional-level outputs in image captioning tasks. CapAgent automatically transforms user-provided simple instructions into detailed, professional instructions, enabling precise and context-aware caption generation. By leveraging multimodal large language models (MLL…
▽ More
The Controllable Image Captioning Agent (CapAgent) is an innovative system designed to bridge the gap between user simplicity and professional-level outputs in image captioning tasks. CapAgent automatically transforms user-provided simple instructions into detailed, professional instructions, enabling precise and context-aware caption generation. By leveraging multimodal large language models (MLLMs) and external tools such as object detection tool and search engines, the system ensures that captions adhere to specified guidelines, including sentiment, keywords, focus, and formatting. CapAgent transparently controls each step of the captioning process, and showcases its reasoning and tool usage at every step, fostering user trust and engagement. The project code is available at https://github.com/xin-ran-w/CapAgent.
△ Less
Submitted 14 December, 2024;
originally announced December 2024.
-
How Do Your Code LLMs Perform? Empowering Code Instruction Tuning with High-Quality Data
Authors:
Yejie Wang,
Keqing He,
Dayuan Fu,
Zhuoma Gongque,
Heyang Xu,
Yanxu Chen,
Zhexu Wang,
Yujia Fu,
Guanting Dong,
Muxi Diao,
Jingang Wang,
Mengdi Zhang,
Xunliang Cai,
Weiran Xu
Abstract:
Recently, there has been a growing interest in studying how to construct better code instruction tuning data. However, we observe Code models trained with these datasets exhibit high performance on HumanEval but perform worse on other benchmarks such as LiveCodeBench. Upon further investigation, we find that many datasets suffer from severe data leakage. After cleaning up most of the leaked data,…
▽ More
Recently, there has been a growing interest in studying how to construct better code instruction tuning data. However, we observe Code models trained with these datasets exhibit high performance on HumanEval but perform worse on other benchmarks such as LiveCodeBench. Upon further investigation, we find that many datasets suffer from severe data leakage. After cleaning up most of the leaked data, some well-known high-quality datasets perform poorly. This discovery reveals a new challenge: identifying which dataset genuinely qualify as high-quality code instruction data. To address this, we propose an efficient code data pruning strategy for selecting good samples. Our approach is based on three dimensions: instruction complexity, response quality, and instruction diversity. Based on our selected data, we present XCoder, a family of models finetuned from LLaMA3. Our experiments show XCoder achieves new state-of-the-art performance using fewer training data, which verify the effectiveness of our data strategy. Moreover, we perform a comprehensive analysis on the data composition and find existing code datasets have different characteristics according to their construction methods, which provide new insights for future code LLMs. Our models and dataset are released in https://github.com/banksy23/XCoder
△ Less
Submitted 5 September, 2024;
originally announced September 2024.
-
SEAS: Self-Evolving Adversarial Safety Optimization for Large Language Models
Authors:
Muxi Diao,
Rumei Li,
Shiyang Liu,
Guogang Liao,
Jingang Wang,
Xunliang Cai,
Weiran Xu
Abstract:
As large language models (LLMs) continue to advance in capability and influence, ensuring their security and preventing harmful outputs has become crucial. A promising approach to address these concerns involves training models to automatically generate adversarial prompts for red teaming. However, the evolving subtlety of vulnerabilities in LLMs challenges the effectiveness of current adversarial…
▽ More
As large language models (LLMs) continue to advance in capability and influence, ensuring their security and preventing harmful outputs has become crucial. A promising approach to address these concerns involves training models to automatically generate adversarial prompts for red teaming. However, the evolving subtlety of vulnerabilities in LLMs challenges the effectiveness of current adversarial methods, which struggle to specifically target and explore the weaknesses of these models. To tackle these challenges, we introduce the $\mathbf{S}\text{elf-}\mathbf{E}\text{volving }\mathbf{A}\text{dversarial }\mathbf{S}\text{afety }\mathbf{(SEAS)}$ optimization framework, which enhances security by leveraging data generated by the model itself. SEAS operates through three iterative stages: Initialization, Attack, and Adversarial Optimization, refining both the Red Team and Target models to improve robustness and safety. This framework reduces reliance on manual testing and significantly enhances the security capabilities of LLMs. Our contributions include a novel adversarial framework, a comprehensive safety dataset, and after three iterations, the Target model achieves a security level comparable to GPT-4, while the Red Team model shows a marked increase in attack success rate (ASR) against advanced models. Our code and datasets are released at https://SEAS-LLM.github.io/.
△ Less
Submitted 23 December, 2024; v1 submitted 5 August, 2024;
originally announced August 2024.
-
We-Math: Does Your Large Multimodal Model Achieve Human-like Mathematical Reasoning?
Authors:
Runqi Qiao,
Qiuna Tan,
Guanting Dong,
Minhui Wu,
Chong Sun,
Xiaoshuai Song,
Zhuoma GongQue,
Shanglin Lei,
Zhe Wei,
Miaoxuan Zhang,
Runfeng Qiao,
Yifan Zhang,
Xiao Zong,
Yida Xu,
Muxi Diao,
Zhimin Bao,
Chen Li,
Honggang Zhang
Abstract:
Visual mathematical reasoning, as a fundamental visual reasoning ability, has received widespread attention from the Large Multimodal Models (LMMs) community. Existing benchmarks, such as MathVista and MathVerse, focus more on the result-oriented performance but neglect the underlying principles in knowledge acquisition and generalization. Inspired by human-like mathematical reasoning, we introduc…
▽ More
Visual mathematical reasoning, as a fundamental visual reasoning ability, has received widespread attention from the Large Multimodal Models (LMMs) community. Existing benchmarks, such as MathVista and MathVerse, focus more on the result-oriented performance but neglect the underlying principles in knowledge acquisition and generalization. Inspired by human-like mathematical reasoning, we introduce WE-MATH, the first benchmark specifically designed to explore the problem-solving principles beyond end-to-end performance. We meticulously collect and categorize 6.5K visual math problems, spanning 67 hierarchical knowledge concepts and five layers of knowledge granularity. We decompose composite problems into sub-problems according to the required knowledge concepts and introduce a novel four-dimensional metric, namely Insufficient Knowledge (IK), Inadequate Generalization (IG), Complete Mastery (CM), and Rote Memorization (RM), to hierarchically assess inherent issues in LMMs' reasoning process. With WE-MATH, we conduct a thorough evaluation of existing LMMs in visual mathematical reasoning and reveal a negative correlation between solving steps and problem-specific performance. We confirm the IK issue of LMMs can be effectively improved via knowledge augmentation strategies. More notably, the primary challenge of GPT-4o has significantly transitioned from IK to IG, establishing it as the first LMM advancing towards the knowledge generalization stage. In contrast, other LMMs exhibit a marked inclination towards Rote Memorization - they correctly solve composite problems involving multiple knowledge concepts yet fail to answer sub-problems. We anticipate that WE-MATH will open new pathways for advancements in visual mathematical reasoning for LMMs. The WE-MATH data and evaluation code are available at https://github.com/We-Math/We-Math.
△ Less
Submitted 1 July, 2024;
originally announced July 2024.
-
CS-Bench: A Comprehensive Benchmark for Large Language Models towards Computer Science Mastery
Authors:
Xiaoshuai Song,
Muxi Diao,
Guanting Dong,
Zhengyang Wang,
Yujia Fu,
Runqi Qiao,
Zhexu Wang,
Dayuan Fu,
Huangxuan Wu,
Bin Liang,
Weihao Zeng,
Yejie Wang,
Zhuoma GongQue,
Jianing Yu,
Qiuna Tan,
Weiran Xu
Abstract:
Large language models (LLMs) have demonstrated significant potential in advancing various fields of research and society. However, the current community of LLMs overly focuses on benchmarks for analyzing specific foundational skills (e.g. mathematics and code generation), neglecting an all-round evaluation of the computer science field. To bridge this gap, we introduce CS-Bench, the first multilin…
▽ More
Large language models (LLMs) have demonstrated significant potential in advancing various fields of research and society. However, the current community of LLMs overly focuses on benchmarks for analyzing specific foundational skills (e.g. mathematics and code generation), neglecting an all-round evaluation of the computer science field. To bridge this gap, we introduce CS-Bench, the first multilingual (English, Chinese, French, German) benchmark dedicated to evaluating the performance of LLMs in computer science. CS-Bench comprises approximately 10K meticulously curated test samples, covering 26 subfields across 4 key areas of computer science, encompassing various task forms and divisions of knowledge and reasoning. Utilizing CS-Bench, we conduct a comprehensive evaluation of over 30 mainstream LLMs, revealing the relationship between CS performance and model scales. We also quantitatively analyze the reasons for failures in existing LLMs and highlight directions for improvements, including knowledge supplementation and CS-specific reasoning. Further cross-capability experiments show a high correlation between LLMs' capabilities in computer science and their abilities in mathematics and coding. Moreover, expert LLMs specialized in mathematics and coding also demonstrate strong performances in several CS subfields. Looking ahead, we envision CS-Bench serving as a cornerstone for LLM applications in the CS field and paving new avenues in assessing LLMs' diverse reasoning capabilities. The CS-Bench data and evaluation code are available at https://github.com/csbench/csbench.
△ Less
Submitted 28 February, 2025; v1 submitted 12 June, 2024;
originally announced June 2024.
-
A Novel Stochastic Transformer-based Approach for Post-Traumatic Stress Disorder Detection using Audio Recording of Clinical Interviews
Authors:
Mamadou Dia,
Ghazaleh Khodabandelou,
Alice Othmani
Abstract:
Post-traumatic stress disorder (PTSD) is a mental disorder that can be developed after witnessing or experiencing extremely traumatic events. PTSD can affect anyone, regardless of ethnicity, or culture. An estimated one in every eleven people will experience PTSD during their lifetime. The Clinician-Administered PTSD Scale (CAPS) and the PTSD Check List for Civilians (PCL-C) interviews are gold st…
▽ More
Post-traumatic stress disorder (PTSD) is a mental disorder that can be developed after witnessing or experiencing extremely traumatic events. PTSD can affect anyone, regardless of ethnicity, or culture. An estimated one in every eleven people will experience PTSD during their lifetime. The Clinician-Administered PTSD Scale (CAPS) and the PTSD Check List for Civilians (PCL-C) interviews are gold standards in the diagnosis of PTSD. These questionnaires can be fooled by the subject's responses. This work proposes a deep learning-based approach that achieves state-of-the-art performances for PTSD detection using audio recordings during clinical interviews. Our approach is based on MFCC low-level features extracted from audio recordings of clinical interviews, followed by deep high-level learning using a Stochastic Transformer. Our proposed approach achieves state-of-the-art performances with an RMSE of 2.92 on the eDAIC dataset thanks to the stochastic depth, stochastic deep learning layers, and stochastic activation function.
△ Less
Submitted 28 March, 2024;
originally announced March 2024.
-
Greedy selection of optimal location of sensors for uncertainty reduction in seismic moment tensor inversion
Authors:
Ben Mansour Dia,
Michael Fehler,
SanLinn I. Kaka,
Andrea Scarinci,
Umair bin Waheed,
Chen Gu
Abstract:
We address an optimal sensor placement problem through Bayesian experimental design for seismic full waveform inversion for the recovery of the associated moment tensor. The objective is that of optimally choosing the location of the sensors (stations) from which to collect the observed data. The Shannon expected information gain is used as the objective function to search for the optimal network…
▽ More
We address an optimal sensor placement problem through Bayesian experimental design for seismic full waveform inversion for the recovery of the associated moment tensor. The objective is that of optimally choosing the location of the sensors (stations) from which to collect the observed data. The Shannon expected information gain is used as the objective function to search for the optimal network of sensors. A closed form for such objective is available due to the linear structure of the forward problem, as well as the Gaussian modeling of the observational errors and prior distribution. The resulting problem being inherently combinatorial, a greedy algorithm is deployed to sequentially select the sensor locations that form the best network for learning the moment tensor. Numerical results are presented and analyzed under several instances of the problem, including: use of full three-dimensional velocity-models, cases in which the earthquake-source location is unknown, as well as moment tensor inversion under model misspecification
△ Less
Submitted 3 March, 2024;
originally announced March 2024.
-
Multi-Perspective Consistency Enhances Confidence Estimation in Large Language Models
Authors:
Pei Wang,
Yejie Wang,
Muxi Diao,
Keqing He,
Guanting Dong,
Weiran Xu
Abstract:
In the deployment of large language models (LLMs), accurate confidence estimation is critical for assessing the credibility of model predictions. However, existing methods often fail to overcome the issue of overconfidence on incorrect answers. In this work, we focus on improving the confidence estimation of large language models. Considering the fragility of self-awareness in language models, we…
▽ More
In the deployment of large language models (LLMs), accurate confidence estimation is critical for assessing the credibility of model predictions. However, existing methods often fail to overcome the issue of overconfidence on incorrect answers. In this work, we focus on improving the confidence estimation of large language models. Considering the fragility of self-awareness in language models, we introduce a Multi-Perspective Consistency (MPC) method. We leverage complementary insights from different perspectives within models (MPC-Internal) and across different models (MPC-Across) to mitigate the issue of overconfidence arising from a singular viewpoint. The experimental results on eight publicly available datasets show that our MPC achieves state-of-the-art performance. Further analyses indicate that MPC can mitigate the problem of overconfidence and is effectively scalable to other models.
△ Less
Submitted 17 February, 2024;
originally announced February 2024.
-
DolphCoder: Echo-Locating Code Large Language Models with Diverse and Multi-Objective Instruction Tuning
Authors:
Yejie Wang,
Keqing He,
Guanting Dong,
Pei Wang,
Weihao Zeng,
Muxi Diao,
Yutao Mou,
Mengdi Zhang,
Jingang Wang,
Xunliang Cai,
Weiran Xu
Abstract:
Code Large Language Models (Code LLMs) have demonstrated outstanding performance in code-related tasks. Several instruction tuning approaches have been proposed to boost the code generation performance of pre-trained Code LLMs. In this paper, we introduce a diverse instruction model (DolphCoder) with self-evaluating for code generation. It learns diverse instruction targets and combines a code eva…
▽ More
Code Large Language Models (Code LLMs) have demonstrated outstanding performance in code-related tasks. Several instruction tuning approaches have been proposed to boost the code generation performance of pre-trained Code LLMs. In this paper, we introduce a diverse instruction model (DolphCoder) with self-evaluating for code generation. It learns diverse instruction targets and combines a code evaluation objective to enhance its code generation ability. Our model achieves superior performance on the HumanEval and MBPP benchmarks, demonstrating new insights for future code instruction tuning work. Our key findings are: (1) Augmenting more diverse responses with distinct reasoning paths increases the code capability of LLMs. (2) Improving one's ability to evaluate the correctness of code solutions also enhances their ability to create it.
△ Less
Submitted 14 February, 2024;
originally announced February 2024.
-
Forward-backward Gaussian variational inference via JKO in the Bures-Wasserstein Space
Authors:
Michael Diao,
Krishnakumar Balasubramanian,
Sinho Chewi,
Adil Salim
Abstract:
Variational inference (VI) seeks to approximate a target distribution $π$ by an element of a tractable family of distributions. Of key interest in statistics and machine learning is Gaussian VI, which approximates $π$ by minimizing the Kullback-Leibler (KL) divergence to $π$ over the space of Gaussians. In this work, we develop the (Stochastic) Forward-Backward Gaussian Variational Inference (FB-G…
▽ More
Variational inference (VI) seeks to approximate a target distribution $π$ by an element of a tractable family of distributions. Of key interest in statistics and machine learning is Gaussian VI, which approximates $π$ by minimizing the Kullback-Leibler (KL) divergence to $π$ over the space of Gaussians. In this work, we develop the (Stochastic) Forward-Backward Gaussian Variational Inference (FB-GVI) algorithm to solve Gaussian VI. Our approach exploits the composite structure of the KL divergence, which can be written as the sum of a smooth term (the potential) and a non-smooth term (the entropy) over the Bures-Wasserstein (BW) space of Gaussians endowed with the Wasserstein distance. For our proposed algorithm, we obtain state-of-the-art convergence guarantees when $π$ is log-smooth and log-concave, as well as the first convergence guarantees to first-order stationary solutions when $π$ is only log-smooth.
△ Less
Submitted 10 April, 2023;
originally announced April 2023.
-
A continuation method in Bayesian inference
Authors:
Ben Mansour Dia
Abstract:
We present a continuation method that entails generating a sequence of transition probability density functions from the prior to the posterior in the context of Bayesian inference for parameter estimation problems. The characterization of transition distributions, by tempering the likelihood function, results in a homogeneous nonlinear partial integro-differential equation whose existence and uni…
▽ More
We present a continuation method that entails generating a sequence of transition probability density functions from the prior to the posterior in the context of Bayesian inference for parameter estimation problems. The characterization of transition distributions, by tempering the likelihood function, results in a homogeneous nonlinear partial integro-differential equation whose existence and uniqueness of solutions are addressed. The posterior probability distribution comes as the interpretation of the final state of the path of transition distributions. A computationally stable scaling domain for the likelihood is explored for the approximation of the expected deviance, where we manage to hold back all the evaluations of the forward predictive model at the prior stage. It follows the computational tractability of the posterior distribution and opens access to the posterior distribution for direct samplings. To get a solution formulation of the expected deviance, we derive a partial differential equation governing the moments generating function of the log-likelihood. We show also that a spectral formulation of the expected deviance can be obtained for low-dimensional problems under certain conditions. The computational efficiency of the proposed method is demonstrated through three differents numerical examples that focus on analyzing the computational bias generated by the method, assessing the continuation method in the Bayesian inference with non-Gaussian noise, and evaluating its ability to invert a multimodal parameter of interest.
△ Less
Submitted 26 November, 2019;
originally announced November 2019.
-
Galaxy Image Simulation Using Progressive GANs
Authors:
Mohamad Dia,
Elodie Savary,
Martin Melchior,
Frederic Courbin
Abstract:
In this work, we provide an efficient and realistic data-driven approach to simulate astronomical images using deep generative models from machine learning. Our solution is based on a variant of the generative adversarial network (GAN) with progressive training methodology and Wasserstein cost function. The proposed solution generates naturalistic images of galaxies that show complex structures an…
▽ More
In this work, we provide an efficient and realistic data-driven approach to simulate astronomical images using deep generative models from machine learning. Our solution is based on a variant of the generative adversarial network (GAN) with progressive training methodology and Wasserstein cost function. The proposed solution generates naturalistic images of galaxies that show complex structures and high diversity, which suggests that data-driven simulations using machine learning can replace many of the expensive model-driven methods used in astronomical data processing.
△ Less
Submitted 26 September, 2019;
originally announced September 2019.
-
Rank-one matrix estimation: analysis of algorithmic and information theoretic limits by the spatial coupling method
Authors:
Jean Barbier,
Mohamad Dia,
Nicolas Macris,
Florent Krzakala,
Lenka Zdeborová
Abstract:
Factorizing low-rank matrices is a problem with many applications in machine learning and statistics, ranging from sparse PCA to community detection and sub-matrix localization. For probabilistic models in the Bayes optimal setting, general expressions for the mutual information have been proposed using powerful heuristic statistical physics computations via the replica and cavity methods, and pro…
▽ More
Factorizing low-rank matrices is a problem with many applications in machine learning and statistics, ranging from sparse PCA to community detection and sub-matrix localization. For probabilistic models in the Bayes optimal setting, general expressions for the mutual information have been proposed using powerful heuristic statistical physics computations via the replica and cavity methods, and proven in few specific cases by a variety of methods. Here, we use the spatial coupling methodology developed in the framework of error correcting codes, to rigorously derive the mutual information for the symmetric rank-one case. We characterize the detectability phase transitions in a large set of estimation problems, where we show that there exists a gap between what currently known polynomial algorithms (in particular spectral methods and approximate message-passing) can do and what is expected information theoretically. Moreover, we show that the computational gap vanishes for the proposed spatially coupled model, a promising feature with many possible applications. Our proof technique has an interest on its own and exploits three essential ingredients: the interpolation method first introduced in statistical physics, the analysis of approximate message-passing algorithms first introduced in compressive sensing, and the theory of threshold saturation for spatially coupled systems first developed in coding theory. Our approach is very generic and can be applied to many other open problems in statistical estimation where heuristic statistical physics predictions are available.
△ Less
Submitted 6 December, 2018;
originally announced December 2018.
-
Multilevel Double Loop Monte Carlo and Stochastic Collocation Methods with Importance Sampling for Bayesian Optimal Experimental Design
Authors:
Joakim Beck,
Ben Mansour Dia,
Luis F. R. Espath,
Raul Tempone
Abstract:
An optimal experimental set-up maximizes the value of data for statistical inferences and predictions. The efficiency of strategies for finding optimal experimental set-ups is particularly important for experiments that are time-consuming or expensive to perform. For instance, in the situation when the experiments are modeled by Partial Differential Equations (PDEs), multilevel methods have been p…
▽ More
An optimal experimental set-up maximizes the value of data for statistical inferences and predictions. The efficiency of strategies for finding optimal experimental set-ups is particularly important for experiments that are time-consuming or expensive to perform. For instance, in the situation when the experiments are modeled by Partial Differential Equations (PDEs), multilevel methods have been proven to dramatically reduce the computational complexity of their single-level counterparts when estimating expected values. For a setting where PDEs can model experiments, we propose two multilevel methods for estimating a popular design criterion known as the expected information gain in simulation-based Bayesian optimal experimental design. The expected information gain criterion is of a nested expectation form, and only a handful of multilevel methods have been proposed for problems of such form. We propose a Multilevel Double Loop Monte Carlo (MLDLMC), which is a multilevel strategy with Double Loop Monte Carlo (DLMC), and a Multilevel Double Loop Stochastic Collocation (MLDLSC), which performs a high-dimensional integration by deterministic quadrature on sparse grids. For both methods, the Laplace approximation is used for importance sampling that significantly reduces the computational work of estimating inner expectations. The optimal values of the method parameters are determined by minimizing the average computational work, subject to satisfying the desired error tolerance. The computational efficiencies of the methods are demonstrated by estimating the expected information gain for Bayesian inference of the fiber orientation in composite laminate materials from an electrical impedance tomography experiment. MLDLSC performs better than MLDLMC when the regularity of the quantity of interest, with respect to the additive noise and the unknown parameters, can be exploited.
△ Less
Submitted 3 February, 2020; v1 submitted 28 November, 2018;
originally announced November 2018.
-
Nesterov-aided Stochastic Gradient Methods using Laplace Approximation for Bayesian Design Optimization
Authors:
Andre Gustavo Carlon,
Ben Mansour Dia,
Luis FR Espath,
Rafael Holdorf Lopez,
Raul Tempone
Abstract:
Finding the best setup for experiments is the primary concern for Optimal Experimental Design (OED). Here, we focus on the Bayesian experimental design problem of finding the setup that maximizes the Shannon expected information gain. We use the stochastic gradient descent and its accelerated counterpart, which employs Nesterov's method, to solve the optimization problem in OED. We adapt a restart…
▽ More
Finding the best setup for experiments is the primary concern for Optimal Experimental Design (OED). Here, we focus on the Bayesian experimental design problem of finding the setup that maximizes the Shannon expected information gain. We use the stochastic gradient descent and its accelerated counterpart, which employs Nesterov's method, to solve the optimization problem in OED. We adapt a restart technique, originally proposed for the acceleration in deterministic optimization, to improve stochastic optimization methods. We combine these optimization methods with three estimators of the objective function: the double-loop Monte Carlo estimator (DLMC), the Monte Carlo estimator using the Laplace approximation for the posterior distribution (MCLA) and the double-loop Monte Carlo estimator with Laplace-based importance sampling (DLMCIS). Using stochastic gradient methods and Laplace-based estimators together allows us to use expensive and complex models, such as those that require solving partial differential equations (PDEs). From a theoretical viewpoint, we derive an explicit formula to compute the gradient estimator of the Monte Carlo methods, including MCLA and DLMCIS. From a computational standpoint, we study four examples: three based on analytical functions and one using the finite element method. The last example is an electrical impedance tomography experiment based on the complete electrode model. In these examples, the accelerated stochastic gradient descent method using MCLA converges to local maxima with up to five orders of magnitude fewer model evaluations than gradient descent with DLMC.
△ Less
Submitted 26 February, 2020; v1 submitted 2 July, 2018;
originally announced July 2018.
-
A Compressed Sensing Approach for Distribution Matching
Authors:
Mohamad Dia,
Vahid Aref,
Laurent Schmalen
Abstract:
In this work, we formulate the fixed-length distribution matching as a Bayesian inference problem. Our proposed solution is inspired from the compressed sensing paradigm and the sparse superposition (SS) codes. First, we introduce sparsity in the binary source via position modulation (PM). We then present a simple and exact matcher based on Gaussian signal quantization. At the receiver, the dematc…
▽ More
In this work, we formulate the fixed-length distribution matching as a Bayesian inference problem. Our proposed solution is inspired from the compressed sensing paradigm and the sparse superposition (SS) codes. First, we introduce sparsity in the binary source via position modulation (PM). We then present a simple and exact matcher based on Gaussian signal quantization. At the receiver, the dematcher exploits the sparsity in the source and performs low-complexity dematching based on generalized approximate message-passing (GAMP). We show that GAMP dematcher and spatial coupling lead to asymptotically optimal performance, in the sense that the rate tends to the entropy of the target distribution with vanishing reconstruction error in a proper limit. Furthermore, we assess the performance of the dematcher on practical Hadamard-based operators. A remarkable feature of our proposed solution is the possibility to: i) perform matching at the symbol level (nonbinary); ii) perform joint channel coding and matching.
△ Less
Submitted 25 November, 2018; v1 submitted 2 April, 2018;
originally announced April 2018.
-
Fast Bayesian experimental design: Laplace-based importance sampling for the expected information gain
Authors:
Joakim Beck,
Ben Mansour Dia,
Luis FR Espath,
Quan Long,
Raul Tempone
Abstract:
In calculating expected information gain in optimal Bayesian experimental design, the computation of the inner loop in the classical double-loop Monte Carlo requires a large number of samples and suffers from underflow if the number of samples is small. These drawbacks can be avoided by using an importance sampling approach. We present a computationally efficient method for optimal Bayesian experi…
▽ More
In calculating expected information gain in optimal Bayesian experimental design, the computation of the inner loop in the classical double-loop Monte Carlo requires a large number of samples and suffers from underflow if the number of samples is small. These drawbacks can be avoided by using an importance sampling approach. We present a computationally efficient method for optimal Bayesian experimental design that introduces importance sampling based on the Laplace method to the inner loop. We derive the optimal values for the method parameters in which the average computational cost is minimized according to the desired error tolerance. We use three numerical examples to demonstrate the computational efficiency of our method compared with the classical double-loop Monte Carlo, and a more recent single-loop Monte Carlo method that uses the Laplace method as an approximation of the return value of the inner loop. The first example is a scalar problem that is linear in the uncertain parameter. The second example is a nonlinear scalar problem. The third example deals with the optimal sensor placement for an electrical impedance tomography experiment to recover the fiber orientation in laminate composites.
△ Less
Submitted 10 October, 2017;
originally announced October 2017.
-
Universal Sparse Superposition Codes with Spatial Coupling and GAMP Decoding
Authors:
Jean Barbier,
Mohamad Dia,
Nicolas Macris
Abstract:
Sparse superposition codes, or sparse regression codes, constitute a new class of codes which was first introduced for communication over the additive white Gaussian noise (AWGN) channel. It has been shown that such codes are capacity-achieving over the AWGN channel under optimal maximum-likelihood decoding as well as under various efficient iterative decoding schemes equipped with power allocatio…
▽ More
Sparse superposition codes, or sparse regression codes, constitute a new class of codes which was first introduced for communication over the additive white Gaussian noise (AWGN) channel. It has been shown that such codes are capacity-achieving over the AWGN channel under optimal maximum-likelihood decoding as well as under various efficient iterative decoding schemes equipped with power allocation or spatially coupled constructions. Here, we generalize the analysis of these codes to a much broader setting that includes all memoryless channels. We show, for a large class of memoryless channels, that spatial coupling allows an efficient decoder, based on the generalized approximate message-passing (GAMP) algorithm, to reach the potential (or Bayes optimal) threshold of the underlying (or uncoupled) code ensemble. Moreover, we argue that spatially coupled sparse superposition codes universally achieve capacity under GAMP decoding by showing, through analytical computations, that the error floor vanishes and the potential threshold tends to capacity as one of the code parameter goes to infinity. Furthermore, we provide a closed form formula for the algorithmic threshold of the underlying code ensemble in terms of a Fisher information. Relating an algorithmic threshold to a Fisher information has theoretical as well as practical importance. Our proof relies on the state evolution analysis and uses the potential method developed in the theory of low-density parity-check (LDPC) codes and compressed sensing.
△ Less
Submitted 8 November, 2018; v1 submitted 13 July, 2017;
originally announced July 2017.
-
Mutual Information and Optimality of Approximate Message-Passing in Random Linear Estimation
Authors:
Jean Barbier,
Nicolas Macris,
Mohamad Dia,
Florent Krzakala
Abstract:
We consider the estimation of a signal from the knowledge of its noisy linear random Gaussian projections. A few examples where this problem is relevant are compressed sensing, sparse superposition codes, and code division multiple access. There has been a number of works considering the mutual information for this problem using the replica method from statistical physics. Here we put these consid…
▽ More
We consider the estimation of a signal from the knowledge of its noisy linear random Gaussian projections. A few examples where this problem is relevant are compressed sensing, sparse superposition codes, and code division multiple access. There has been a number of works considering the mutual information for this problem using the replica method from statistical physics. Here we put these considerations on a firm rigorous basis. First, we show, using a Guerra-Toninelli type interpolation, that the replica formula yields an upper bound to the exact mutual information. Secondly, for many relevant practical cases, we present a converse lower bound via a method that uses spatial coupling, state evolution analysis and the I-MMSE theorem. This yields a single letter formula for the mutual information and the minimal-mean-square error for random Gaussian linear estimation of all discrete bounded signals. In addition, we prove that the low complexity approximate message-passing algorithm is optimal outside of the so-called hard phase, in the sense that it asymptotically reaches the minimal-mean-square error. In this work spatial coupling is used primarily as a proof technique. However our results also prove two important features of spatially coupled noisy linear random Gaussian estimation. First there is no algorithmically hard phase. This means that for such systems approximate message-passing always reaches the minimal-mean-square error. Secondly, in a proper limit the mutual information associated to such systems is the same as the one of uncoupled linear random Gaussian estimation.
△ Less
Submitted 28 August, 2020; v1 submitted 20 January, 2017;
originally announced January 2017.
-
Generalized Approximate Message-Passing Decoder for Universal Sparse Superposition Codes
Authors:
Erdem Biyik,
Jean Barbier,
Mohamad Dia
Abstract:
Sparse superposition (SS) codes were originally proposed as a capacity-achieving communication scheme over the additive white Gaussian noise channel (AWGNC) [1]. Very recently, it was discovered that these codes are universal, in the sense that they achieve capacity over any memoryless channel under generalized approximate message-passing (GAMP) decoding [2], although this decoder has never been s…
▽ More
Sparse superposition (SS) codes were originally proposed as a capacity-achieving communication scheme over the additive white Gaussian noise channel (AWGNC) [1]. Very recently, it was discovered that these codes are universal, in the sense that they achieve capacity over any memoryless channel under generalized approximate message-passing (GAMP) decoding [2], although this decoder has never been stated for SS codes. In this contribution we introduce the GAMP decoder for SS codes, we confirm empirically the universality of this communication scheme through its study on various channels and we provide the main analysis tools: state evolution and potential. We also compare the performance of GAMP with the Bayes-optimal MMSE decoder. We empirically illustrate that despite the presence of a phase transition preventing GAMP to reach the optimal performance, spatial coupling allows to boost the performance that eventually tends to capacity in a proper limit. We also prove that, in contrast with the AWGNC case, SS codes for binary input channels have a vanishing error floor in the limit of large codewords. Moreover, the performance of Hadamard-based encoders is assessed for practical implementations.
△ Less
Submitted 13 January, 2017;
originally announced January 2017.
-
The Mutual Information in Random Linear Estimation
Authors:
Jean Barbier,
Mohamad Dia,
Nicolas Macris,
Florent Krzakala
Abstract:
We consider the estimation of a signal from the knowledge of its noisy linear random Gaussian projections, a problem relevant in compressed sensing, sparse superposition codes or code division multiple access just to cite few. There has been a number of works considering the mutual information for this problem using the heuristic replica method from statistical physics. Here we put these considera…
▽ More
We consider the estimation of a signal from the knowledge of its noisy linear random Gaussian projections, a problem relevant in compressed sensing, sparse superposition codes or code division multiple access just to cite few. There has been a number of works considering the mutual information for this problem using the heuristic replica method from statistical physics. Here we put these considerations on a firm rigorous basis. First, we show, using a Guerra-type interpolation, that the replica formula yields an upper bound to the exact mutual information. Secondly, for many relevant practical cases, we present a converse lower bound via a method that uses spatial coupling, state evolution analysis and the I-MMSE theorem. This yields, in particular, a single letter formula for the mutual information and the minimal-mean-square error for random Gaussian linear estimation of all discrete bounded signals.
△ Less
Submitted 6 September, 2016; v1 submitted 8 July, 2016;
originally announced July 2016.
-
Mutual information for symmetric rank-one matrix estimation: A proof of the replica formula
Authors:
Jean Barbier,
Mohamad Dia,
Nicolas Macris,
Florent Krzakala,
Thibault Lesieur,
Lenka Zdeborova
Abstract:
Factorizing low-rank matrices has many applications in machine learning and statistics. For probabilistic models in the Bayes optimal setting, a general expression for the mutual information has been proposed using heuristic statistical physics computations, and proven in few specific cases. Here, we show how to rigorously prove the conjectured formula for the symmetric rank-one case. This allows…
▽ More
Factorizing low-rank matrices has many applications in machine learning and statistics. For probabilistic models in the Bayes optimal setting, a general expression for the mutual information has been proposed using heuristic statistical physics computations, and proven in few specific cases. Here, we show how to rigorously prove the conjectured formula for the symmetric rank-one case. This allows to express the minimal mean-square-error and to characterize the detectability phase transitions in a large set of estimation problems ranging from community detection to sparse PCA. We also show that for a large set of parameters, an iterative algorithm called approximate message-passing is Bayes optimal. There exists, however, a gap between what currently known polynomial algorithms can do and what is expected information theoretically. Additionally, the proof technique has an interest of its own and exploits three essential ingredients: the interpolation method introduced in statistical physics by Guerra, the analysis of the approximate message-passing algorithm and the theory of spatial coupling and threshold saturation in coding. Our approach is generic and applicable to other open problems in statistical estimation where heuristic statistical physics predictions are available.
△ Less
Submitted 13 June, 2016;
originally announced June 2016.
-
Threshold Saturation of Spatially Coupled Sparse Superposition Codes for All Memoryless Channels
Authors:
Jean Barbier,
Mohamad Dia,
Nicolas Macris
Abstract:
We recently proved threshold saturation for spatially coupled sparse superposition codes on the additive white Gaussian noise channel. Here we generalize our analysis to a much broader setting. We show for any memoryless channel that spatial coupling allows generalized approximate message-passing (GAMP) decoding to reach the potential (or Bayes optimal) threshold of the code ensemble. Moreover in…
▽ More
We recently proved threshold saturation for spatially coupled sparse superposition codes on the additive white Gaussian noise channel. Here we generalize our analysis to a much broader setting. We show for any memoryless channel that spatial coupling allows generalized approximate message-passing (GAMP) decoding to reach the potential (or Bayes optimal) threshold of the code ensemble. Moreover in the large input alphabet size limit: i) the GAMP algorithmic threshold of the underlying (or uncoupled) code ensemble is simply expressed as a Fisher information; ii) the potential threshold tends to Shannon's capacity. Although we focus on coding for sake of coherence with our previous results, the framework and methods are very general and hold for a wide class of generalized estimation problems with random linear mixing.
△ Less
Submitted 15 March, 2016;
originally announced March 2016.
-
Proof of Threshold Saturation for Spatially Coupled Sparse Superposition Codes
Authors:
Jean Barbier,
Mohamad Dia,
Nicolas Macris
Abstract:
Recently, a new class of codes, called sparse superposition or sparse regression codes, has been proposed for communication over the AWGN channel. It has been proven that they achieve capacity using power allocation and various forms of iterative decoding. Empirical evidence has also strongly suggested that the codes achieve capacity when spatial coupling and approximate message passing decoding a…
▽ More
Recently, a new class of codes, called sparse superposition or sparse regression codes, has been proposed for communication over the AWGN channel. It has been proven that they achieve capacity using power allocation and various forms of iterative decoding. Empirical evidence has also strongly suggested that the codes achieve capacity when spatial coupling and approximate message passing decoding are used, without need of power allocation. In this note we prove that state evolution (which tracks message passing) indeed saturates the potential threshold of the underlying code ensemble, which approaches in a proper limit the optimal threshold. Our proof uses ideas developed in the theory of low-density parity-check codes and compressive sensing.
△ Less
Submitted 6 March, 2016;
originally announced March 2016.
-
Mean-Field Games for Marriage
Authors:
Dario Bauso,
Ben Mansour Dia,
Boualem Djehiche,
Hamidou Tembine,
Raul Tempone
Abstract:
This article examines mean-field games for marriage. The results support the argument that optimizing the long-term well-being through effort and social feeling state distribution (mean-field) will help to stabilize marriage. However, if the cost of effort is very high, the couple fluctuates in a bad feeling state or the marriage breaks down. We then examine the influence of society on a couple us…
▽ More
This article examines mean-field games for marriage. The results support the argument that optimizing the long-term well-being through effort and social feeling state distribution (mean-field) will help to stabilize marriage. However, if the cost of effort is very high, the couple fluctuates in a bad feeling state or the marriage breaks down. We then examine the influence of society on a couple using mean field sentimental games. We show that, in mean-field equilibrium, the optimal effort is always higher than the one-shot optimal effort. We illustrate numerically the influence of the couple's network on their feeling states and their well-being.
△ Less
Submitted 13 April, 2014;
originally announced April 2014.
