id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2501.18873
Best Policy Learning from Trajectory Preference Feedback
cs.LG
We address the problem of best policy identification in preference-based reinforcement learning (PbRL), where learning occurs from noisy binary preferences over trajectory pairs rather than explicit numerical rewards. This approach is useful for post-training optimization of generative AI models during multi-turn user interactions, where preference feedback is more robust than handcrafted reward models. In this setting, learning is driven by both an offline preference dataset -- collected from a rater of unknown 'competence' -- and online data collected with pure exploration. Since offline datasets may exhibit out-of-distribution (OOD) biases, principled online data collection is necessary. To address this, we propose Posterior Sampling for Preference Learning ($\mathsf{PSPL}$), a novel algorithm inspired by Top-Two Thompson Sampling, that maintains independent posteriors over the true reward model and transition dynamics. We provide the first theoretical guarantees for PbRL in this setting, establishing an upper bound on the simple Bayesian regret of $\mathsf{PSPL}$. Since the exact algorithm can be computationally impractical, we also provide an approximate version that outperforms existing baselines.
2501.18875
Self-Supervised Learning Using Nonlinear Dependence
cs.LG cs.CV stat.ML
Self-supervised learning has gained significant attention in contemporary applications, particularly due to the scarcity of labeled data. While existing SSL methodologies primarily address feature variance and linear correlations, they often neglect the intricate relations between samples and the nonlinear dependencies inherent in complex data. In this paper, we introduce Correlation-Dependence Self-Supervised Learning (CDSSL), a novel framework that unifies and extends existing SSL paradigms by integrating both linear correlations and nonlinear dependencies, encapsulating sample-wise and feature-wise interactions. Our approach incorporates the Hilbert-Schmidt Independence Criterion (HSIC) to robustly capture nonlinear dependencies within a Reproducing Kernel Hilbert Space, enriching representation learning. Experimental evaluations on diverse benchmarks demonstrate the efficacy of CDSSL in improving representation quality.
2501.18876
QMe14S, A Comprehensive and Efficient Spectral Dataset for Small Organic Molecules
physics.chem-ph cs.LG
Developing machine learning protocols for molecular simulations requires comprehensive and efficient datasets. Here we introduce the QMe14S dataset, comprising 186,102 small organic molecules featuring 14 elements (H, B, C, N, O, F, Al, Si, P, S, Cl, As, Se, Br) and 47 functional groups. Using density functional theory at the B3LYP/TZVP level, we optimized the geometries and calculated properties including energy, atomic charge, atomic force, dipole moment, quadrupole moment, polarizability, octupole moment, first hyperpolarizability, and Hessian. At the same level, we obtained the harmonic IR, Raman and NMR spectra. Furthermore, we conducted ab initio molecular dynamics simulations to generate dynamic configurations and extract nonequilibrium properties, including energy, forces, and Hessians. By leveraging our E(3)-equivariant message-passing neural network (DetaNet), we demonstrated that models trained on QMe14S outperform those trained on the previously developed QM9S dataset in simulating molecular spectra. The QMe14S dataset thus serves as a comprehensive benchmark for molecular simulations, offering valuable insights into structure-property relationships.
2501.18877
Distorting Embedding Space for Safety: A Defense Mechanism for Adversarially Robust Diffusion Models
cs.CV cs.CR cs.LG
Text-to-image diffusion models show remarkable generation performance following text prompts, but risk generating Not Safe For Work (NSFW) contents from unsafe prompts. Existing approaches, such as prompt filtering or concept unlearning, fail to defend against adversarial attacks while maintaining benign image quality. In this paper, we propose a novel approach called Distorting Embedding Space (DES), a text encoder-based defense mechanism that effectively tackles these issues through innovative embedding space control. DES transforms unsafe embeddings, extracted from a text encoder using unsafe prompts, toward carefully calculated safe embedding regions to prevent unsafe contents generation, while reproducing the original safe embeddings. DES also neutralizes the nudity embedding, extracted using prompt ``nudity", by aligning it with neutral embedding to enhance robustness against adversarial attacks. These methods ensure both robust defense and high-quality image generation. Additionally, DES can be adopted in a plug-and-play manner and requires zero inference overhead, facilitating its deployment. Extensive experiments on diverse attack types, including black-box and white-box scenarios, demonstrate DES's state-of-the-art performance in both defense capability and benign image generation quality. Our model is available at https://github.com/aei13/DES.
2501.18879
Understanding Generalization in Physics Informed Models through Affine Variety Dimensions
cs.LG math.ST stat.ML stat.TH
In recent years, physics-informed machine learning has gained significant attention for its ability to enhance statistical performance and sample efficiency by integrating physical structures into machine learning models. These structures, such as differential equations, conservation laws, and symmetries, serve as inductive biases that can improve the generalization capacity of the hybrid model. However, the mechanisms by which these physical structures enhance generalization capacity are not fully understood, limiting the ability to guarantee the performance of the models. In this study, we show that the generalization performance of linear regressors incorporating differential equation structures is determined by the dimension of the associated affine variety, rather than the number of parameters. This finding enables a unified analysis of various equations, including nonlinear ones. We introduce a method to approximate the dimension of the affine variety and provide experimental evidence to validate our theoretical insights.
2501.18880
RLS3: RL-Based Synthetic Sample Selection to Enhance Spatial Reasoning in Vision-Language Models for Indoor Autonomous Perception
cs.CV cs.LG
Vision-language model (VLM) fine-tuning for application-specific visual grounding based on natural language instructions has become one of the most popular approaches for learning-enabled autonomous systems. However, such fine-tuning relies heavily on high-quality datasets to achieve successful performance in various downstream tasks. Additionally, VLMs often encounter limitations due to insufficient and imbalanced fine-tuning data. To address these issues, we propose a new generalizable framework to improve VLM fine-tuning by integrating it with a reinforcement learning (RL) agent. Our method utilizes the RL agent to manipulate objects within an indoor setting to create synthetic data for fine-tuning to address certain vulnerabilities of the VLM. Specifically, we use the performance of the VLM to provide feedback to the RL agent to generate informative data that efficiently fine-tune the VLM over the targeted task (e.g. spatial reasoning). The key contribution of this work is developing a framework where the RL agent serves as an informative data sampling tool and assists the VLM in order to enhance performance and address task-specific vulnerabilities. By targeting the data sampling process to address the weaknesses of the VLM, we can effectively train a more context-aware model. In addition, generating synthetic data allows us to have precise control over each scene and generate granular ground truth captions. Our results show that the proposed data generation approach improves the spatial reasoning performance of VLMs, which demonstrates the benefits of using RL-guided data generation in vision-language tasks.
2501.18883
Can We Predict the Effect of Prompts?
cs.SE cs.LG
Large Language Models (LLMs) are machine learning models that have seen widespread adoption due to their capability of handling previously difficult tasks. LLMs, due to their training, are sensitive to how exactly a question is presented, also known as prompting. However, prompting well is challenging, as it has been difficult to uncover principles behind prompting -- generally, trial-and-error is the most common way of improving prompts, despite its significant computational cost. In this context, we argue it would be useful to perform `predictive prompt analysis', in which an automated technique would perform a quick analysis of a prompt and predict how the LLM would react to it, relative to a goal provided by the user. As a demonstration of the concept, we present Syntactic Prevalence Analyzer (SPA), a predictive prompt analysis approach based on sparse autoencoders (SAEs). SPA accurately predicted how often an LLM would generate target syntactic structures during code synthesis, with up to 0.994 Pearson correlation between the predicted and actual prevalence of the target structure. At the same time, SPA requires only 0.4\% of the time it takes to run the LLM on a benchmark. As LLMs are increasingly used during and integrated into modern software development, our proposed predictive prompt analysis concept has the potential to significantly ease the use of LLMs for both practitioners and researchers.
2501.18887
Building Bridges, Not Walls -- Advancing Interpretability by Unifying Feature, Data, and Model Component Attribution
cs.LG cs.AI
The increasing complexity of AI systems has made understanding their behavior a critical challenge. Numerous methods have been developed to attribute model behavior to three key aspects: input features, training data, and internal model components. However, these attribution methods are studied and applied rather independently, resulting in a fragmented landscape of approaches and terminology. This position paper argues that feature, data, and component attribution methods share fundamental similarities, and bridging them can benefit interpretability research. We conduct a detailed analysis of successful methods of these three attribution aspects and present a unified view to demonstrate that these seemingly distinct methods employ similar approaches, such as perturbations, gradients, and linear approximations, differing primarily in their perspectives rather than core techniques. Our unified perspective enhances understanding of existing attribution methods, identifies shared concepts and challenges, makes this field more accessible to newcomers, and highlights new directions not only for attribution and interpretability but also for broader AI research, including model editing, steering, and regulation.
2501.18889
Fully Distributed and Quantized Algorithm for MPC-based Autonomous Vehicle Platooning Optimization
eess.SY cs.MA cs.SY eess.SP math.OC
Intelligent transportation systems have recently emerged to address the growing interest for safer, more efficient, and sustainable transportation solutions. In this direction, this paper presents distributed algorithms for control and optimization over vehicular networks. First, we formulate the autonomous vehicle platooning framework based on model-predictive-control (MPC) strategies and present its objective optimization as a cooperative quadratic cost function. Then, we propose a distributed algorithm to locally optimize this objective at every vehicle subject to data quantization over the communication network of vehicles. In contrast to most existing literature that assumes ideal communication channels, log-scale data quantization over the network is addressed in this work, which is more realistic and practical. In particular, we show by simulation that the proposed log-quantized algorithm reaches optimal convergence with less residual and optimality gap. This outperforms the existing literature considering uniform quantization which leads to a large optimality gap and residual.
2501.18890
Distributed Observer Design for Tracking Platoon of Connected and Autonomous Vehicles
eess.SY cs.MA cs.SY eess.SP math.OC
Intelligent transportation systems (ITS) aim to advance innovative strategies relating to different modes of transport, traffic management, and autonomous vehicles. This paper studies the platoon of connected and autonomous vehicles (CAV) and proposes a distributed observer to track the state of the CAV dynamics. First, we model the CAV dynamics via an LTI interconnected system. Then, a consensus-based strategy is proposed to infer the state of the CAV dynamics based on local information exchange over the communication network of vehicles. A linear-matrix-inequality (LMI) technique is adopted for the block-diagonal observer gain design such that this gain is associated in a distributed way and locally to every vehicle. The distributed observer error dynamics is then shown to follow the structure of the Kronecker matrix product of the system dynamics and the adjacency matrix of the CAV network. The notions of survivable network design and redundant observer scheme are further discussed in the paper to address resilience to link and node failure. Finally, we verify our theoretical contributions via numerical simulations.
2501.18891
CAAT-EHR: Cross-Attentional Autoregressive Transformer for Multimodal Electronic Health Record Embeddings
cs.LG
Electronic health records (EHRs) provide a comprehensive source of longitudinal patient data, encompassing structured modalities such as laboratory results, imaging data, and vital signs, and unstructured clinical notes. These datasets, after necessary preprocessing to clean and format the data for analysis, often remain in their raw EHR form, representing numerical or categorical values without further transformation into task-agnostic embeddings. While such raw EHR data enables predictive modeling, its reliance on manual feature engineering or downstream task-specific optimization limits its utility for general-purpose applications. Deep learning (DL) techniques, such as recurrent neural networks (RNNs) and Transformers, have facilitated predictive tasks like disease progression and diagnosis prediction. However, these methods often struggle to fully exploit the temporal and multimodal dependencies inherent in EHR data due to their reliance on pre-processed but untransformed raw EHR inputs. In this study, we introduce CAAT-EHR, a novel architecture designed to bridge this gap by generating robust, task-agnostic longitudinal embeddings from raw EHR data. CAAT-EHR leverages self- and cross-attention mechanisms in its encoder to integrate temporal and contextual relationships across multiple modalities, transforming the data into enriched embeddings that capture complex dependencies. An autoregressive decoder complements the encoder by predicting future time points data during pre-training, ensuring that the resulting embeddings maintain temporal consistency and alignment. CAAT-EHR eliminates the need for manual feature engineering and enables seamless transferability across diverse downstream tasks. Extensive evaluations on benchmark datasets, demonstrate the superiority of CAAT-EHR-generated embeddings over pre-processed raw EHR data and other baseline approaches.
2501.18893
A machine learning approach for Premature Coronary Artery Disease Diagnosis according to Different Ethnicities in Iran
cs.LG
Premature coronary artery disease (PCAD) refers to the early onset of the disease, usually before the age of 55 for men and 65 for women. Coronary Artery Disease (CAD) develops when coronary arteries, the major blood vessels supplying the heart with blood, oxygen, and nutrients, become clogged or diseased. This is often due to many risk factors, including lifestyle and cardiometabolic ones, but few studies were done on ethnicity as one of these risk factors, especially in PCAD. In this study, we tested the rank of ethnicity among the major risk factors of PCAD, including age, gender, body mass index (BMI), visceral obesity presented as waist circumference (WC), diabetes mellitus (DM), high blood pressure (HBP), high low-density lipoprotein cholesterol (LDL-C), and smoking in a large national sample of patients with PCAD from different ethnicities. All patients who met the age criteria underwent coronary angiography to confirm CAD diagnosis. The weight of ethnicity was compared to the other eight features using feature weighting algorithms in PCAD diagnosis. In addition, we conducted an experiment where we ran predictive models (classification algorithms) to predict PCAD. We compared the performance of these models under two conditions: we trained the classification algorithms, including or excluding ethnicity. This study analyzed various factors to determine their predictive power influencing PCAD prediction. Among these factors, gender and age were the most significant predictors, with ethnicity being the third most important. The results also showed that if ethnicity is used as one of the input risk factors for classification algorithms, it can improve their efficiency. Our results show that ethnicity ranks as an influential factor in predicting PCAD. Therefore, it needs to be addressed in the PCAD diagnostic and preventive measures.
2501.18895
Efficient Supernet Training with Orthogonal Softmax for Scalable ASR Model Compression
cs.CL
ASR systems are deployed across diverse environments, each with specific hardware constraints. We use supernet training to jointly train multiple encoders of varying sizes, enabling dynamic model size adjustment to fit hardware constraints without redundant training. Moreover, we introduce a novel method called OrthoSoftmax, which applies multiple orthogonal softmax functions to efficiently identify optimal subnets within the supernet, avoiding resource-intensive search. This approach also enables more flexible and precise subnet selection by allowing selection based on various criteria and levels of granularity. Our results with CTC on Librispeech and TED-LIUM-v2 show that FLOPs-aware component-wise selection achieves the best overall performance. With the same number of training updates from one single job, WERs for all model sizes are comparable to or slightly better than those of individually trained models. Furthermore, we analyze patterns in the selected components and reveal interesting insights.
2501.18897
Trustworthy Evaluation of Generative AI Models
stat.ML cs.LG
Generative AI (GenAI) models have recently achieved remarkable empirical performance in various applications, however, their evaluations yet lack uncertainty quantification. In this paper, we propose a method to compare two generative models based on an unbiased estimator of their relative performance gap. Statistically, our estimator achieves parametric convergence rate and asymptotic normality, which enables valid inference. Computationally, our method is efficient and can be accelerated by parallel computing and leveraging pre-storing intermediate results. On simulated datasets with known ground truth, we show our approach effectively controls type I error and achieves power comparable with commonly used metrics. Furthermore, we demonstrate the performance of our method in evaluating diffusion models on real image datasets with statistical confidence.
2501.18898
GestureLSM: Latent Shortcut based Co-Speech Gesture Generation with Spatial-Temporal Modeling
cs.CV cs.GR
Controlling human gestures based on speech signals presents a significant challenge in computer vision. While existing works did preliminary studies of generating holistic co-speech gesture from speech, the spatial interaction of each body region during the speech remains barely explored. This leads to wield body part interactions given the speech signal. Furthermore, the slow generation speed limits the construction of real-world digital avatars. To resolve these problems, we propose \textbf{GestureLSM}, a Latent Shortcut based approach for Co-Speech Gesture Generation with spatial-temporal modeling. We tokenize various body regions and explicitly model their interactions with spatial and temporal attention. To achieve real-time gesture generations, we exam the denoising patterns and design an effective time distribution to speed up sampling while improve the generation quality for shortcut model. Extensive quantitative and qualitative experiments demonstrate the effectiveness of GestureLSM, showcasing its potential for various applications in the development of digital humans and embodied agents. Project Page: https://andypinxinliu.github.io/GestureLSM
2501.18899
Minimum Time Strategies for a Differential Drive Robot Escaping from a Circular Detection Region
cs.RO math.OC
A Differential Drive Robot (DDR) located inside a circular detection region in the plane wants to escape from it in minimum time. Various robotics applications can be modeled like the previous problem, such as a DDR escaping as soon as possible from a forbidden/dangerous region in the plane or running out from the sensor footprint of an unmanned vehicle flying at a constant altitude. In this paper, we find the motion strategies to accomplish its goal under two scenarios. In one, the detection region moves slower than the DDR and seeks to prevent escape; in another, its position is fixed. We formulate the problem as a zero-sum pursuit-evasion game, and using differential games theory, we compute the players' time-optimal motion strategies. Given the DDR's speed advantage, it can always escape by translating away from the center of the detection region at maximum speed. In this work, we show that the previous strategy could be optimal in some cases; however, other motion strategies emerge based on the player's speed ratio and the players' initial configurations.
2501.18901
Lightspeed Geometric Dataset Distance via Sliced Optimal Transport
cs.LG cs.AI stat.CO stat.ME stat.ML
We introduce sliced optimal transport dataset distance (s-OTDD), a model-agnostic, embedding-agnostic approach for dataset comparison that requires no training, is robust to variations in the number of classes, and can handle disjoint label sets. The core innovation is Moment Transform Projection (MTP), which maps a label, represented as a distribution over features, to a real number. Using MTP, we derive a data point projection that transforms datasets into one-dimensional distributions. The s-OTDD is defined as the expected Wasserstein distance between the projected distributions, with respect to random projection parameters. Leveraging the closed form solution of one-dimensional optimal transport, s-OTDD achieves (near-)linear computational complexity in the number of data points and feature dimensions and is independent of the number of classes. With its geometrically meaningful projection, s-OTDD strongly correlates with the optimal transport dataset distance while being more efficient than existing dataset discrepancy measures. Moreover, it correlates well with the performance gap in transfer learning and classification accuracy in data augmentation.
2501.18911
Integrated Communication and Binary State Detection Under Unequal Error Constraints
cs.IT eess.SP math.IT
This work considers a problem of integrated sensing and communication (ISAC) in which the goal of sensing is to detect a binary state. Unlike most approaches that minimize the total detection error probability, in our work, we disaggregate the error probability into false alarm and missed detection probabilities and investigate their information-theoretic three-way tradeoff including communication data rate. We consider a broadcast channel that consists of a transmitter, a communication receiver, and a detector where the receiver's and the detector's channels are affected by an unknown binary state. We consider and present results on two different state-dependent models. In the first setting, the state is fixed throughout the entire transmission, for which we fully characterize the optimal three-way tradeoff between the coding rate for communication and the two possibly nonidentical error exponents for sensing in the asymptotic regime. The achievability and converse proofs rely on the analysis of the cumulant-generating function of the log-likelihood ratio. In the second setting, the state changes every symbol in an independently and identically distributed (i.i.d.) manner, for which we characterize the optimal tradeoff region based on the analysis of the receiver operating characteristic (ROC) curves.
2501.18912
Analyzing Classroom Interaction Data Using Prompt Engineering and Network Analysis
stat.AP cs.SI
Classroom interactions play a vital role in developing critical thinking, collaborative problem-solving abilities, and enhanced learning outcomes. While analyzing these interactions is crucial for improving educational practices, the examination of classroom dialogues presents significant challenges due to the complexity and high-dimensionality of conversational data. This study presents an integrated framework that combines prompt engineering with network analysis to investigate classroom interactions comprehensively. Our approach automates utterance classification through prompt engineering, enabling efficient and scalable dialogue analysis without requiring pre-labeled datasets. The classified interactions are subsequently transformed into network representations, facilitating the analysis of classroom dynamics as structured social networks. To uncover complex interaction patterns and how underlying interaction structures relate to student learning, we utilize network mediation analysis. In this approach, latent interaction structures, derived from the additive and multiplicative effects network (AMEN) model that places students within a latent social space, act as mediators. In particular, we investigate how the gender gap in mathematics performance may be mediated by students' classroom interaction structures.
2501.18913
Rethinking Diffusion Posterior Sampling: From Conditional Score Estimator to Maximizing a Posterior
cs.CV
Recent advancements in diffusion models have been leveraged to address inverse problems without additional training, and Diffusion Posterior Sampling (DPS) (Chung et al., 2022a) is among the most popular approaches. Previous analyses suggest that DPS accomplishes posterior sampling by approximating the conditional score. While in this paper, we demonstrate that the conditional score approximation employed by DPS is not as effective as previously assumed, but rather aligns more closely with the principle of maximizing a posterior (MAP). This assertion is substantiated through an examination of DPS on 512x512 ImageNet images, revealing that: 1) DPS's conditional score estimation significantly diverges from the score of a well-trained conditional diffusion model and is even inferior to the unconditional score; 2) The mean of DPS's conditional score estimation deviates significantly from zero, rendering it an invalid score estimation; 3) DPS generates high-quality samples with significantly lower diversity. In light of the above findings, we posit that DPS more closely resembles MAP than a conditional score estimator, and accordingly propose the following enhancements to DPS: 1) we explicitly maximize the posterior through multi-step gradient ascent and projection; 2) we utilize a light-weighted conditional score estimator trained with only 100 images and 8 GPU hours. Extensive experimental results indicate that these proposed improvements significantly enhance DPS's performance. The source code for these improvements is provided in https://github.com/tongdaxu/Rethinking-Diffusion-Posterior-Sampling-From-Conditional-Score-Estimator-to-Maximizing-a-Posterior.
2501.18914
Scaling Laws for Differentially Private Language Models
cs.LG cs.CR
Scaling laws have emerged as important components of large language model (LLM) training as they can predict performance gains through scale, and provide guidance on important hyper-parameter choices that would otherwise be expensive. LLMs also rely on large, high-quality training datasets, like those sourced from (sometimes sensitive) user data. Training models on this sensitive user data requires careful privacy protections like differential privacy (DP). However, the dynamics of DP training are significantly different, and consequently their scaling laws are not yet fully understood. In this work, we establish scaling laws that accurately model the intricacies of DP LLM training, providing a complete picture of the compute-privacy-utility tradeoffs and the optimal training configurations in many settings.
2501.18915
An Invitation to Neuroalgebraic Geometry
cs.LG math.AG
In this expository work, we promote the study of function spaces parameterized by machine learning models through the lens of algebraic geometry. To this end, we focus on algebraic models, such as neural networks with polynomial activations, whose associated function spaces are semi-algebraic varieties. We outline a dictionary between algebro-geometric invariants of these varieties, such as dimension, degree, and singularities, and fundamental aspects of machine learning, such as sample complexity, expressivity, training dynamics, and implicit bias. Along the way, we review the literature and discuss ideas beyond the algebraic domain. This work lays the foundations of a research direction bridging algebraic geometry and deep learning, that we refer to as neuroalgebraic geometry.
2501.18916
LLM Program Optimization via Retrieval Augmented Search
cs.LG
With the advent of large language models (LLMs), there has been a great deal of interest in applying them to solve difficult programming tasks. Recent work has demonstrated their potential at program optimization, a key challenge in programming languages research. We propose a blackbox adaptation method called Retrieval Augmented Search (RAS) that performs beam search over candidate optimizations; at each step, it retrieves in-context examples from a given training dataset of slow-fast program pairs to guide the LLM. Critically, we find that performing contextual retrieval based on an LLM-generated natural language description significantly outperforms retrieval based on the source code. In addition, we propose a method called AEGIS for improving interpretability by decomposing training examples into "atomic edits" that are significantly more incremental in nature. We show that RAS performs 1.8$\times$ better than prior state-of-the-art blackbox adaptation strategies, and that AEGIS performs 1.37$\times$ better while performing significantly smaller edits.
2501.18919
Deepfake Detection of Singing Voices With Whisper Encodings
cs.SD cs.AI eess.AS
The deepfake generation of singing vocals is a concerning issue for artists in the music industry. In this work, we propose a singing voice deepfake detection (SVDD) system, which uses noise-variant encodings of open-AI's Whisper model. As counter-intuitive as it may sound, even though the Whisper model is known to be noise-robust, the encodings are rich in non-speech information, and are noise-variant. This leads us to evaluate Whisper encodings as feature representations for the SVDD task. Therefore, in this work, the SVDD task is performed on vocals and mixtures, and the performance is evaluated in \%EER over varying Whisper model sizes and two classifiers- CNN and ResNet34, under different testing conditions.
2501.18921
Full-scale Representation Guided Network for Retinal Vessel Segmentation
eess.IV cs.CV
The U-Net architecture and its variants have remained state-of-the-art (SOTA) for retinal vessel segmentation over the past decade. In this study, we introduce a Full Scale Guided Network (FSG-Net), where the feature representation network with modernized convolution blocks extracts full-scale information and the guided convolution block refines that information. Attention-guided filter is introduced to the guided convolution block under the interpretation that the filter behaves like the unsharp mask filter. Passing full-scale information to the attention block allows for the generation of improved attention maps, which are then passed to the attention-guided filter, resulting in performance enhancement of the segmentation network. The structure preceding the guided convolution block can be replaced by any U-Net variant, which enhances the scalability of the proposed approach. For a fair comparison, we re-implemented recent studies available in public repositories to evaluate their scalability and reproducibility. Our experiments also show that the proposed network demonstrates competitive results compared to current SOTA models on various public datasets. Ablation studies demonstrate that the proposed model is competitive with much smaller parameter sizes. Lastly, by applying the proposed model to facial wrinkle segmentation, we confirmed the potential for scalability to similar tasks in other domains. Our code is available on https://github.com/ZombaSY/FSG-Net-pytorch.
2501.18922
KBQA-o1: Agentic Knowledge Base Question Answering with Monte Carlo Tree Search
cs.CL cs.AI cs.DB
Knowledge Base Question Answering (KBQA) aims to answer natural language questions with a large-scale structured knowledge base (KB). Despite advancements with large language models (LLMs), KBQA still faces challenges in weak KB awareness, imbalance between effectiveness and efficiency, and high reliance on annotated data. To address these challenges, we propose KBQA-o1, a novel agentic KBQA method with Monte Carlo Tree Search (MCTS). It introduces a ReAct-based agent process for stepwise logical form generation with KB environment exploration. Moreover, it employs MCTS, a heuristic search method driven by policy and reward models, to balance agentic exploration's performance and search space. With heuristic exploration, KBQA-o1 generates high-quality annotations for further improvement by incremental fine-tuning. Experimental results show that KBQA-o1 outperforms previous low-resource KBQA methods with limited annotated data, boosting Llama-3.1-8B model's GrailQA F1 performance to 78.5% compared to 48.5% of the previous sota method with GPT-3.5-turbo.
2501.18924
Language Games as the Pathway to Artificial Superhuman Intelligence
cs.AI cs.CL cs.MA
The evolution of large language models (LLMs) toward artificial superhuman intelligence (ASI) hinges on data reproduction, a cyclical process in which models generate, curate and retrain on novel data to refine capabilities. Current methods, however, risk getting stuck in a data reproduction trap: optimizing outputs within fixed human-generated distributions in a closed loop leads to stagnation, as models merely recombine existing knowledge rather than explore new frontiers. In this paper, we propose language games as a pathway to expanded data reproduction, breaking this cycle through three mechanisms: (1) \textit{role fluidity}, which enhances data diversity and coverage by enabling multi-agent systems to dynamically shift roles across tasks; (2) \textit{reward variety}, embedding multiple feedback criteria that can drive complex intelligent behaviors; and (3) \textit{rule plasticity}, iteratively evolving interaction constraints to foster learnability, thereby injecting continual novelty. By scaling language games into global sociotechnical ecosystems, human-AI co-evolution generates unbounded data streams that drive open-ended exploration. This framework redefines data reproduction not as a closed loop but as an engine for superhuman intelligence.
2501.18929
Training-free Quantum-Inspired Image Edge Extraction Method
cs.CV
Edge detection is a cornerstone of image processing, yet existing methods often face critical limitations. Traditional deep learning edge detection methods require extensive training datasets and fine-tuning, while classical techniques often fail in complex or noisy scenarios, limiting their real-world applicability. To address these limitations, we propose a training-free, quantum-inspired edge detection model. Our approach integrates classical Sobel edge detection, the Schr\"odinger wave equation refinement, and a hybrid framework combining Canny and Laplacian operators. By eliminating the need for training, the model is lightweight and adaptable to diverse applications. The Schr\"odinger wave equation refines gradient-based edge maps through iterative diffusion, significantly enhancing edge precision. The hybrid framework further strengthens the model by synergistically combining local and global features, ensuring robustness even under challenging conditions. Extensive evaluations on datasets like BIPED, Multicue, and NYUD demonstrate superior performance of the proposed model, achieving state-of-the-art metrics, including ODS, OIS, AP, and F-measure. Noise robustness experiments highlight its reliability, showcasing its practicality for real-world scenarios. Due to its versatile and adaptable nature, our model is well-suited for applications such as medical imaging, autonomous systems, and environmental monitoring, setting a new benchmark for edge detection.
2501.18935
TabFSBench: Tabular Benchmark for Feature Shifts in Open Environment
cs.LG
Tabular data is widely utilized in various machine learning tasks. Current tabular learning research predominantly focuses on closed environments, while in real-world applications, open environments are often encountered, where distribution and feature shifts occur, leading to significant degradation in model performance. Previous research has primarily concentrated on mitigating distribution shifts, whereas feature shifts, a distinctive and unexplored challenge of tabular data, have garnered limited attention. To this end, this paper conducts the first comprehensive study on feature shifts in tabular data and introduces the first tabular feature-shift benchmark (TabFSBench). TabFSBench evaluates impacts of four distinct feature-shift scenarios on four tabular model categories across various datasets and assesses the performance of large language models (LLMs) and tabular LLMs in the tabular benchmark for the first time. Our study demonstrates three main observations: (1) most tabular models have the limited applicability in feature-shift scenarios; (2) the shifted feature set importance has a linear relationship with model performance degradation; (3) model performance in closed environments correlates with feature-shift performance. Future research direction is also explored for each observation. TabFSBench is released for public access by using a few lines of Python codes at https://github.com/LAMDASZ-ML/TabFSBench.
2501.18936
Adaptive Prompt: Unlocking the Power of Visual Prompt Tuning
cs.LG cs.CV
Visual Prompt Tuning (VPT) has recently emerged as a powerful method for adapting pre-trained vision models to downstream tasks. By introducing learnable prompt tokens as task-specific instructions, VPT effectively guides pre-trained transformer models with minimal overhead. Despite its empirical success, a comprehensive theoretical understanding of VPT remains an active area of research. Building on recent insights into the connection between mixture of experts and prompt-based approaches, we identify a key limitation in VPT: the restricted functional expressiveness in prompt formulation. To address this limitation, we propose Visual Adaptive Prompt Tuning (VAPT), a new generation of prompts that redefines prompts as adaptive functions of the input. Our theoretical analysis shows that this simple yet intuitive approach achieves optimal sample efficiency. Empirical results on VTAB-1K and FGVC further demonstrate VAPT's effectiveness, with performance gains of 7.34% and 1.04% over fully fine-tuning baselines, respectively. Notably, VAPT also surpasses VPT by a substantial margin while using fewer parameters. These results highlight both the effectiveness and efficiency of our method and pave the way for future research to explore the potential of adaptive prompts.
2501.18940
TV-Dialogue: Crafting Theme-Aware Video Dialogues with Immersive Interaction
cs.CV
Recent advancements in LLMs have accelerated the development of dialogue generation across text and images, yet video-based dialogue generation remains underexplored and presents unique challenges. In this paper, we introduce Theme-aware Video Dialogue Crafting (TVDC), a novel task aimed at generating new dialogues that align with video content and adhere to user-specified themes. We propose TV-Dialogue, a novel multi-modal agent framework that ensures both theme alignment (i.e., the dialogue revolves around the theme) and visual consistency (i.e., the dialogue matches the emotions and behaviors of characters in the video) by enabling real-time immersive interactions among video characters, thereby accurately understanding the video content and generating new dialogue that aligns with the given themes. To assess the generated dialogues, we present a multi-granularity evaluation benchmark with high accuracy, interpretability and reliability, demonstrating the effectiveness of TV-Dialogue on self-collected dataset over directly using existing LLMs. Extensive experiments reveal that TV-Dialogue can generate dialogues for videos of any length and any theme in a zero-shot manner without training. Our findings underscore the potential of TV-Dialogue for various applications, such as video re-creation, film dubbing and its use in downstream multimodal tasks.
2501.18942
Open-Source Autonomous Driving Software Platforms: Comparison of Autoware and Apollo
cs.RO
Full-stack autonomous driving system spans diverse technological domains-including perception, planning, and control-that each require in-depth research. Moreover, validating such technologies of the system necessitates extensive supporting infrastructure, from simulators and sensors to high-definition maps. These complexities with barrier to entry pose substantial limitations for individual developers and research groups. Recently, open-source autonomous driving software platforms have emerged to address this challenge by providing autonomous driving technologies and practical supporting infrastructure for implementing and evaluating autonomous driving functionalities. Among the prominent open-source platforms, Autoware and Apollo are frequently adopted in both academia and industry. While previous studies have assessed each platform independently, few have offered a quantitative and detailed head-to-head comparison of their capabilities. In this paper, we systematically examine the core modules of Autoware and Apollo and evaluate their middleware performance to highlight key differences. These insights serve as a practical reference for researchers and engineers, guiding them in selecting the most suitable platform for their specific development environments and advancing the field of full-stack autonomous driving system.
2501.18943
HeLiOS: Heterogeneous LiDAR Place Recognition via Overlap-based Learning and Local Spherical Transformer
cs.RO
LiDAR place recognition is a crucial module in localization that matches the current location with previously observed environments. Most existing approaches in LiDAR place recognition dominantly focus on the spinning type LiDAR to exploit its large FOV for matching. However, with the recent emergence of various LiDAR types, the importance of matching data across different LiDAR types has grown significantly-a challenge that has been largely overlooked for many years. To address these challenges, we introduce HeLiOS, a deep network tailored for heterogeneous LiDAR place recognition, which utilizes small local windows with spherical transformers and optimal transport-based cluster assignment for robust global descriptors. Our overlap-based data mining and guided-triplet loss overcome the limitations of traditional distance-based mining and discrete class constraints. HeLiOS is validated on public datasets, demonstrating performance in heterogeneous LiDAR place recognition while including an evaluation for long-term recognition, showcasing its ability to handle unseen LiDAR types. We release the HeLiOS code as an open source for the robotics community at https://github.com/minwoo0611/HeLiOS.
2501.18944
O-MAPL: Offline Multi-agent Preference Learning
cs.LG cs.MA
Inferring reward functions from demonstrations is a key challenge in reinforcement learning (RL), particularly in multi-agent RL (MARL), where large joint state-action spaces and complex inter-agent interactions complicate the task. While prior single-agent studies have explored recovering reward functions and policies from human preferences, similar work in MARL is limited. Existing methods often involve separate stages of supervised reward learning and MARL algorithms, leading to unstable training. In this work, we introduce a novel end-to-end preference-based learning framework for cooperative MARL, leveraging the underlying connection between reward functions and soft Q-functions. Our approach uses a carefully-designed multi-agent value decomposition strategy to improve training efficiency. Extensive experiments on SMAC and MAMuJoCo benchmarks show that our algorithm outperforms existing methods across various tasks.
2501.18945
Solving Inverse Problem for Multi-armed Bandits via Convex Optimization
cs.CE cs.LG math.OC q-bio.NC
We consider the inverse problem of multi-armed bandits (IMAB) that are widely used in neuroscience and psychology research for behavior modelling. We first show that the IMAB problem is not convex in general, but can be relaxed to a convex problem via variable transformation. Based on this result, we propose a two-step sequential heuristic for (approximately) solving the IMAB problem. We discuss a condition where our method provides global solution to the IMAB problem with certificate, as well as approximations to further save computing time. Numerical experiments indicate that our heuristic method is more robust than directly solving the IMAB problem via repeated local optimization, and can achieve the performance of Monte Carlo methods within a significantly decreased running time. We provide the implementation of our method based on CVXPY, which allows straightforward application by users not well versed in convex optimization.
2501.18950
Fantastic Targets for Concept Erasure in Diffusion Models and Where To Find Them
cs.LG cs.AI cs.CV
Concept erasure has emerged as a promising technique for mitigating the risk of harmful content generation in diffusion models by selectively unlearning undesirable concepts. The common principle of previous works to remove a specific concept is to map it to a fixed generic concept, such as a neutral concept or just an empty text prompt. In this paper, we demonstrate that this fixed-target strategy is suboptimal, as it fails to account for the impact of erasing one concept on the others. To address this limitation, we model the concept space as a graph and empirically analyze the effects of erasing one concept on the remaining concepts. Our analysis uncovers intriguing geometric properties of the concept space, where the influence of erasing a concept is confined to a local region. Building on this insight, we propose the Adaptive Guided Erasure (AGE) method, which \emph{dynamically} selects optimal target concepts tailored to each undesirable concept, minimizing unintended side effects. Experimental results show that AGE significantly outperforms state-of-the-art erasure methods on preserving unrelated concepts while maintaining effective erasure performance. Our code is published at {https://github.com/tuananhbui89/Adaptive-Guided-Erasure}.
2501.18954
LLMDet: Learning Strong Open-Vocabulary Object Detectors under the Supervision of Large Language Models
cs.CV
Recent open-vocabulary detectors achieve promising performance with abundant region-level annotated data. In this work, we show that an open-vocabulary detector co-training with a large language model by generating image-level detailed captions for each image can further improve performance. To achieve the goal, we first collect a dataset, GroundingCap-1M, wherein each image is accompanied by associated grounding labels and an image-level detailed caption. With this dataset, we finetune an open-vocabulary detector with training objectives including a standard grounding loss and a caption generation loss. We take advantage of a large language model to generate both region-level short captions for each region of interest and image-level long captions for the whole image. Under the supervision of the large language model, the resulting detector, LLMDet, outperforms the baseline by a clear margin, enjoying superior open-vocabulary ability. Further, we show that the improved LLMDet can in turn build a stronger large multi-modal model, achieving mutual benefits. The code, model, and dataset is available at https://github.com/iSEE-Laboratory/LLMDet.
2501.18955
Deep Learning based Quasi-consciousness Training for Robot Intelligent Model
cs.RO cs.AI
This paper explores a deep learning based robot intelligent model that renders robots learn and reason for complex tasks. First, by constructing a network of environmental factor matrix to stimulate the learning process of the robot intelligent model, the model parameters must be subjected to coarse & fine tuning to optimize the loss function for minimizing the loss score, meanwhile robot intelligent model can fuse all previously known concepts together to represent things never experienced before, which need robot intelligent model can be generalized extensively. Secondly, in order to progressively develop a robot intelligent model with primary consciousness, every robot must be subjected to at least 1~3 years of special school for training anthropomorphic behaviour patterns to understand and process complex environmental information and make rational decisions. This work explores and delivers the potential application of deep learning-based quasi-consciousness training in the field of robot intelligent model.
2501.18956
Differentiable Simulation of Soft Robots with Frictional Contacts
cs.RO
In recent years, soft robotics simulators have evolved to offer various functionalities, including the simulation of different material types (e.g., elastic, hyper-elastic) and actuation methods (e.g., pneumatic, cable-driven, servomotor). These simulators also provide tools for various tasks, such as calibration, design, and control. However, efficiently and accurately computing derivatives within these simulators remains a challenge, particularly in the presence of physical contact interactions. Incorporating these derivatives can, for instance, significantly improve the convergence speed of control methods like reinforcement learning and trajectory optimization, enable gradient-based techniques for design, or facilitate end-to-end machine-learning approaches for model reduction. This paper addresses these challenges by introducing a unified method for computing the derivatives of mechanical equations within the finite element method framework, including contact interactions modeled as a nonlinear complementarity problem. The proposed approach handles both collision and friction phases, accounts for their nonsmooth dynamics, and leverages the sparsity introduced by mesh-based models. Its effectiveness is demonstrated through several examples of controlling and calibrating soft systems.
2501.18957
Intrinsic Tensor Field Propagation in Large Language Models: A Novel Approach to Contextual Information Flow
cs.CL
Context propagation remains a central challenge in language model architectures, particularly in tasks requiring the retention of long-range dependencies. Conventional attention mechanisms, while effective in many applications, exhibit limitations in maintaining coherent contextual representations over extended sequences due to their reliance on discrete token interactions. A novel approach is introduced through the formulation of Intrinsic Tensor Field Propagation (ITFP), which models contextual relationships as continuous tensor fields distributed across token embeddings. The propagation dynamics are governed through differential equations that enable a structured flow of contextual information, augmenting the standard attention mechanism to enhance coherence and recall. A series of experiments conducted on an open-source transformer-based model demonstrate that ITFP provides measurable improvements in contextual retention, dependency resolution, and inference stability across various linguistic structures. Comparisons with baseline models reveal a reduction in syntactic inconsistencies and factual errors, while ablation studies indicate that the choice of propagation depth and integration strength significantly impacts model performance. Additional evaluations assessing domain generalization suggest that ITFP effectively adapts across different text genres, reinforcing its applicability beyond conventional language modeling tasks. Although computational trade-offs are introduced through the inclusion of tensor field computations, empirical findings suggest that the benefits in accuracy and coherence outweigh the increased processing demands.
2501.18959
Enhancing Neural Function Approximation: The XNet Outperforming KAN
cs.LG cs.AI
XNet is a single-layer neural network architecture that leverages Cauchy integral-based activation functions for high-order function approximation. Through theoretical analysis, we show that the Cauchy activation functions used in XNet can achieve arbitrary-order polynomial convergence, fundamentally outperforming traditional MLPs and Kolmogorov-Arnold Networks (KANs) that rely on increased depth or B-spline activations. Our extensive experiments on function approximation, PDE solving, and reinforcement learning demonstrate XNet's superior performance - reducing approximation error by up to 50000 times and accelerating training by up to 10 times compared to existing approaches. These results establish XNet as a highly efficient architecture for both scientific computing and AI applications.
2501.18962
Spend Wisely: Maximizing Post-Training Gains in Iterative Synthetic Data Boostrapping
cs.LG
Modern foundation models often undergo iterative ``bootstrapping'' in their post-training phase: a model generates synthetic data, an external verifier filters out low-quality samples, and the high-quality subset is used for further fine-tuning. Over multiple iterations, the model's performance improves--raising a crucial question: how should the total budget on generation and training be allocated across iterations to maximize final performance? In this work, we develop a theoretical framework to analyze budget allocation strategies. Specifically, we show that constant policies fail to converge with high probability, while increasing policies--particularly exponential growth policies--exhibit significant theoretical advantages. Experiments on image denoising with diffusion probabilistic models and math reasoning with large language models show that both exponential and polynomial growth policies consistently outperform constant policies, with exponential policies often providing more stable performance.
2501.18963
Optimizing Through Change: Bounds and Recommendations for Time-Varying Bayesian Optimization Algorithms
stat.ML cs.LG
Time-Varying Bayesian Optimization (TVBO) is the go-to framework for optimizing a time-varying, expensive, noisy black-box function. However, most of the solutions proposed so far either rely on unrealistic assumptions on the nature of the objective function or do not offer any theoretical guarantees. We propose the first analysis that asymptotically bounds the cumulative regret of TVBO algorithms under mild and realistic assumptions only. In particular, we provide an algorithm-independent lower regret bound and an upper regret bound that holds for a large class of TVBO algorithms. Based on this analysis, we formulate recommendations for TVBO algorithms and show how an algorithm (BOLT) that follows them performs better than the state-of-the-art of TVBO through experiments on synthetic and real-world problems.
2501.18965
The Surprising Agreement Between Convex Optimization Theory and Learning-Rate Scheduling for Large Model Training
cs.LG math.OC stat.ML
We show that learning-rate schedules for large model training behave surprisingly similar to a performance bound from non-smooth convex optimization theory. We provide a bound for the constant schedule with linear cooldown; in particular, the practical benefit of cooldown is reflected in the bound due to the absence of logarithmic terms. Further, we show that this surprisingly close match between optimization theory and practice can be exploited for learning-rate tuning: we achieve noticeable improvements for training 124M and 210M Llama-type models by (i) extending the schedule for continued training with optimal learning-rate, and (ii) transferring the optimal learning-rate across schedules.
2501.18972
BCAT: A Block Causal Transformer for PDE Foundation Models for Fluid Dynamics
cs.LG cs.NA math.NA
We introduce BCAT, a PDE foundation model designed for autoregressive prediction of solutions to two dimensional fluid dynamics problems. Our approach uses a block causal transformer architecture to model next frame predictions, leveraging previous frames as contextual priors rather than relying solely on sub-frames or pixel-based inputs commonly used in image generation methods. This block causal framework more effectively captures the spatial dependencies inherent in nonlinear spatiotemporal dynamics and physical phenomena. In an ablation study, next frame prediction demonstrated a 2.9x accuracy improvement over next token prediction. BCAT is trained on a diverse range of fluid dynamics datasets, including incompressible and compressible Navier-Stokes equations across various geometries and parameter regimes, as well as the shallow-water equations. The model's performance was evaluated on 6 distinct downstream prediction tasks and tested on about 8K trajectories to measure robustness on a variety of fluid dynamics simulations. BCAT achieved an average relative error of 1.92% across all evaluation tasks, outperforming prior approaches on standard benchmarks.
2501.18973
GPO-VAE: Modeling Explainable Gene Perturbation Responses utilizing GRN-Aligned Parameter Optimization
cs.LG cs.AI
Motivation: Predicting cellular responses to genetic perturbations is essential for understanding biological systems and developing targeted therapeutic strategies. While variational autoencoders (VAEs) have shown promise in modeling perturbation responses, their limited explainability poses a significant challenge, as the learned features often lack clear biological meaning. Nevertheless, model explainability is one of the most important aspects in the realm of biological AI. One of the most effective ways to achieve explainability is incorporating the concept of gene regulatory networks (GRNs) in designing deep learning models such as VAEs. GRNs elicit the underlying causal relationships between genes and are capable of explaining the transcriptional responses caused by genetic perturbation treatments. Results: We propose GPO-VAE, an explainable VAE enhanced by GRN-aligned Parameter Optimization that explicitly models gene regulatory networks in the latent space. Our key approach is to optimize the learnable parameters related to latent perturbation effects towards GRN-aligned explainability. Experimental results on perturbation prediction show our model achieves state-of-the-art performance in predicting transcriptional responses across multiple benchmark datasets. Furthermore, additional results on evaluating the GRN inference task reveal our model's ability to generate meaningful GRNs compared to other methods. According to qualitative analysis, GPO-VAE posseses the ability to construct biologically explainable GRNs that align with experimentally validated regulatory pathways. GPO-VAE is available at https://github.com/dmis-lab/GPO-VAE
2501.18975
Meta-learning of shared linear representations beyond well-specified linear regression
cs.LG stat.ML
Motivated by multi-task and meta-learning approaches, we consider the problem of learning structure shared by tasks or users, such as shared low-rank representations or clustered structures. While all previous works focus on well-specified linear regression, we consider more general convex objectives, where the structural low-rank and cluster assumptions are expressed on the optima of each function. We show that under mild assumptions such as \textit{Hessian concentration} and \textit{noise concentration at the optimum}, rank and clustered regularized estimators recover such structure, provided the number of samples per task and the number of tasks are large enough. We then study the problem of recovering the subspace in which all the solutions lie, in the setting where there is only a single sample per task: we show that in that case, the rank-constrained estimator can recover the subspace, but that the number of tasks needs to scale exponentially large with the dimension of the subspace. Finally, we provide a polynomial-time algorithm via nuclear norm constraints for learning a shared linear representation in the context of convex learning objectives.
2501.18977
Blocked Bloom Filters with Choices
cs.DB cs.DS
Probabilistic filters are approximate set membership data structures that represent a set of keys in small space, and answer set membership queries without false negative answers, but with a certain allowed false positive probability. Such filters are widely used in database systems, networks, storage systems and in biological sequence analysis because of their fast query times and low space requirements. Starting with Bloom filters in the 1970s, many filter data structures have been developed, each with its own advantages and disadvantages, e.g., Blocked Bloom filters, Cuckoo filters, XOR filters, Ribbon filters, and more. We introduce Blocked Bloom filters with choices that work similarly to Blocked Bloom filters, except that for each key there are two (or more) alternative choices of blocks where the key's information may be stored. The result is a filter that partially inherits the advantages of a Blocked Bloom filter, such as the ability to insert keys rapidly online or the ability to slightly overload the filter with only a small penalty to the false positive rate. At the same time, it avoids the major disadvantage of a Blocked Bloom filter, namely the larger space consumption. Our new data structure uses less space at the same false positive rate, or has a lower false positive rate at the same space consumption as a Blocked Bloom filter. We discuss the methodology, engineered implementation, a detailed performance evaluation and use cases in bioinformatics of Blocked Bloom filters with choices, showing that they can be of practical value. The implementation of the evaluated filters and the workflows used are provided via Gitlab at https://gitlab.com/rahmannlab/blowchoc-filters.
2501.18980
Symmetric Pruning of Large Language Models
cs.LG cs.AI
Popular post-training pruning methods such as Wanda and RIA are known for their simple, yet effective, designs that have shown exceptional empirical performance. Wanda optimizes performance through calibrated activations during pruning, while RIA emphasizes the relative, rather than absolute, importance of weight elements. Despite their practical success, a thorough theoretical foundation explaining these outcomes has been lacking. This paper introduces new theoretical insights that redefine the standard minimization objective for pruning, offering a deeper understanding of the factors contributing to their success. Our study extends beyond these insights by proposing complementary strategies that consider both input activations and weight significance. We validate these approaches through rigorous experiments, demonstrating substantial enhancements over existing methods. Furthermore, we introduce a novel training-free fine-tuning approach $R^2$-DSnoT that incorporates relative weight importance and a regularized decision boundary within a dynamic pruning-and-growing framework, significantly outperforming strong baselines and establishing a new state of the art.
2501.18982
OmniPhysGS: 3D Constitutive Gaussians for General Physics-Based Dynamics Generation
cs.CV
Recently, significant advancements have been made in the reconstruction and generation of 3D assets, including static cases and those with physical interactions. To recover the physical properties of 3D assets, existing methods typically assume that all materials belong to a specific predefined category (e.g., elasticity). However, such assumptions ignore the complex composition of multiple heterogeneous objects in real scenarios and tend to render less physically plausible animation given a wider range of objects. We propose OmniPhysGS for synthesizing a physics-based 3D dynamic scene composed of more general objects. A key design of OmniPhysGS is treating each 3D asset as a collection of constitutive 3D Gaussians. For each Gaussian, its physical material is represented by an ensemble of 12 physical domain-expert sub-models (rubber, metal, honey, water, etc.), which greatly enhances the flexibility of the proposed model. In the implementation, we define a scene by user-specified prompts and supervise the estimation of material weighting factors via a pretrained video diffusion model. Comprehensive experiments demonstrate that OmniPhysGS achieves more general and realistic physical dynamics across a broader spectrum of materials, including elastic, viscoelastic, plastic, and fluid substances, as well as interactions between different materials. Our method surpasses existing methods by approximately 3% to 16% in metrics of visual quality and text alignment.
2501.18984
Context Matters: Query-aware Dynamic Long Sequence Modeling of Gigapixel Images
cs.CV
Whole slide image (WSI) analysis presents significant computational challenges due to the massive number of patches in gigapixel images. While transformer architectures excel at modeling long-range correlations through self-attention, their quadratic computational complexity makes them impractical for computational pathology applications. Existing solutions like local-global or linear self-attention reduce computational costs but compromise the strong modeling capabilities of full self-attention. In this work, we propose Querent, i.e., the query-aware long contextual dynamic modeling framework, which maintains the expressive power of full self-attention while achieving practical efficiency. Our method adaptively predicts which surrounding regions are most relevant for each patch, enabling focused yet unrestricted attention computation only with potentially important contexts. By using efficient region-wise metadata computation and importance estimation, our approach dramatically reduces computational overhead while preserving global perception to model fine-grained patch correlations. Through comprehensive experiments on biomarker prediction, gene mutation prediction, cancer subtyping, and survival analysis across over 10 WSI datasets, our method demonstrates superior performance compared to the state-of-the-art approaches. Code will be made available at https://github.com/dddavid4real/Querent.
2501.18989
Extension of Optimal Locally Repairable codes
cs.IT math.IT
Recent studies have delved into the construction of locally repairable codes (LRCs) with optimal minimum distance from function fields. In this paper, we present several novel constructions by extending the findings of optimally designed locally repairable codes documented in the literature. Let $C$ denote an optimal LRC of locality $r$, implying that every repairable block of $C$ is a $[r+1, r]$ MDS code, and $C$ maximizes its minimum distance. By extending a single coordinate of one of these blocks, we demonstrate that the resulting code remains an optimally designed locally repairable code. This suggests that the maximal length of an optimal LRC from rational function fields can be extended up to $q+2$ over a finite field $\mathbb{F}_q$. In addition, we give a new construction of optimal $(r, 3)$-LRC by extending one coordinate in each block within $C$. Furthermore, we propose a novel family of LRCs with Roth-Lempel type that are optimal under certain conditions. Finally, we explore optimal LRCs derived from elliptic function fields and extend a single coordinate of such codes. This approach leads us to confirm that the new codes are also optimal, thereby allowing their lengths to reach $q + 2\sqrt{q} - 2r - 2$ with locality $r$. We also consider the construction of optimal $(r, 3)$-LRC in elliptic function fields, with exploring one more condition.
2501.18990
Permutation-Based Rank Test in the Presence of Discretization and Application in Causal Discovery with Mixed Data
cs.LG
Recent advances have shown that statistical tests for the rank of cross-covariance matrices play an important role in causal discovery. These rank tests include partial correlation tests as special cases and provide further graphical information about latent variables. Existing rank tests typically assume that all the continuous variables can be perfectly measured, and yet, in practice many variables can only be measured after discretization. For example, in psychometric studies, the continuous level of certain personality dimensions of a person can only be measured after being discretized into order-preserving options such as disagree, neutral, and agree. Motivated by this, we propose Mixed data Permutation-based Rank Test (MPRT), which properly controls the statistical errors even when some or all variables are discretized. Theoretically, we establish the exchangeability and estimate the asymptotic null distribution by permutations; as a consequence, MPRT can effectively control the Type I error in the presence of discretization while previous methods cannot. Empirically, our method is validated by extensive experiments on synthetic data and real-world data to demonstrate its effectiveness as well as applicability in causal discovery.
2501.18991
Optimal Transport-based Conformal Prediction
stat.ML cs.LG
Conformal Prediction (CP) is a principled framework for quantifying uncertainty in blackbox learning models, by constructing prediction sets with finite-sample coverage guarantees. Traditional approaches rely on scalar nonconformity scores, which fail to fully exploit the geometric structure of multivariate outputs, such as in multi-output regression or multiclass classification. Recent methods addressing this limitation impose predefined convex shapes for the prediction sets, potentially misaligning with the intrinsic data geometry. We introduce a novel CP procedure handling multivariate score functions through the lens of optimal transport. Specifically, we leverage Monge-Kantorovich vector ranks and quantiles to construct prediction region with flexible, potentially non-convex shapes, better suited to the complex uncertainty patterns encountered in multivariate learning tasks. We prove that our approach ensures finite-sample, distribution-free coverage properties, similar to typical CP methods. We then adapt our method for multi-output regression and multiclass classification, and also propose simple adjustments to generate adaptive prediction regions with asymptotic conditional coverage guarantees. Finally, we evaluate our method on practical regression and classification problems, illustrating its advantages in terms of (conditional) coverage and efficiency.
2501.18993
Visual Autoregressive Modeling for Image Super-Resolution
cs.CV
Image Super-Resolution (ISR) has seen significant progress with the introduction of remarkable generative models. However, challenges such as the trade-off issues between fidelity and realism, as well as computational complexity, have also posed limitations on their application. Building upon the tremendous success of autoregressive models in the language domain, we propose \textbf{VARSR}, a novel visual autoregressive modeling for ISR framework with the form of next-scale prediction. To effectively integrate and preserve semantic information in low-resolution images, we propose using prefix tokens to incorporate the condition. Scale-aligned Rotary Positional Encodings are introduced to capture spatial structures and the diffusion refiner is utilized for modeling quantization residual loss to achieve pixel-level fidelity. Image-based Classifier-free Guidance is proposed to guide the generation of more realistic images. Furthermore, we collect large-scale data and design a training process to obtain robust generative priors. Quantitative and qualitative results show that VARSR is capable of generating high-fidelity and high-realism images with more efficiency than diffusion-based methods. Our codes will be released at https://github.com/qyp2000/VARSR.
2501.18994
VKFPos: A Learning-Based Monocular Positioning with Variational Bayesian Extended Kalman Filter Integration
cs.CV cs.AI
This paper addresses the challenges in learning-based monocular positioning by proposing VKFPos, a novel approach that integrates Absolute Pose Regression (APR) and Relative Pose Regression (RPR) via an Extended Kalman Filter (EKF) within a variational Bayesian inference framework. Our method shows that the essential posterior probability of the monocular positioning problem can be decomposed into APR and RPR components. This decomposition is embedded in the deep learning model by predicting covariances in both APR and RPR branches, allowing them to account for associated uncertainties. These covariances enhance the loss functions and facilitate EKF integration. Experimental evaluations on both indoor and outdoor datasets show that the single-shot APR branch achieves accuracy on par with state-of-the-art methods. Furthermore, for temporal positioning, where consecutive images allow for RPR and EKF integration, VKFPos outperforms temporal APR and model-based integration methods, achieving superior accuracy.
2501.18997
Collaborative Diffusion Model for Recommender System
cs.IR
Diffusion-based recommender systems (DR) have gained increasing attention for their advanced generative and denoising capabilities. However, existing DR face two central limitations: (i) a trade-off between enhancing generative capacity via noise injection and retaining the loss of personalized information. (ii) the underutilization of rich item-side information. To address these challenges, we present a Collaborative Diffusion model for Recommender System (CDiff4Rec). Specifically, CDiff4Rec generates pseudo-users from item features and leverages collaborative signals from both real and pseudo personalized neighbors identified through behavioral similarity, thereby effectively reconstructing nuanced user preferences. Experimental results on three public datasets show that CDiff4Rec outperforms competitors by effectively mitigating the loss of personalized information through the integration of item content and collaborative signals.
2501.18998
Adversarial Attacks on AI-Generated Text Detection Models: A Token Probability-Based Approach Using Embeddings
cs.CL cs.AI cs.LG
In recent years, text generation tools utilizing Artificial Intelligence (AI) have occasionally been misused across various domains, such as generating student reports or creative writings. This issue prompts plagiarism detection services to enhance their capabilities in identifying AI-generated content. Adversarial attacks are often used to test the robustness of AI-text generated detectors. This work proposes a novel textual adversarial attack on the detection models such as Fast-DetectGPT. The method employs embedding models for data perturbation, aiming at reconstructing the AI generated texts to reduce the likelihood of detection of the true origin of the texts. Specifically, we employ different embedding techniques, including the Tsetlin Machine (TM), an interpretable approach in machine learning for this purpose. By combining synonyms and embedding similarity vectors, we demonstrates the state-of-the-art reduction in detection scores against Fast-DetectGPT. Particularly, in the XSum dataset, the detection score decreased from 0.4431 to 0.2744 AUROC, and in the SQuAD dataset, it dropped from 0.5068 to 0.3532 AUROC.
2501.19001
Quantum SMOTE with Angular Outliers: Redefining Minority Class Handling
quant-ph cs.LG
This paper introduces Quantum-SMOTEV2, an advanced variant of the Quantum-SMOTE method, leveraging quantum computing to address class imbalance in machine learning datasets without K-Means clustering. Quantum-SMOTEV2 synthesizes data samples using swap tests and quantum rotation centered around a single data centroid, concentrating on the angular distribution of minority data points and the concept of angular outliers (AOL). Experimental results show significant enhancements in model performance metrics at moderate SMOTE levels (30-36%), which previously required up to 50% with the original method. Quantum-SMOTEV2 maintains essential features of its predecessor (arXiv:2402.17398), such as rotation angle, minority percentage, and splitting factor, allowing for tailored adaptation to specific dataset needs. The method is scalable, utilizing compact swap tests and low depth quantum circuits to accommodate a large number of features. Evaluation on the public Cell-to-Cell Telecom dataset with Random Forest (RF), K-Nearest Neighbours (KNN) Classifier, and Neural Network (NN) illustrates that integrating Angular Outliers modestly boosts classification metrics like accuracy, F1 Score, AUC-ROC, and AUC-PR across different proportions of synthetic data, highlighting the effectiveness of Quantum-SMOTEV2 in enhancing model performance for edge cases.
2501.19003
Virtual airways heatmaps to optimize point of entry location in lung biopsy planning systems
cs.CV cs.AI
Purpose: We present a virtual model to optimize point of entry (POE) in lung biopsy planning systems. Our model allows to compute the quality of a biopsy sample taken from potential POE, taking into account the margin of error that arises from discrepancies between the orientation in the planning simulation and the actual orientation during the operation. Additionally, the study examines the impact of the characteristics of the lesion. Methods: The quality of the biopsy is given by a heatmap projected onto the skeleton of a patient-specific model of airways. The skeleton provides a 3D representation of airways structure, while the heatmap intensity represents the potential amount of tissue that it could be extracted from each POE. This amount of tissue is determined by the intersection of the lesion with a cone that represents the uncertainty area in the introduction of biopsy instruments. The cone, lesion, and skeleton are modelled as graphical objects that define a 3D scene of the intervention. Results: We have simulated different settings of the intervention scene from a single anatomy extracted from a CT scan and two lesions with regular and irregular shapes. The different scenarios are simulated by systematic rotation of each lesion placed at different distances from airways. Analysis of the heatmaps for the different settings show a strong impact of lesion orientation for irregular shape and the distance for both shapes. Conclusion: The proposed heatmaps help to visually assess the optimal POE and identify whether multiple optimal POEs exist in different zones of the bronchi. They also allow us to model the maximum allowable error in navigation systems and study which variables have the greatest influence on the success of the operation. Additionally, they help determine at what point this influence could potentially jeopardize the operation.
2501.19004
CPU vs. GPU for Community Detection: Performance Insights from GVE-Louvain and $\nu$-Louvain
cs.DC cs.SI
Community detection involves identifying natural divisions in networks, a crucial task for many large-scale applications. This report presents GVE-Louvain, one of the most efficient multicore implementations of the Louvain algorithm, a high-quality method for community detection. Running on a dual 16-core Intel Xeon Gold 6226R server, GVE-Louvain outperforms Vite, Grappolo, NetworKit Louvain, and cuGraph Louvain (on an NVIDIA A100 GPU) by factors of 50x, 22x, 20x, and 5.8x, respectively, achieving a processing rate of 560M edges per second on a 3.8B-edge graph. Additionally, it scales efficiently, improving performance by 1.6x for every thread doubling. The paper also presents $\nu$-Louvain, a GPU-based implementation. When evaluated on an NVIDIA A100 GPU, $\nu$-Louvain performs only on par with GVE-Louvain, largely due to reduced workload and parallelism in later algorithmic passes. These results suggest that CPUs, with their flexibility in handling irregular workloads, may be better suited for community detection tasks.
2501.19010
DyPCL: Dynamic Phoneme-level Contrastive Learning for Dysarthric Speech Recognition
cs.CL cs.SD eess.AS
Dysarthric speech recognition often suffers from performance degradation due to the intrinsic diversity of dysarthric severity and extrinsic disparity from normal speech. To bridge these gaps, we propose a Dynamic Phoneme-level Contrastive Learning (DyPCL) method, which leads to obtaining invariant representations across diverse speakers. We decompose the speech utterance into phoneme segments for phoneme-level contrastive learning, leveraging dynamic connectionist temporal classification alignment. Unlike prior studies focusing on utterance-level embeddings, our granular learning allows discrimination of subtle parts of speech. In addition, we introduce dynamic curriculum learning, which progressively transitions from easy negative samples to difficult-to-distinguishable negative samples based on phonetic similarity of phoneme. Our approach to training by difficulty levels alleviates the inherent variability of speakers, better identifying challenging speeches. Evaluated on the UASpeech dataset, DyPCL outperforms baseline models, achieving an average 22.10\% relative reduction in word error rate (WER) across the overall dysarthria group.
2501.19012
Importing Phantoms: Measuring LLM Package Hallucination Vulnerabilities
cs.LG cs.CL cs.CR
Large Language Models (LLMs) have become an essential tool in the programmer's toolkit, but their tendency to hallucinate code can be used by malicious actors to introduce vulnerabilities to broad swathes of the software supply chain. In this work, we analyze package hallucination behaviour in LLMs across popular programming languages examining both existing package references and fictional dependencies. By analyzing this package hallucination behaviour we find potential attacks and suggest defensive strategies to defend against these attacks. We discover that package hallucination rate is predicated not only on model choice, but also programming language, model size, and specificity of the coding task request. The Pareto optimality boundary between code generation performance and package hallucination is sparsely populated, suggesting that coding models are not being optimized for secure code. Additionally, we find an inverse correlation between package hallucination rate and the HumanEval coding benchmark, offering a heuristic for evaluating the propensity of a model to hallucinate packages. Our metrics, findings and analyses provide a base for future models, securing AI-assisted software development workflows against package supply chain attacks.
2501.19013
On the efficiency of explicit and semi-explicit immersed boundary finite element methods for wave propagation problems
cs.CE cs.NA math.NA
Immersed boundary methods have attracted substantial interest in the last decades due to their potential for computations involving complex geometries. Often these cannot be efficiently discretized using boundary-fitted finite elements. Immersed boundary methods provide a simple and fully automatic discretization based on Cartesian grids and tailored quadrature schemes that account for the geometric model. It can thus be described independently of the grid, e.g., by image data obtained from computed tomography scans. The drawback of such a discretization lies in the potentially small overlap between certain elements in the grid and the geometry. These badly cut elements with small physical support pose a particular challenge for nonlinear and/or dynamic simulations. In this work, we focus on problems in structural dynamics and acoustics and concentrate on solving them with explicit time-marching schemes. In this context, badly cut elements can lead to unfeasibly small critical time step sizes. We investigate the performance of implicit-explicit time marching schemes and two stabilization methods developed in previous works as potential remedies. While these have been studied before with regard to their effectiveness in increasing the critical time step size, their numerical efficiency has only been considered in terms of accuracy per degree of freedom. In this paper, we evaluate the computation time required for a given accuracy, which depends not only on the number of degrees of freedom but also on the selected spatial discretization, the sparsity patterns of the system matrices, and the employed time-marching scheme.
2501.19016
Modelling Infodemics on a Global Scale: A 30 Countries Study using Epidemiological and Social Listening Data
cs.SI physics.soc-ph
Infodemics are a threat to public health, arising from multiple interacting phenomena occurring both online and offline. The continuous feedback loops between the digital information ecosystem and offline contingencies make infodemics particularly challenging to define operationally, measure, and eventually model in quantitative terms. In this study, we present evidence of the effect of various epidemic-related variables on the dynamics of infodemics, using a robust modelling framework applied to data from 30 countries across diverse income groups. We use WHO COVID-19 surveillance data on new cases and deaths, vaccination data from the Oxford COVID-19 Government Response Tracker, infodemic data (volume of public conversations and social media content) from the WHO EARS platform, and Google Trends data to represent information demand. Our findings show that new deaths are the strongest predictor of the infodemic, measured as new document production including social media content and public conversations, and that the epidemic burden in neighbouring countries appears to have a greater impact on document production than the domestic one. Building on these results, we propose a taxonomy that highlights country-specific discrepancies between the evolution of the infodemic and the epidemic. Further, an analysis of the temporal evolution of the relationship between the two phenomena quantifies how much the discussions around vaccine rollouts may have shaped the development of the infodemic. The insights from our quantitative model contribute to advancing infodemic research, highlighting the importance of a holistic approach integrating both online and offline dimensions.
2501.19017
Calling a Spade a Heart: Gaslighting Multimodal Large Language Models via Negation
cs.CL
Multimodal Large Language Models (MLLMs) have exhibited remarkable advancements in integrating different modalities, excelling in complex understanding and generation tasks. Despite their success, MLLMs remain vulnerable to conversational adversarial inputs, particularly negation arguments. This paper systematically evaluates state-of-the-art MLLMs across diverse benchmarks, revealing significant performance drops when negation arguments are introduced to initially correct responses. We show critical vulnerabilities in the reasoning and alignment mechanisms of these models. Proprietary models such as GPT-4o and Claude-3.5-Sonnet demonstrate better resilience compared to open-source counterparts like Qwen2-VL and LLaVA. However, all evaluated MLLMs struggle to maintain logical consistency under negation arguments during conversation. This paper aims to offer valuable insights for improving the robustness of MLLMs against adversarial inputs, contributing to the development of more reliable and trustworthy multimodal AI systems.
2501.19018
Scalable Multi-phase Word Embedding Using Conjunctive Propositional Clauses
cs.LG cs.CL
The Tsetlin Machine (TM) architecture has recently demonstrated effectiveness in Machine Learning (ML), particularly within Natural Language Processing (NLP). It has been utilized to construct word embedding using conjunctive propositional clauses, thereby significantly enhancing our understanding and interpretation of machine-derived decisions. The previous approach performed the word embedding over a sequence of input words to consolidate the information into a cohesive and unified representation. However, that approach encounters scalability challenges as the input size increases. In this study, we introduce a novel approach incorporating two-phase training to discover contextual embeddings of input sequences. Specifically, this method encapsulates the knowledge for each input word within the dataset's vocabulary, subsequently constructing embeddings for a sequence of input words utilizing the extracted knowledge. This technique not only facilitates the design of a scalable model but also preserves interpretability. Our experimental findings revealed that the proposed method yields competitive performance compared to the previous approaches, demonstrating promising results in contrast to human-generated benchmarks. Furthermore, we applied the proposed approach to sentiment analysis on the IMDB dataset, where the TM embedding and the TM classifier, along with other interpretable classifiers, offered a transparent end-to-end solution with competitive performance.
2501.19022
On the Impact of Noise in Differentially Private Text Rewriting
cs.CL
The field of text privatization often leverages the notion of $\textit{Differential Privacy}$ (DP) to provide formal guarantees in the rewriting or obfuscation of sensitive textual data. A common and nearly ubiquitous form of DP application necessitates the addition of calibrated noise to vector representations of text, either at the data- or model-level, which is governed by the privacy parameter $\varepsilon$. However, noise addition almost undoubtedly leads to considerable utility loss, thereby highlighting one major drawback of DP in NLP. In this work, we introduce a new sentence infilling privatization technique, and we use this method to explore the effect of noise in DP text rewriting. We empirically demonstrate that non-DP privatization techniques excel in utility preservation and can find an acceptable empirical privacy-utility trade-off, yet cannot outperform DP methods in empirical privacy protections. Our results highlight the significant impact of noise in current DP rewriting mechanisms, leading to a discussion of the merits and challenges of DP in NLP, as well as the opportunities that non-DP methods present.
2501.19025
Recognize then Resolve: A Hybrid Framework for Understanding Interaction and Cooperative Conflict Resolution in Mixed Traffic
cs.MA
A lack of understanding of interactions and the inability to effectively resolve conflicts continue to impede the progress of Connected Autonomous Vehicles (CAVs) in their interactions with Human-Driven Vehicles (HDVs). To address this challenge, we propose the Recognize then Resolve (RtR) framework. First, a Bilateral Intention Progression Graph (BIPG) is constructed based on CAV-HDV interaction data to model the evolution of interactions and identify potential HDV intentions. Three typical interaction breakdown scenarios are then categorized, and key moments are defined for triggering cooperative conflict resolution. On this basis, a constrained Monte Carlo Tree Search (MCTS) algorithm is introduced to determine the optimal passage order while accommodating HDV intentions. Experimental results demonstrate that the proposed RtR framework outperforms other cooperative approaches in terms of safety and efficiency across various penetration rates, achieving results close to consistent cooperation while significantly reducing computational resources. Our code and data are available at: https://github.com/FanGShiYuu/RtR-Recognize-then-Resolve/.
2501.19027
True Online TD-Replan(lambda) Achieving Planning through Replaying
cs.LG
In this paper, we develop a new planning method that extends the capabilities of the true online TD to allow an agent to efficiently replay all or part of its past experience, online in the sequence that they appear with, either in each step or sparsely according to the usual {\lambda} parameter. In this new method that we call True Online TD-Replan({\lambda}), the {\lambda} parameter plays a new role in specifying the density of the replay process in addition to the usual role of specifying the depth of the target's updates. We demonstrate that, for problems that benefit from experience replay, our new method outperforms true online TD({\lambda}), albeit quadratic in complexity due to its replay capabilities. In addition, we demonstrate that our method outperforms other methods with similar quadratic complexity such as Dyna Planning and TD({\lambda})-Replan algorithms. We test our method on two benchmarking environments, a random walk problem that uses simple binary features and a myoelectric control domain that uses both simple sEMG features and deeply extracted features to showcase its capabilities.
2501.19032
Error Slice Discovery via Manifold Compactness
cs.LG
Despite the great performance of deep learning models in many areas, they still make mistakes and underperform on certain subsets of data, i.e. error slices. Given a trained model, it is important to identify its semantically coherent error slices that are easy to interpret, which is referred to as the error slice discovery problem. However, there is no proper metric of slice coherence without relying on extra information like predefined slice labels. Current evaluation of slice coherence requires access to predefined slices formulated by metadata like attributes or subclasses. Its validity heavily relies on the quality and abundance of metadata, where some possible patterns could be ignored. Besides, current algorithms cannot directly incorporate the constraint of coherence into their optimization objective due to the absence of an explicit coherence metric, which could potentially hinder their effectiveness. In this paper, we propose manifold compactness, a coherence metric without reliance on extra information by incorporating the data geometry property into its design, and experiments on typical datasets empirically validate the rationality of the metric. Then we develop Manifold Compactness based error Slice Discovery (MCSD), a novel algorithm that directly treats risk and coherence as the optimization objective, and is flexible to be applied to models of various tasks. Extensive experiments on the benchmark and case studies on other typical datasets demonstrate the superiority of MCSD.
2501.19034
XRF V2: A Dataset for Action Summarization with Wi-Fi Signals, and IMUs in Phones, Watches, Earbuds, and Glasses
cs.CV
Human Action Recognition (HAR) plays a crucial role in applications such as health monitoring, smart home automation, and human-computer interaction. While HAR has been extensively studied, action summarization, which involves identifying and summarizing continuous actions, remains an emerging task. This paper introduces the novel XRF V2 dataset, designed for indoor daily activity Temporal Action Localization (TAL) and action summarization. XRF V2 integrates multimodal data from Wi-Fi signals, IMU sensors (smartphones, smartwatches, headphones, and smart glasses), and synchronized video recordings, offering a diverse collection of indoor activities from 16 volunteers across three distinct environments. To tackle TAL and action summarization, we propose the XRFMamba neural network, which excels at capturing long-term dependencies in untrimmed sensory sequences and outperforms state-of-the-art methods, such as ActionFormer and WiFiTAD. We envision XRF V2 as a valuable resource for advancing research in human action localization, action forecasting, pose estimation, multimodal foundation models pre-training, synthetic data generation, and more.
2501.19035
SynthmanticLiDAR: A Synthetic Dataset for Semantic Segmentation on LiDAR Imaging
cs.CV
Semantic segmentation on LiDAR imaging is increasingly gaining attention, as it can provide useful knowledge for perception systems and potential for autonomous driving. However, collecting and labeling real LiDAR data is an expensive and time-consuming task. While datasets such as SemanticKITTI have been manually collected and labeled, the introduction of simulation tools such as CARLA, has enabled the creation of synthetic datasets on demand. In this work, we present a modified CARLA simulator designed with LiDAR semantic segmentation in mind, with new classes, more consistent object labeling with their counterparts from real datasets such as SemanticKITTI, and the possibility to adjust the object class distribution. Using this tool, we have generated SynthmanticLiDAR, a synthetic dataset for semantic segmentation on LiDAR imaging, designed to be similar to SemanticKITTI, and we evaluate its contribution to the training process of different semantic segmentation algorithms by using a naive transfer learning approach. Our results show that incorporating SynthmanticLiDAR into the training process improves the overall performance of tested algorithms, proving the usefulness of our dataset, and therefore, our adapted CARLA simulator. The dataset and simulator are available in https://github.com/vpulab/SynthmanticLiDAR.
2501.19036
RedundancyLens: Revealing and Exploiting Visual Token Processing Redundancy for Efficient Decoder-Only MLLMs
cs.CV
Current Multimodal Large Language Model (MLLM) architectures face a critical tradeoff between performance and efficiency: decoder-only architectures achieve higher performance but lower efficiency, while cross-attention-based architectures offer greater efficiency but lower performance. The key distinction lies in how visual tokens are processed. Decoder-only architectures apply self-attention and FFN operations on visual tokens, while cross-attention architectures skip these computations. To investigate whether redundancy exists in this computationally expensive process, we propose a training-free framework for analyzing trained MLLMs. It consists of Probe-Activated Dynamic FFN and Hollow Attention, which enable adjustable reductions in computations for visual tokens, as well as a Layer Ranking Algorithm that prioritizes layers for these reductions. Extensive experiments demonstrate substantial, structured, and clustered redundancy unique to decoder-only MLLMs, offering valuable insights for future MLLM architecture design. Furthermore, by leveraging our reduction framework as a training-free inference acceleration approach, we achieve performance comparable to or better than state-of-the-art methods while remaining compatible with them. Code will be publicly available at https://github.com/L-Hugh/RedundancyLens.
2501.19038
Conformal Prediction in Hierarchical Classification
stat.ML cs.LG
Conformal prediction has emerged as a widely used framework for constructing valid prediction sets in classification and regression tasks. In this work, we extend the split conformal prediction framework to hierarchical classification, where prediction sets are commonly restricted to internal nodes of a predefined hierarchy, and propose two computationally efficient inference algorithms. The first algorithm returns internal nodes as prediction sets, while the second relaxes this restriction, using the notion of representation complexity, yielding a more general and combinatorial inference problem, but smaller set sizes. Empirical evaluations on several benchmark datasets demonstrate the effectiveness of the proposed algorithms in achieving nominal coverage.
2501.19040
Towards the Worst-case Robustness of Large Language Models
cs.LG
Recent studies have revealed the vulnerability of Large Language Models (LLMs) to adversarial attacks, where the adversary crafts specific input sequences to induce harmful, violent, private, or incorrect outputs. Although various defenses have been proposed, they have not been evaluated by strong adaptive attacks, leaving the worst-case robustness of LLMs still intractable. By developing a stronger white-box attack, our evaluation results indicate that most typical defenses achieve nearly 0\% robustness.To solve this, we propose \textit{DiffTextPure}, a general defense that diffuses the (adversarial) input prompt using any pre-defined smoothing distribution, and purifies the diffused input using a pre-trained language model. Theoretically, we derive tight robustness lower bounds for all smoothing distributions using Fractal Knapsack or 0-1 Knapsack solvers. Under this framework, we certify the robustness of a specific case -- smoothing LLMs using a uniform kernel -- against \textit{any possible attack} with an average $\ell_0$ perturbation of 2.02 or an average suffix length of 6.41.
2501.19042
Swarm-Gen: Fast Generation of Diverse Feasible Swarm Behaviors
cs.RO cs.AI
Coordination behavior in robot swarms is inherently multi-modal in nature. That is, there are numerous ways in which a swarm of robots can avoid inter-agent collisions and reach their respective goals. However, the problem of generating diverse and feasible swarm behaviors in a scalable manner remains largely unaddressed. In this paper, we fill this gap by combining generative models with a safety-filter (SF). Specifically, we sample diverse trajectories from a learned generative model which is subsequently projected onto the feasible set using the SF. We experiment with two choices for generative models, namely: Conditional Variational Autoencoder (CVAE) and Vector-Quantized Variational Autoencoder (VQ-VAE). We highlight the trade-offs these two models provide in terms of computation time and trajectory diversity. We develop a custom solver for our SF and equip it with a neural network that predicts context-specific initialization. Thecinitialization network is trained in a self-supervised manner, taking advantage of the differentiability of the SF solver. We provide two sets of empirical results. First, we demonstrate that we can generate a large set of multi-modal, feasible trajectories, simulating diverse swarm behaviors, within a few tens of milliseconds. Second, we show that our initialization network provides faster convergence of our SF solver vis-a-vis other alternative heuristics.
2501.19043
Self-Supervised Cross-Modal Text-Image Time Series Retrieval in Remote Sensing
cs.CV
The development of image time series retrieval (ITSR) methods is a growing research interest in remote sensing (RS). Given a user-defined image time series (i.e., the query time series), the ITSR methods search and retrieve from large archives the image time series that have similar content to the query time series. The existing ITSR methods in RS are designed for unimodal retrieval problems, limiting their usability and versatility. To overcome this issue, as a first time in RS we introduce the task of cross-modal text-ITSR. In particular, we present a self-supervised cross-modal text-image time series retrieval (text-ITSR) method that enables the retrieval of image time series using text sentences as queries, and vice versa. In detail, we focus our attention on text-ITSR in pairs of images (i.e., bitemporal images). The proposed text-ITSR method consists of two key components: 1) modality-specific encoders to model the semantic content of bitemporal images and text sentences with discriminative features; and 2) modality-specific projection heads to align textual and image representations in a shared embedding space. To effectively model the temporal information within the bitemporal images, we introduce two fusion strategies: i) global feature fusion (GFF) strategy that combines global image features through simple yet effective operators; and ii) transformer-based feature fusion (TFF) strategy that leverages transformers for fine-grained temporal integration. Extensive experiments conducted on two benchmark RS archives demonstrate the effectiveness of the proposed method in accurately retrieving semantically relevant bitemporal images (or text sentences) to a query text sentence (or bitemporal image). The code of this work is publicly available at https://git.tu-berlin.de/rsim/cross-modal-text-tsir.
2501.19045
Trajectory Optimization Under Stochastic Dynamics Leveraging Maximum Mean Discrepancy
cs.RO
This paper addresses sampling-based trajectory optimization for risk-aware navigation under stochastic dynamics. Typically such approaches operate by computing $\tilde{N}$ perturbed rollouts around the nominal dynamics to estimate the collision risk associated with a sequence of control commands. We consider a setting where it is expensive to estimate risk using perturbed rollouts, for example, due to expensive collision-checks. We put forward two key contributions. First, we develop an algorithm that distills the statistical information from a larger set of rollouts to a reduced-set with sample size $N<<\tilde{N}$. Consequently, we estimate collision risk using just $N$ rollouts instead of $\tilde{N}$. Second, we formulate a novel surrogate for the collision risk that can leverage the distilled statistical information contained in the reduced-set. We formalize both algorithmic contributions using distribution embedding in Reproducing Kernel Hilbert Space (RKHS) and Maximum Mean Discrepancy (MMD). We perform extensive benchmarking to demonstrate that our MMD-based approach leads to safer trajectories at low sample regime than existing baselines using Conditional Value-at Risk (CVaR) based collision risk estimate.
2501.19047
Understanding Model Calibration -- A gentle introduction and visual exploration of calibration and the expected calibration error (ECE)
stat.ME cs.AI cs.CV cs.LG stat.ML
To be considered reliable, a model must be calibrated so that its confidence in each decision closely reflects its true outcome. In this blogpost we'll take a look at the most commonly used definition for calibration and then dive into a frequently used evaluation measure for model calibration. We'll then cover some of the drawbacks of this measure and how these surfaced the need for additional notions of calibration, which require their own new evaluation measures. This post is not intended to be an in-depth dissection of all works on calibration, nor does it focus on how to calibrate models. Instead, it is meant to provide a gentle introduction to the different notions and their evaluation measures as well as to re-highlight some issues with a measure that is still widely used to evaluate calibration.
2501.19048
The Role of Graph-based MIL and Interventional Training in the Generalization of WSI Classifiers
eess.IV cs.CV
Whole Slide Imaging (WSI), which involves high-resolution digital scans of pathology slides, has become the gold standard for cancer diagnosis, but its gigapixel resolution and the scarcity of annotated datasets present challenges for deep learning models. Multiple Instance Learning (MIL), a widely-used weakly supervised approach, bypasses the need for patch-level annotations. However, conventional MIL methods overlook the spatial relationships between patches, which are crucial for tasks such as cancer grading and diagnosis. To address this, graph-based approaches have gained prominence by incorporating spatial information through node connections. Despite their potential, both MIL and graph-based models are vulnerable to learning spurious associations, like color variations in WSIs, affecting their robustness. In this dissertation, we conduct an extensive comparison of multiple graph construction techniques, MIL models, graph-MIL approaches, and interventional training, introducing a new framework, Graph-based Multiple Instance Learning with Interventional Training (GMIL-IT), for WSI classification. We evaluate their impact on model generalization through domain shift analysis and demonstrate that graph-based models alone achieve the generalization initially anticipated from interventional training. Our code is available here: github.com/ritamartinspereira/GMIL-IT
2501.19050
Norm-Bounded Low-Rank Adaptation
cs.LG
In this work, we propose norm-bounded low-rank adaptation (NB-LoRA) for parameter-efficient fine tuning. We introduce two parameterizations that allow explicit bounds on each singular value of the weight adaptation matrix, which can therefore satisfy any prescribed unitarily invariant norm bound, including the Schatten norms (e.g., nuclear, Frobenius, spectral norm). The proposed parameterizations are unconstrained and complete, i.e. they cover all matrices satisfying the prescribed rank and norm constraints. Experiments on vision fine-tuning benchmarks show that the proposed approach can achieve good adaptation performance while avoiding model catastrophic forgetting and also substantially improve robustness to a wide range of hyper-parameters, including adaptation rank, learning rate and number of training epochs. We also explore applications in privacy-preserving model merging and low-rank matrix completion.
2501.19054
Text-to-CAD Generation Through Infusing Visual Feedback in Large Language Models
cs.CV cs.LG
Creating Computer-Aided Design (CAD) models requires significant expertise and effort. Text-to-CAD, which converts textual descriptions into CAD parametric sequences, is crucial in streamlining this process. Recent studies have utilized ground-truth parametric sequences, known as sequential signals, as supervision to achieve this goal. However, CAD models are inherently multimodal, comprising parametric sequences and corresponding rendered visual objects. Besides,the rendering process from parametric sequences to visual objects is many-to-one. Therefore, both sequential and visual signals are critical for effective training. In this work, we introduce CADFusion, a framework that uses Large Language Models (LLMs) as the backbone and alternates between two training stages: the sequential learning (SL) stage and the visual feedback (VF) stage. In the SL stage, we train LLMs using ground-truth parametric sequences, enabling the generation of logically coherent parametric sequences. In the VF stage, we reward parametric sequences that render into visually preferred objects and penalize those that do not, allowing LLMs to learn how rendered visual objects are perceived and evaluated. These two stages alternate throughout the training, ensuring balanced learning and preserving benefits of both signals. Experiments demonstrate that CADFusion significantly improves performance, both qualitatively and quantitatively.
2501.19055
Towards Physiologically Sensible Predictions via the Rule-based Reinforcement Learning Layer
cs.LG cs.AI
This paper adds to the growing literature of reinforcement learning (RL) for healthcare by proposing a novel paradigm: augmenting any predictor with Rule-based RL Layer (RRLL) that corrects the model's physiologically impossible predictions. Specifically, RRLL takes as input states predicted labels and outputs corrected labels as actions. The reward of the state-action pair is evaluated by a set of general rules. RRLL is efficient, general and lightweight: it does not require heavy expert knowledge like prior work but only a set of impossible transitions. This set is much smaller than all possible transitions; yet it can effectively reduce physiologically impossible mistakes made by the state-of-the-art predictor models. We verify the utility of RRLL on a variety of important healthcare classification problems and observe significant improvements using the same setup, with only the domain-specific set of impossibility changed. In-depth analysis shows that RRLL indeed improves accuracy by effectively reducing the presence of physiologically impossible predictions.
2501.19056
Enabling Autonomic Microservice Management through Self-Learning Agents
cs.SE cs.AI cs.CL cs.MA
The increasing complexity of modern software systems necessitates robust autonomic self-management capabilities. While Large Language Models (LLMs) demonstrate potential in this domain, they often face challenges in adapting their general knowledge to specific service contexts. To address this limitation, we propose ServiceOdyssey, a self-learning agent system that autonomously manages microservices without requiring prior knowledge of service-specific configurations. By leveraging curriculum learning principles and iterative exploration, ServiceOdyssey progressively develops a deep understanding of operational environments, reducing dependence on human input or static documentation. A prototype built with the Sock Shop microservice demonstrates the potential of this approach for autonomic microservice management.
2501.19057
TeZO: Empowering the Low-Rankness on the Temporal Dimension in the Zeroth-Order Optimization for Fine-tuning LLMs
cs.LG
Zeroth-order optimization (ZO) has demonstrated remarkable promise in efficient fine-tuning tasks for Large Language Models (LLMs). In particular, recent advances incorporate the low-rankness of gradients, introducing low-rank ZO estimators to further reduce GPU memory consumption. However, most existing works focus solely on the low-rankness of each individual gradient, overlooking a broader property shared by all gradients throughout the training, i.e., all gradients approximately reside within a similar subspace. In this paper, we consider two factors together and propose a novel low-rank ZO estimator, TeZO, which captures the low-rankness across both the model and temporal dimension. Specifically, we represent ZO perturbations along the temporal dimension as a 3D tensor and employ Canonical Polyadic Decomposition (CPD) to extract each low-rank 2D matrix, significantly reducing the training cost. TeZO can also be easily extended to the Adam variant while consuming less memory than MeZO-SGD, and requiring about only 35% memory of MeZO-Adam. Both comprehensive theoretical analysis and extensive experimental research have validated its efficiency, achieving SOTA-comparable results with lower overhead of time and memory.
2501.19058
Gravity Compensation of the dVRK-Si Patient Side Manipulator based on Dynamic Model Identification
cs.RO cs.SY eess.SY
The da Vinci Research Kit (dVRK, also known as dVRK Classic) is an open-source teleoperated surgical robotic system whose hardware is obtained from the first generation da Vinci Surgical System (Intuitive, Sunnyvale, CA, USA). The dVRK has greatly facilitated research in robot-assisted surgery over the past decade and helped researchers address multiple major challenges in this domain. Recently, the dVRK-Si system, a new version of the dVRK which uses mechanical components from the da Vinci Si Surgical System, became available to the community. The major difference between the first generation da Vinci and the da Vinci Si is in the structural upgrade of the Patient Side Manipulator (PSM). Because of this upgrade, the gravity of the dVRK-Si PSM can no longer be ignored as in the dVRK Classic. The high gravity offset may lead to relatively low control accuracy and longer response time. In addition, although substantial progress has been made in addressing the dynamic model identification problem for the dVRK Classic, further research is required on model-based control for the dVRK-Si, due to differences in mechanical components and the demand for enhanced control performance. To address these problems, in this work, we present (1) a novel full kinematic model of the dVRK-Si PSM, and (2) a gravity compensation approach based on the dynamic model identification.
2501.19059
Controllable Neural Architectures for Multi-Task Control
eess.SY cs.SY
This paper studies a multi-task control problem where multiple linear systems are to be regulated by a single non-linear controller. In particular, motivated by recent advances in multi-task learning and the design of brain-inspired architectures, we consider a neural controller with (smooth) ReLU activation function. The parameters of the controller are a connectivity matrix and a bias vector: although both parameters can be designed, the connectivity matrix is constant while the bias vector can be varied and is used to adapt the controller across different control tasks. The bias vector determines the equilibrium of the neural controller and, consequently, of its linearized dynamics. Our multi-task control strategy consists of designing the connectivity matrix and a set of bias vectors in a way that the linearized dynamics of the neural controller for the different bias vectors provide a good approximation of a set of desired controllers. We show that, by properly choosing the bias vector, the linearized dynamics of the neural controller can replicate the dynamics of any single, linear controller. Further, we design gradient-based algorithms to train the parameters of the neural controller, and we provide upper and lower bounds for the performance of our neural controller. Finally, we validate our results using different numerical examples.
2501.19060
Contrast-Aware Calibration for Fine-Tuned CLIP: Leveraging Image-Text Alignment
cs.CV cs.LG
Vision-language models (VLMs), such as CLIP, have demonstrated exceptional generalization capabilities and can quickly adapt to downstream tasks through prompt fine-tuning. Unfortunately, in classification tasks involving non-training classes, known as open-vocabulary setting, fine-tuned VLMs often overfit to train classes, resulting in a misalignment between confidence scores and actual accuracy on unseen classes, which significantly undermines their reliability in real-world deployments. Existing confidence calibration methods typically require training parameters or analyzing features from the training dataset, restricting their ability to generalize unseen classes without corresponding train data. Moreover, VLM-specific calibration methods rely solely on text features from train classes as calibration indicators, which inherently limits their ability to calibrate train classes. To address these challenges, we propose an effective multimodal calibration method Contrast-Aware Calibration (CAC). Building on the original CLIP's zero-shot adaptability and the conclusion from empirical analysis that poor intra-class and inter-class discriminative ability on unseen classes is the root cause, we calculate calibration weights based on the contrastive difference between the original and fine-tuned CLIP. This method not only adapts to calibrating unseen classes but also overcomes the limitations of previous VLM calibration methods that could not calibrate train classes. In experiments involving 11 datasets with 5 fine-tuning methods, CAC consistently achieved the best calibration effect on both train and unseen classes without sacrificing accuracy and inference speed.
2501.19061
EgoMe: Follow Me via Egocentric View in Real World
cs.CV
When interacting with the real world, human often take the egocentric (first-person) view as a benchmark, naturally transferring behaviors observed from a exocentric (third-person) view to their own. This cognitive theory provides a foundation for researching how robots can more effectively imitate human behavior. However, current research either employs multiple cameras with different views focusing on the same individual's behavior simultaneously or encounters unpair ego-exo view scenarios, there is no effort to fully exploit human cognitive behavior in the real world. To fill this gap, in this paper, we introduce a novel large-scale egocentric dataset, called EgoMe, which towards following the process of human imitation learning via egocentric view in the real world. Our dataset includes 7902 pairs of videos (15804 videos) for diverse daily behaviors in real-world scenarios. For a pair of videos, one video captures a exocentric view of the imitator observing the demonstrator's actions, while the other captures a egocentric view of the imitator subsequently following those actions. Notably, our dataset also contain exo-ego eye gaze, angular velocity, acceleration, magnetic strength and other sensor multi-modal data for assisting in establishing correlations between observing and following process. In addition, we also propose eight challenging benchmark tasks for fully leveraging this data resource and promoting the research of robot imitation learning ability. Extensive statistical analysis demonstrates significant advantages compared to existing datasets. The proposed EgoMe dataset and benchmark will be released soon.
2501.19063
Optimizing Job Allocation using Reinforcement Learning with Graph Neural Networks
cs.LG
Efficient job allocation in complex scheduling problems poses significant challenges in real-world applications. In this report, we propose a novel approach that leverages the power of Reinforcement Learning (RL) and Graph Neural Networks (GNNs) to tackle the Job Allocation Problem (JAP). The JAP involves allocating a maximum set of jobs to available resources while considering several constraints. Our approach enables learning of adaptive policies through trial-and-error interactions with the environment while exploiting the graph-structured data of the problem. By leveraging RL, we eliminate the need for manual annotation, a major bottleneck in supervised learning approaches. Experimental evaluations on synthetic and real-world data demonstrate the effectiveness and generalizability of our proposed approach, outperforming baseline algorithms and showcasing its potential for optimizing job allocation in complex scheduling problems.
2501.19064
Machine Learning in Gamma Astronomy
astro-ph.IM astro-ph.HE cs.LG
The purpose of this paper is to review the most popular deep learning methods used to analyze astroparticle data obtained with Imaging Atmospheric Cherenkov Telescopes and provide references to the original papers.
2501.19065
BEAT: Balanced Frequency Adaptive Tuning for Long-Term Time-Series Forecasting
cs.LG cs.AI
Time-series forecasting is crucial for numerous real-world applications including weather prediction and financial market modeling. While temporal-domain methods remain prevalent, frequency-domain approaches can effectively capture multi-scale periodic patterns, reduce sequence dependencies, and naturally denoise signals. However, existing approaches typically train model components for all frequencies under a unified training objective, often leading to mismatched learning speeds: high-frequency components converge faster and risk overfitting, while low-frequency components underfit due to insufficient training time. To deal with this challenge, we propose BEAT (Balanced frEquency Adaptive Tuning), a novel framework that dynamically monitors the training status for each frequency and adaptively adjusts their gradient updates. By recognizing convergence, overfitting, or underfitting for each frequency, BEAT dynamically reallocates learning priorities, moderating gradients for rapid learners and increasing those for slower ones, alleviating the tension between competing objectives across frequencies and synchronizing the overall learning process. Extensive experiments on seven real-world datasets demonstrate that BEAT consistently outperforms state-of-the-art approaches.
2501.19066
Concept Steerers: Leveraging K-Sparse Autoencoders for Controllable Generations
cs.CV
Despite the remarkable progress in text-to-image generative models, they are prone to adversarial attacks and inadvertently generate unsafe, unethical content. Existing approaches often rely on fine-tuning models to remove specific concepts, which is computationally expensive, lack scalability, and/or compromise generation quality. In this work, we propose a novel framework leveraging k-sparse autoencoders (k-SAEs) to enable efficient and interpretable concept manipulation in diffusion models. Specifically, we first identify interpretable monosemantic concepts in the latent space of text embeddings and leverage them to precisely steer the generation away or towards a given concept (e.g., nudity) or to introduce a new concept (e.g., photographic style). Through extensive experiments, we demonstrate that our approach is very simple, requires no retraining of the base model nor LoRA adapters, does not compromise the generation quality, and is robust to adversarial prompt manipulations. Our method yields an improvement of $\mathbf{20.01\%}$ in unsafe concept removal, is effective in style manipulation, and is $\mathbf{\sim5}$x faster than current state-of-the-art.
2501.19067
Deep Multi-Task Learning Has Low Amortized Intrinsic Dimensionality
cs.LG stat.ML
Deep learning methods are known to generalize well from training to future data, even in an overparametrized regime, where they could easily overfit. One explanation for this phenomenon is that even when their *ambient dimensionality*, (i.e. the number of parameters) is large, the models' *intrinsic dimensionality* is small, i.e. their learning takes place in a small subspace of all possible weight configurations. In this work, we confirm this phenomenon in the setting of *deep multi-task learning*. We introduce a method to parametrize multi-task network directly in the low-dimensional space, facilitated by the use of *random expansions* techniques. We then show that high-accuracy multi-task solutions can be found with much smaller intrinsic dimensionality (fewer free parameters) than what single-task learning requires. Subsequently, we show that the low-dimensional representations in combination with *weight compression* and *PAC-Bayesian* reasoning lead to the first *non-vacuous generalization bounds* for deep multi-task networks.
2501.19069
Improving vision-language alignment with graph spiking hybrid Networks
cs.CV cs.AI
To bridge the semantic gap between vision and language (VL), it is necessary to develop a good alignment strategy, which includes handling semantic diversity, abstract representation of visual information, and generalization ability of models. Recent works use detector-based bounding boxes or patches with regular partitions to represent visual semantics. While current paradigms have made strides, they are still insufficient for fully capturing the nuanced contextual relations among various objects. This paper proposes a comprehensive visual semantic representation module, necessitating the utilization of panoptic segmentation to generate coherent fine-grained semantic features. Furthermore, we propose a novel Graph Spiking Hybrid Network (GSHN) that integrates the complementary advantages of Spiking Neural Networks (SNNs) and Graph Attention Networks (GATs) to encode visual semantic information. Intriguingly, the model not only encodes the discrete and continuous latent variables of instances but also adeptly captures both local and global contextual features, thereby significantly enhancing the richness and diversity of semantic representations. Leveraging the spatiotemporal properties inherent in SNNs, we employ contrastive learning (CL) to enhance the similarity-based representation of embeddings. This strategy alleviates the computational overhead of the model and enriches meaningful visual representations by constructing positive and negative sample pairs. We design an innovative pre-training method, Spiked Text Learning (STL), which uses text features to improve the encoding ability of discrete semantics. Experiments show that the proposed GSHN exhibits promising results on multiple VL downstream tasks.
2501.19072
SpikingSoft: A Spiking Neuron Controller for Bio-inspired Locomotion with Soft Snake Robots
cs.RO cs.LG
Inspired by the dynamic coupling of moto-neurons and physical elasticity in animals, this work explores the possibility of generating locomotion gaits by utilizing physical oscillations in a soft snake by means of a low-level spiking neural mechanism. To achieve this goal, we introduce the Double Threshold Spiking neuron model with adjustable thresholds to generate varied output patterns. This neuron model can excite the natural dynamics of soft robotic snakes, and it enables distinct movements, such as turning or moving forward, by simply altering the neural thresholds. Finally, we demonstrate that our approach, termed SpikingSoft, naturally pairs and integrates with reinforcement learning. The high-level agent only needs to adjust the two thresholds to generate complex movement patterns, thus strongly simplifying the learning of reactive locomotion. Simulation results demonstrate that the proposed architecture significantly enhances the performance of the soft snake robot, enabling it to achieve target objectives with a 21.6% increase in success rate, a 29% reduction in time to reach the target, and smoother movements compared to the vanilla reinforcement learning controllers or Central Pattern Generator controller acting in torque space.
2501.19073
Pareto-frontier Entropy Search with Variational Lower Bound Maximization
cs.LG stat.ML
This study considers multi-objective Bayesian optimization (MOBO) through the information gain of the Pareto-frontier. To calculate the information gain, a predictive distribution conditioned on the Pareto-frontier plays a key role, which is defined as a distribution truncated by the Pareto-frontier. However, it is usually impossible to obtain the entire Pareto-frontier in a continuous domain, and therefore, the complete truncation cannot be known. We consider an approximation of the truncate distribution by using a mixture distribution consisting of two possible approximate truncation obtainable from a subset of the Pareto-frontier, which we call over- and under-truncation. Since the optimal balance of the mixture is unknown beforehand, we propose optimizing the balancing coefficient through the variational lower bound maximization framework, by which the approximation error of the information gain can be minimized. Our empirical evaluation demonstrates the effectiveness of the proposed method particularly when the number of objective functions is large.
2501.19077
Temperature-Annealed Boltzmann Generators
cs.LG
Efficient sampling of unnormalized probability densities such as the Boltzmann distribution of molecular systems is a longstanding challenge. Next to conventional approaches like molecular dynamics or Markov chain Monte Carlo, variational approaches, such as training normalizing flows with the reverse Kullback-Leibler divergence, have been introduced. However, such methods are prone to mode collapse and often do not learn to sample the full configurational space. Here, we present temperature-annealed Boltzmann generators (TA-BG) to address this challenge. First, we demonstrate that training a normalizing flow with the reverse Kullback-Leibler divergence at high temperatures is possible without mode collapse. Furthermore, we introduce a reweighting-based training objective to anneal the distribution to lower target temperatures. We apply this methodology to three molecular systems of increasing complexity and, compared to the baseline, achieve better results in almost all metrics while requiring up to three times fewer target energy evaluations. For the largest system, our approach is the only method that accurately resolves the metastable states of the system.
2501.19080
Differentially Private Policy Gradient
cs.LG
Motivated by the increasing deployment of reinforcement learning in the real world, involving a large consumption of personal data, we introduce a differentially private (DP) policy gradient algorithm. We show that, in this setting, the introduction of Differential Privacy can be reduced to the computation of appropriate trust regions, thus avoiding the sacrifice of theoretical properties of the DP-less methods. Therefore, we show that it is possible to find the right trade-off between privacy noise and trust-region size to obtain a performant differentially private policy gradient algorithm. We then outline its performance empirically on various benchmarks. Our results and the complexity of the tasks addressed represent a significant improvement over existing DP algorithms in online RL.