id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2501.19082
A Bias-Correction Decentralized Stochastic Gradient Algorithm with Momentum Acceleration
cs.LG cs.DC math.OC stat.ML
Distributed stochastic optimization algorithms can simultaneously process large-scale datasets, significantly accelerating model training. However, their effectiveness is often hindered by the sparsity of distributed networks and data heterogeneity. In this paper, we propose a momentum-accelerated distributed stochastic gradient algorithm, termed Exact-Diffusion with Momentum (EDM), which mitigates the bias from data heterogeneity and incorporates momentum techniques commonly used in deep learning to enhance convergence rate. Our theoretical analysis demonstrates that the EDM algorithm converges sub-linearly to the neighborhood of the optimal solution, the radius of which is irrespective of data heterogeneity, when applied to non-convex objective functions; under the Polyak-Lojasiewicz condition, which is a weaker assumption than strong convexity, it converges linearly to the target region. Our analysis techniques employed to handle momentum in complex distributed parameter update structures yield a sufficiently tight convergence upper bound, offering a new perspective for the theoretical analysis of other momentum-based distributed algorithms.
2501.19083
MotionPCM: Real-Time Motion Synthesis with Phased Consistency Model
cs.CV
Diffusion models have become a popular choice for human motion synthesis due to their powerful generative capabilities. However, their high computational complexity and large sampling steps pose challenges for real-time applications. Fortunately, the Consistency Model (CM) provides a solution to greatly reduce the number of sampling steps from hundreds to a few, typically fewer than four, significantly accelerating the synthesis of diffusion models. However, its application to text-conditioned human motion synthesis in latent space remains challenging. In this paper, we introduce \textbf{MotionPCM}, a phased consistency model-based approach designed to improve the quality and efficiency of real-time motion synthesis in latent space.
2501.19084
Laser: Efficient Language-Guided Segmentation in Neural Radiance Fields
cs.CV
In this work, we propose a method that leverages CLIP feature distillation, achieving efficient 3D segmentation through language guidance. Unlike previous methods that rely on multi-scale CLIP features and are limited by processing speed and storage requirements, our approach aims to streamline the workflow by directly and effectively distilling dense CLIP features, thereby achieving precise segmentation of 3D scenes using text. To achieve this, we introduce an adapter module and mitigate the noise issue in the dense CLIP feature distillation process through a self-cross-training strategy. Moreover, to enhance the accuracy of segmentation edges, this work presents a low-rank transient query attention mechanism. To ensure the consistency of segmentation for similar colors under different viewpoints, we convert the segmentation task into a classification task through label volume, which significantly improves the consistency of segmentation in color-similar areas. We also propose a simplified text augmentation strategy to alleviate the issue of ambiguity in the correspondence between CLIP features and text. Extensive experimental results show that our method surpasses current state-of-the-art technologies in both training speed and performance. Our code is available on: https://github.com/xingy038/Laser.git.
2501.19086
Fairness Analysis of CLIP-Based Foundation Models for X-Ray Image Classification
cs.CV cs.AI
X-ray imaging is pivotal in medical diagnostics, offering non-invasive insights into a range of health conditions. Recently, vision-language models, such as the Contrastive Language-Image Pretraining (CLIP) model, have demonstrated potential in improving diagnostic accuracy by leveraging large-scale image-text datasets. However, since CLIP was not initially designed for medical images, several CLIP-like models trained specifically on medical images have been developed. Despite their enhanced performance, issues of fairness - particularly regarding demographic attributes - remain largely unaddressed. In this study, we perform a comprehensive fairness analysis of CLIP-like models applied to X-ray image classification. We assess their performance and fairness across diverse patient demographics and disease categories using zero-shot inference and various fine-tuning techniques, including Linear Probing, Multilayer Perceptron (MLP), Low-Rank Adaptation (LoRA), and full fine-tuning. Our results indicate that while fine-tuning improves model accuracy, fairness concerns persist, highlighting the need for further fairness interventions in these foundational models.
2501.19088
JGHand: Joint-Driven Animatable Hand Avater via 3D Gaussian Splatting
cs.CV
Since hands are the primary interface in daily interactions, modeling high-quality digital human hands and rendering realistic images is a critical research problem. Furthermore, considering the requirements of interactive and rendering applications, it is essential to achieve real-time rendering and driveability of the digital model without compromising rendering quality. Thus, we propose Jointly 3D Gaussian Hand (JGHand), a novel joint-driven 3D Gaussian Splatting (3DGS)-based hand representation that renders high-fidelity hand images in real-time for various poses and characters. Distinct from existing articulated neural rendering techniques, we introduce a differentiable process for spatial transformations based on 3D key points. This process supports deformations from the canonical template to a mesh with arbitrary bone lengths and poses. Additionally, we propose a real-time shadow simulation method based on per-pixel depth to simulate self-occlusion shadows caused by finger movements. Finally, we embed the hand prior and propose an animatable 3DGS representation of the hand driven solely by 3D key points. We validate the effectiveness of each component of our approach through comprehensive ablation studies. Experimental results on public datasets demonstrate that JGHand achieves real-time rendering speeds with enhanced quality, surpassing state-of-the-art methods.
2501.19089
Understanding Oversmoothing in GNNs as Consensus in Opinion Dynamics
cs.LG
In contrast to classes of neural networks where the learned representations become increasingly expressive with network depth, the learned representations in graph neural networks (GNNs), tend to become increasingly similar. This phenomena, known as oversmoothing, is characterized by learned representations that cannot be reliably differentiated leading to reduced predictive performance. In this paper, we propose an analogy between oversmoothing in GNNs and consensus or agreement in opinion dynamics. Through this analogy, we show that the message passing structure of recent continuous-depth GNNs is equivalent to a special case of opinion dynamics (i.e., linear consensus models) which has been theoretically proven to converge to consensus (i.e., oversmoothing) for all inputs. Using the understanding developed through this analogy, we design a new continuous-depth GNN model based on nonlinear opinion dynamics and prove that our model, which we call behavior-inspired message passing neural network (BIMP) circumvents oversmoothing for general inputs. Through extensive experiments, we show that BIMP is robust to oversmoothing and adversarial attack, and consistently outperforms competitive baselines on numerous benchmarks.
2501.19090
Pivoting Factorization: A Compact Meta Low-Rank Representation of Sparsity for Efficient Inference in Large Language Models
cs.LG
The rapid growth of Large Language Models has driven demand for effective model compression techniques to reduce memory and computation costs. Low-rank pruning has gained attention for its tensor coherence and GPU compatibility across all densities. However, low-rank pruning has struggled to match the performance of semi-structured pruning, often doubling perplexity (PPL) at similar densities. In this paper, we propose Pivoting Factorization (PIFA), a novel lossless meta low-rank representation that unsupervisedly learns a compact form of any low-rank representation, effectively eliminating redundant information. PIFA identifies pivot rows (linearly independent rows) and expresses non-pivot rows as linear combinations, achieving an additional 24.2\% memory savings and 24.6\% faster inference over low-rank layers at r/d = 0.5, thereby significantly enhancing performance at the same density. To mitigate the performance degradation caused by low-rank pruning, we introduce a novel, retraining-free low-rank reconstruction method that minimizes error accumulation (M). MPIFA, combining M and PIFA into an end-to-end framework, significantly outperforms existing low-rank pruning methods and, for the first time, achieves performance comparable to semi-structured pruning, while surpassing it in GPU efficiency and compatibility.
2501.19091
FL-APU: A Software Architecture to Ease Practical Implementation of Cross-Silo Federated Learning
cs.DC cs.LG
Federated Learning (FL) is an upcoming technology that is increasingly applied in real-world applications. Early applications focused on cross-device scenarios, where many participants with limited resources train machine learning (ML) models together, e.g., in the case of Google's GBoard. Contrarily, cross-silo scenarios have only few participants but with many resources, e.g., in the healthcare domain. Despite such early efforts, FL is still rarely used in practice and best practices are, hence, missing. For new applications, in our case inter-organizational cross-silo applications, overcoming this lack of role models is a significant challenge. In order to ease the use of FL in real-world cross-silo applications, we here propose a scenario-based architecture for the practical use of FL in the context of multiple companies collaborating to improve the quality of their ML models. The architecture emphasizes the collaboration between the participants and the FL server and extends basic interactions with domain-specific features. First, it combines governance with authentication, creating an environment where only trusted participants can join. Second, it offers traceability of governance decisions and tracking of training processes, which are also crucial in a production environment. Beyond presenting the architectural design, we analyze requirements for the real-world use of FL and evaluate the architecture with a scenario-based analysis method.
2501.19093
Improving Low-Resource Sequence Labeling with Knowledge Fusion and Contextual Label Explanations
cs.CL
Sequence labeling remains a significant challenge in low-resource, domain-specific scenarios, particularly for character-dense languages like Chinese. Existing methods primarily focus on enhancing model comprehension and improving data diversity to boost performance. However, these approaches still struggle with inadequate model applicability and semantic distribution biases in domain-specific contexts. To overcome these limitations, we propose a novel framework that combines an LLM-based knowledge enhancement workflow with a span-based Knowledge Fusion for Rich and Efficient Extraction (KnowFREE) model. Our workflow employs explanation prompts to generate precise contextual interpretations of target entities, effectively mitigating semantic biases and enriching the model's contextual understanding. The KnowFREE model further integrates extension label features, enabling efficient nested entity extraction without relying on external knowledge during inference. Experiments on multiple Chinese domain-specific sequence labeling datasets demonstrate that our approach achieves state-of-the-art performance, effectively addressing the challenges posed by low-resource settings.
2501.19094
Ambient Denoising Diffusion Generative Adversarial Networks for Establishing Stochastic Object Models from Noisy Image Data
cs.CV eess.IV
It is widely accepted that medical imaging systems should be objectively assessed via task-based image quality (IQ) measures that ideally account for all sources of randomness in the measured image data, including the variation in the ensemble of objects to be imaged. Stochastic object models (SOMs) that can randomly draw samples from the object distribution can be employed to characterize object variability. To establish realistic SOMs for task-based IQ analysis, it is desirable to employ experimental image data. However, experimental image data acquired from medical imaging systems are subject to measurement noise. Previous work investigated the ability of deep generative models (DGMs) that employ an augmented generative adversarial network (GAN), AmbientGAN, for establishing SOMs from noisy measured image data. Recently, denoising diffusion models (DDMs) have emerged as a leading DGM for image synthesis and can produce superior image quality than GANs. However, original DDMs possess a slow image-generation process because of the Gaussian assumption in the denoising steps. More recently, denoising diffusion GAN (DDGAN) was proposed to permit fast image generation while maintain high generated image quality that is comparable to the original DDMs. In this work, we propose an augmented DDGAN architecture, Ambient DDGAN (ADDGAN), for learning SOMs from noisy image data. Numerical studies that consider clinical computed tomography (CT) images and digital breast tomosynthesis (DBT) images are conducted. The ability of the proposed ADDGAN to learn realistic SOMs from noisy image data is demonstrated. It has been shown that the ADDGAN significantly outperforms the advanced AmbientGAN models for synthesizing high resolution medical images with complex textures.
2501.19095
PathE: Leveraging Entity-Agnostic Paths for Parameter-Efficient Knowledge Graph Embeddings
cs.AI cs.LG
Knowledge Graphs (KGs) store human knowledge in the form of entities (nodes) and relations, and are used extensively in various applications. KG embeddings are an effective approach to addressing tasks like knowledge discovery, link prediction, and reasoning. This is often done by allocating and learning embedding tables for all or a subset of the entities. As this scales linearly with the number of entities, learning embedding models in real-world KGs with millions of nodes can be computationally intractable. To address this scalability problem, our model, PathE, only allocates embedding tables for relations (which are typically orders of magnitude fewer than the entities) and requires less than 25% of the parameters of previous parameter efficient methods. Rather than storing entity embeddings, we learn to compute them by leveraging multiple entity-relation paths to contextualise individual entities within triples. Evaluated on four benchmarks, PathE achieves state-of-the-art performance in relation prediction, and remains competitive in link prediction on path-rich KGs while training on consumer-grade hardware. We perform ablation experiments to test our design choices and analyse the sensitivity of the model to key hyper-parameters. PathE is efficient and cost-effective for relationally diverse and well-connected KGs commonly found in real-world applications.
2501.19098
$\infty$-Video: A Training-Free Approach to Long Video Understanding via Continuous-Time Memory Consolidation
cs.CV cs.LG
Current video-language models struggle with long-video understanding due to limited context lengths and reliance on sparse frame subsampling, often leading to information loss. This paper introduces $\infty$-Video, which can process arbitrarily long videos through a continuous-time long-term memory (LTM) consolidation mechanism. Our framework augments video Q-formers by allowing them to process unbounded video contexts efficiently and without requiring additional training. Through continuous attention, our approach dynamically allocates higher granularity to the most relevant video segments, forming "sticky" memories that evolve over time. Experiments with Video-LLaMA and VideoChat2 demonstrate improved performance in video question-answering tasks, showcasing the potential of continuous-time LTM mechanisms to enable scalable and training-free comprehension of long videos.
2501.19099
Unraveling Zeroth-Order Optimization through the Lens of Low-Dimensional Structured Perturbations
cs.LG
Zeroth-order (ZO) optimization has emerged as a promising alternative to gradient-based backpropagation methods, particularly for black-box optimization and large language model (LLM) fine-tuning. However, ZO methods suffer from slow convergence due to high-variance stochastic gradient estimators. While structured perturbations, such as sparsity and low-rank constraints, have been explored to mitigate these issues, their effectiveness remains highly under-explored. In this work, we develop a unified theoretical framework that analyzes both the convergence and generalization properties of ZO optimization under structured perturbations. We show that high dimensionality is the primary bottleneck and introduce the notions of \textit{stable rank} and \textit{effective overlap} to explain how structured perturbations reduce gradient noise and accelerate convergence. Using the uniform stability under our framework, we then provide the first theoretical justification for why these perturbations enhance generalization. Additionally, through empirical analysis, we identify that \textbf{block coordinate descent} (BCD) to be an effective structured perturbation method. Extensive experiments show that, compared to existing alternatives, memory-efficient ZO (MeZO) with BCD (\textit{MeZO-BCD}) can provide improved converge with a faster wall-clock time/iteration by up to $\times\textbf{2.09}$ while yielding similar or better accuracy.
2501.19102
Reinforcement Learning on Reconfigurable Hardware: Overcoming Material Variability in Laser Material Processing
cs.LG
Ensuring consistent processing quality is challenging in laser processes due to varying material properties and surface conditions. Although some approaches have shown promise in solving this problem via automation, they often rely on predetermined targets or are limited to simulated environments. To address these shortcomings, we propose a novel real-time reinforcement learning approach for laser process control, implemented on a Field Programmable Gate Array to achieve real-time execution. Our experimental results from laser welding tests on stainless steel samples with a range of surface roughnesses validated the method's ability to adapt autonomously, without relying on reward engineering or prior setup information. Specifically, the algorithm learned the correct power profile for each unique surface characteristic, demonstrating significant improvements over hand-engineered optimal constant power strategies -- up to 23% better performance on rougher surfaces and 7% on mixed surfaces. This approach represents a significant advancement in automating and optimizing laser processes, with potential applications across multiple industries.
2501.19104
Neural Collapse Beyond the Unconstrained Features Model: Landscape, Dynamics, and Generalization in the Mean-Field Regime
cs.LG
Neural Collapse is a phenomenon where the last-layer representations of a well-trained neural network converge to a highly structured geometry. In this paper, we focus on its first (and most basic) property, known as NC1: the within-class variability vanishes. While prior theoretical studies establish the occurrence of NC1 via the data-agnostic unconstrained features model, our work adopts a data-specific perspective, analyzing NC1 in a three-layer neural network, with the first two layers operating in the mean-field regime and followed by a linear layer. In particular, we establish a fundamental connection between NC1 and the loss landscape: we prove that points with small empirical loss and gradient norm (thus, close to being stationary) approximately satisfy NC1, and the closeness to NC1 is controlled by the residual loss and gradient norm. We then show that (i) gradient flow on the mean squared error converges to NC1 solutions with small empirical loss, and (ii) for well-separated data distributions, both NC1 and vanishing test loss are achieved simultaneously. This aligns with the empirical observation that NC1 emerges during training while models attain near-zero test error. Overall, our results demonstrate that NC1 arises from gradient training due to the properties of the loss landscape, and they show the co-occurrence of NC1 and small test error for certain data distributions.
2501.19105
Relating Misfit to Gain in Weak-to-Strong Generalization Beyond the Squared Loss
cs.LG math.PR
The paradigm of weak-to-strong generalization constitutes the training of a strong AI model on data labeled by a weak AI model, with the goal that the strong model nevertheless outperforms its weak supervisor on the target task of interest. For the setting of real-valued regression with the squared loss, recent work quantitatively characterizes the gain in performance of the strong model over the weak model in terms of the misfit between the strong and weak model. We generalize such a characterization to learning tasks whose loss functions correspond to arbitrary Bregman divergences when the strong class is convex. This extends the misfit-based characterization of performance gain in weak-to-strong generalization to classification tasks, as the cross-entropy loss can be expressed in terms of a Bregman divergence. In most practical scenarios, however, the strong model class may not be convex. We therefore weaken this assumption and study weak-to-strong generalization for convex combinations of $k$ strong models in the strong class, in the concrete setting of classification. This allows us to obtain a similar misfit-based characterization of performance gain, upto an additional error term that vanishes as $k$ gets large. Our theoretical findings are supported by thorough experiments on synthetic as well as real-world datasets.
2501.19107
Brain-inspired sparse training enables Transformers and LLMs to perform as fully connected
cs.LG
This study aims to enlarge our current knowledge on application of brain-inspired network science principles for training artificial neural networks (ANNs) with sparse connectivity. Dynamic sparse training (DST) can reduce the computational demands in ANNs, but faces difficulties to keep peak performance at high sparsity levels. The Cannistraci-Hebb training (CHT) is a brain-inspired method for growing connectivity in DST. CHT leverages a gradient-free, topology-driven link regrowth, which has shown ultra-sparse (1% connectivity or lower) advantage across various tasks compared to fully connected networks. Yet, CHT suffers two main drawbacks: (i) its time complexity is O(Nd^3) - N node network size, d node degree - hence it can apply only to ultra-sparse networks. (ii) it selects top link prediction scores, which is inappropriate for the early training epochs, when the network presents unreliable connections. We propose a GPU-friendly approximation of the CH link predictor, which reduces the computational complexity to O(N^3), enabling a fast implementation of CHT in large-scale models. We introduce the Cannistraci-Hebb training soft rule (CHTs), which adopts a strategy for sampling connections in both link removal and regrowth, balancing the exploration and exploitation of network topology. To improve performance, we integrate CHTs with a sigmoid gradual density decay (CHTss). Empirical results show that, using 1% of connections, CHTs outperforms fully connected networks in MLP on visual classification tasks, compressing some networks to < 30% nodes. Using 5% of the connections, CHTss outperforms fully connected networks in two Transformer-based machine translation tasks. Using 30% of the connections, CHTss achieves superior performance compared to other dynamic sparse training methods in language modeling, and it surpasses the fully connected counterpart in zero-shot evaluations.
2501.19111
A Benchmark for Incremental Micro-expression Recognition
cs.CV cs.AI
Micro-expression recognition plays a pivotal role in understanding hidden emotions and has applications across various fields. Traditional recognition methods assume access to all training data at once, but real-world scenarios involve continuously evolving data streams. To respond to the requirement of adapting to new data while retaining previously learned knowledge, we introduce the first benchmark specifically designed for incremental micro-expression recognition. Our contributions include: Firstly, we formulate the incremental learning setting tailored for micro-expression recognition. Secondly, we organize sequential datasets with carefully curated learning orders to reflect real-world scenarios. Thirdly, we define two cross-evaluation-based testing protocols, each targeting distinct evaluation objectives. Finally, we provide six baseline methods and their corresponding evaluation results. This benchmark lays the groundwork for advancing incremental micro-expression recognition research. All source code used in this study will be publicly available at https://github.com/ZhengQinLai/IMER-benchmark.
2501.19112
Logical Modalities within the European AI Act: An Analysis
cs.AI cs.CY cs.LO
The paper presents a comprehensive analysis of the European AI Act in terms of its logical modalities, with the aim of preparing its formal representation, for example, within the logic-pluralistic Knowledge Engineering Framework and Methodology (LogiKEy). LogiKEy develops computational tools for normative reasoning based on formal methods, employing Higher-Order Logic (HOL) as a unifying meta-logic to integrate diverse logics through shallow semantic embeddings. This integration is facilitated by Isabelle/HOL, a proof assistant tool equipped with several automated theorem provers. The modalities within the AI Act and the logics suitable for their representation are discussed. For a selection of these logics, embeddings in HOL are created, which are then used to encode sample paragraphs. Initial experiments evaluate the suitability of these embeddings for automated reasoning, and highlight key challenges on the way to more robust reasoning capabilities.
2501.19113
Genetic AI: Evolutionary Simulation for Data Analysis
cs.NE
We introduce Genetic AI, a novel method for data analysis by evolutionary simulations. The method can be applied to data of any domain and allows for a data-less training of AI models. Without employing predefined rules or training data, Genetic AI first converts the input data into genes and organisms. In a simulation from first principles, these genes and organisms compete for fitness, where their behavior is governed by universal evolutionary strategies. Investigating evolutionary stable equilibriums, Genetic AI helps understanding correlations and symmetries in general input data. Several numerical experiments demonstrate the dynamics of exemplary systems.
2501.19114
Principal Components for Neural Network Initialization
cs.LG cs.AI
Principal Component Analysis (PCA) is a commonly used tool for dimension reduction and denoising. Therefore, it is also widely used on the data prior to training a neural network. However, this approach can complicate the explanation of explainable AI (XAI) methods for the decision of the model. In this work, we analyze the potential issues with this approach and propose Principal Components-based Initialization (PCsInit), a strategy to incorporate PCA into the first layer of a neural network via initialization of the first layer in the network with the principal components, and its two variants PCsInit-Act and PCsInit-Sub. Explanations using these strategies are as direct and straightforward as for neural networks and are simpler than using PCA prior to training a neural network on the principal components. Moreover, as will be illustrated in the experiments, such training strategies can also allow further improvement of training via backpropagation.
2501.19116
A Theoretical Justification for Asymmetric Actor-Critic Algorithms
cs.LG stat.ML
In reinforcement learning for partially observable environments, many successful algorithms were developed within the asymmetric learning paradigm. This paradigm leverages additional state information available at training time for faster learning. Although the proposed learning objectives are usually theoretically sound, these methods still lack a theoretical justification for their potential benefits. We propose such a justification for asymmetric actor-critic algorithms with linear function approximators by adapting a finite-time convergence analysis to this setting. The resulting finite-time bound reveals that the asymmetric critic eliminates an error term arising from aliasing in the agent state.
2501.19122
FedRTS: Federated Robust Pruning via Combinatorial Thompson Sampling
cs.LG cs.AI
Federated Learning (FL) enables collaborative model training across distributed clients without data sharing, but its high computational and communication demands strain resource-constrained devices. While existing methods use dynamic pruning to improve efficiency by periodically adjusting sparse model topologies while maintaining sparsity, these approaches suffer from issues such as greedy adjustments, unstable topologies, and communication inefficiency, resulting in less robust models and suboptimal performance under data heterogeneity and partial client availability. To address these challenges, we propose Federated Robust pruning via combinatorial Thompson Sampling (FedRTS), a novel framework designed to develop robust sparse models. FedRTS enhances robustness and performance through its Thompson Sampling-based Adjustment (TSAdj) mechanism, which uses probabilistic decisions informed by stable, farsighted information instead of deterministic decisions reliant on unstable and myopic information in previous methods. Extensive experiments demonstrate that FedRTS achieves state-of-the-art performance in computer vision and natural language processing tasks while reducing communication costs, particularly excelling in scenarios with heterogeneous data distributions and partial client participation. Our codes are available at: https://github.com/Little0o0/FedRTS
2501.19125
Upper Bounds on the Minimum Distance of Structured LDPC Codes
cs.IT math.IT
We investigate the minimum distance of structured binary Low-Density Parity-Check (LDPC) codes whose parity-check matrices are of the form $[\mathbf{C} \vert \mathbf{M}]$ where $\mathbf{C}$ is circulant and of column weight $2$, and $\mathbf{M}$ has fixed column weight $r \geq 3$ and row weight at least $1$. These codes are of interest because they are LDPC codes which come with a natural linear-time encoding algorithm. We show that the minimum distance of these codes is in $O(n^{\frac{r-2}{r-1} + \epsilon})$, where $n$ is the code length and $\epsilon > 0$ is arbitrarily small. This improves the previously known upper bound in $O(n^{\frac{r-1}{r}})$ on the minimum distance of such codes.
2501.19128
Shaping Sparse Rewards in Reinforcement Learning: A Semi-supervised Approach
cs.LG cs.AI
In many real-world scenarios, reward signal for agents are exceedingly sparse, making it challenging to learn an effective reward function for reward shaping. To address this issue, our approach performs reward shaping not only by utilizing non-zero-reward transitions but also by employing the Semi-Supervised Learning (SSL) technique combined with a novel data augmentation to learn trajectory space representations from the majority of transitions, zero-reward transitions, thereby improving the efficacy of reward shaping. Experimental results in Atari and robotic manipulation demonstrate that our method effectively generalizes reward shaping to sparse reward scenarios, achieving up to four times better performance in reaching higher best scores compared to curiosity-driven methods. The proposed double entropy data augmentation enhances performance, showcasing a 15.8\% increase in best score over other augmentation methods.
2501.19129
RGB-Event ISP: The Dataset and Benchmark
cs.CV eess.IV
Event-guided imaging has received significant attention due to its potential to revolutionize instant imaging systems. However, the prior methods primarily focus on enhancing RGB images in a post-processing manner, neglecting the challenges of image signal processor (ISP) dealing with event sensor and the benefits events provide for reforming the ISP process. To achieve this, we conduct the first research on event-guided ISP. First, we present a new event-RAW paired dataset, collected with a novel but still confidential sensor that records pixel-level aligned events and RAW images. This dataset includes 3373 RAW images with 2248 x 3264 resolution and their corresponding events, spanning 24 scenes with 3 exposure modes and 3 lenses. Second, we propose a conventional ISP pipeline to generate good RGB frames as reference. This conventional ISP pipleline performs basic ISP operations, e.g.demosaicing, white balancing, denoising and color space transforming, with a ColorChecker as reference. Third, we classify the existing learnable ISP methods into 3 classes, and select multiple methods to train and evaluate on our new dataset. Lastly, since there is no prior work for reference, we propose a simple event-guided ISP method and test it on our dataset. We further put forward key technical challenges and future directions in RGB-Event ISP. In summary, to the best of our knowledge, this is the very first research focusing on event-guided ISP, and we hope it will inspire the community. The code and dataset are available at: https://github.com/yunfanLu/RGB-Event-ISP.
2501.19133
Decorrelated Soft Actor-Critic for Efficient Deep Reinforcement Learning
cs.LG cs.AI
The effectiveness of credit assignment in reinforcement learning (RL) when dealing with high-dimensional data is influenced by the success of representation learning via deep neural networks, and has implications for the sample efficiency of deep RL algorithms. Input decorrelation has been previously introduced as a method to speed up optimization in neural networks, and has proven impactful in both efficient deep learning and as a method for effective representation learning for deep RL algorithms. We propose a novel approach to online decorrelation in deep RL based on the decorrelated backpropagation algorithm that seamlessly integrates the decorrelation process into the RL training pipeline. Decorrelation matrices are added to each layer, which are updated using a separate decorrelation learning rule that minimizes the total decorrelation loss across all layers, in parallel to minimizing the usual RL loss. We used our approach in combination with the soft actor-critic (SAC) method, which we refer to as decorrelated soft actor-critic (DSAC). Experiments on the Atari 100k benchmark with DSAC shows, compared to the regular SAC baseline, faster training in five out of the seven games tested and improved reward performance in two games with around 50% reduction in wall-clock time, while maintaining performance levels on the other games. These results demonstrate the positive impact of network-wide decorrelation in deep RL for speeding up its sample efficiency through more effective credit assignment.
2501.19134
Mixed Feelings: Cross-Domain Sentiment Classification of Patient Feedback
cs.CL
Sentiment analysis of patient feedback from the public health domain can aid decision makers in evaluating the provided services. The current paper focuses on free-text comments in patient surveys about general practitioners and psychiatric healthcare, annotated with four sentence-level polarity classes -- positive, negative, mixed and neutral -- while also attempting to alleviate data scarcity by leveraging general-domain sources in the form of reviews. For several different architectures, we compare in-domain and out-of-domain effects, as well as the effects of training joint multi-domain models.
2501.19137
A Metric for the Balance of Information in Graph Learning
cs.LG cs.AI
Graph learning on molecules makes use of information from both the molecular structure and the features attached to that structure. Much work has been conducted on biasing either towards structure or features, with the aim that bias bolsters performance. Identifying which information source a dataset favours, and therefore how to approach learning that dataset, is an open issue. Here we propose Noise-Noise Ratio Difference (NNRD), a quantitative metric for whether there is more useful information in structure or features. By employing iterative noising on features and structure independently, leaving the other intact, NNRD measures the degradation of information in each. We employ NNRD over a range of molecular tasks, and show that it corresponds well to a loss of information, with intuitive results that are more expressive than simple performance aggregates. Our future work will focus on expanding data domains, tasks and types, as well as refining our choice of baseline model.
2501.19140
Transformation trees -- documentation of multimodal image registration
cs.CV
The paper presents proposals for the application of a tree structure to the documentation of a set of transformations obtained as a result of various registrations of multimodal images obtained in coordinate systems associated with acquisition devices and being registered in one patient-specific coordinate system. A special file format .dpw (digital patient workspace) is introduced. Examples of different registrations yielded from orthodontic analysis and showing main aspects of the usage of tree structure are illustrated in dpVision software.
2501.19143
Imitation Game for Adversarial Disillusion with Multimodal Generative Chain-of-Thought Role-Play
cs.AI cs.CR cs.CV
As the cornerstone of artificial intelligence, machine perception confronts a fundamental threat posed by adversarial illusions. These adversarial attacks manifest in two primary forms: deductive illusion, where specific stimuli are crafted based on the victim model's general decision logic, and inductive illusion, where the victim model's general decision logic is shaped by specific stimuli. The former exploits the model's decision boundaries to create a stimulus that, when applied, interferes with its decision-making process. The latter reinforces a conditioned reflex in the model, embedding a backdoor during its learning phase that, when triggered by a stimulus, causes aberrant behaviours. The multifaceted nature of adversarial illusions calls for a unified defence framework, addressing vulnerabilities across various forms of attack. In this study, we propose a disillusion paradigm based on the concept of an imitation game. At the heart of the imitation game lies a multimodal generative agent, steered by chain-of-thought reasoning, which observes, internalises and reconstructs the semantic essence of a sample, liberated from the classic pursuit of reversing the sample to its original state. As a proof of concept, we conduct experimental simulations using a multimodal generative dialogue agent and evaluates the methodology under a variety of attack scenarios.
2501.19145
Improving Multi-Label Contrastive Learning by Leveraging Label Distribution
cs.LG cs.AI cs.CV
In multi-label learning, leveraging contrastive learning to learn better representations faces a key challenge: selecting positive and negative samples and effectively utilizing label information. Previous studies selected positive and negative samples based on the overlap between labels and used them for label-wise loss balancing. However, these methods suffer from a complex selection process and fail to account for the varying importance of different labels. To address these problems, we propose a novel method that improves multi-label contrastive learning through label distribution. Specifically, when selecting positive and negative samples, we only need to consider whether there is an intersection between labels. To model the relationships between labels, we introduce two methods to recover label distributions from logical labels, based on Radial Basis Function (RBF) and contrastive loss, respectively. We evaluate our method on nine widely used multi-label datasets, including image and vector datasets. The results demonstrate that our method outperforms state-of-the-art methods in six evaluation metrics.
2501.19148
Constant-Factor Distortion Mechanisms for $k$-Committee Election
cs.GT cs.DS cs.MA
In the $k$-committee election problem, we wish to aggregate the preferences of $n$ agents over a set of alternatives and select a committee of $k$ alternatives that minimizes the cost incurred by the agents. While we typically assume that agent preferences are captured by a cardinal utility function, in many contexts we only have access to ordinal information, namely the agents' rankings over the outcomes. As preference rankings are not as expressive as cardinal utilities, a loss of efficiency is inevitable, and is quantified by the notion of \emph{distortion}. We study the problem of electing a $k$-committee that minimizes the sum of the $\ell$-largest costs incurred by the agents, when agents and candidates are embedded in a metric space. This problem is called the $\ell$-centrum problem and captures both the utilitarian and egalitarian objectives. When $k \geq 2$, it is not possible to compute a bounded-distortion committee using purely ordinal information. We develop the first algorithms (that we call mechanisms) for the $\ell$-centrum problem (when $k \geq 2$), which achieve $O(1)$-distortion while eliciting only a very limited amount of cardinal information via value queries. We obtain two types of query-complexity guarantees: $O(\log k \log n)$ queries \emph{per agent}, and $O(k^2 \log^2 n)$ queries \emph{in total} (while achieving $O(1)$-distortion in both cases). En route, we give a simple adaptive-sampling algorithm for the $\ell$-centrum $k$-clustering problem.
2501.19149
On the inductive bias of infinite-depth ResNets and the bottleneck rank
cs.LG cs.AI stat.ML
We compute the minimum-norm weights of a deep linear ResNet, and find that the inductive bias of this architecture lies between minimizing nuclear norm and rank. This implies that, with appropriate hyperparameters, deep nonlinear ResNets have an inductive bias towards minimizing bottleneck rank.
2501.19153
Test-Time Training Scaling for Chemical Exploration in Drug Design
cs.LG
Chemical language models for molecular design have the potential to find solutions to multi-parameter optimization problems in drug discovery via reinforcement learning (RL). A key requirement to achieve this is the capacity to "search" chemical space to identify all molecules of interest. Here, we propose a challenging new benchmark to discover dissimilar molecules that possess similar bioactivity, a common scenario in drug discovery, but a hard problem to optimize. We show that a population of RL agents can solve the benchmark, while a single agent cannot. We also find that cooperative strategies are not significantly better than independent agents. Moreover, the performance on the benchmark scales log-linearly with the number of independent agents, showing a test-time training scaling law for chemical language models.
2501.19155
SWAT: Sliding Window Adversarial Training for Gradual Domain Adaptation
cs.CV cs.AI
Domain shifts are critical issues that harm the performance of machine learning. Unsupervised Domain Adaptation (UDA) mitigates this issue but suffers when the domain shifts are steep and drastic. Gradual Domain Adaptation (GDA) alleviates this problem in a mild way by gradually adapting from the source to the target domain using multiple intermediate domains. In this paper, we propose Sliding Window Adversarial Training (SWAT) for Gradual Domain Adaptation. SWAT uses the construction of adversarial streams to connect the feature spaces of the source and target domains. In order to gradually narrow the small gap between adjacent intermediate domains, a sliding window paradigm is designed that moves along the adversarial stream. When the window moves to the end of the stream, i.e., the target domain, the domain shift is drastically reduced. Extensive experiments are conducted on public GDA benchmarks, and the results demonstrate that the proposed SWAT significantly outperforms the state-of-the-art approaches. The implementation is available at: https://anonymous.4open.science/r/SWAT-8677.
2501.19158
A theoretical framework for overfitting in energy-based modeling
cs.LG cond-mat.dis-nn cond-mat.stat-mech
We investigate the impact of limited data on training pairwise energy-based models for inverse problems aimed at identifying interaction networks. Utilizing the Gaussian model as testbed, we dissect training trajectories across the eigenbasis of the coupling matrix, exploiting the independent evolution of eigenmodes and revealing that the learning timescales are tied to the spectral decomposition of the empirical covariance matrix. We see that optimal points for early stopping arise from the interplay between these timescales and the initial conditions of training. Moreover, we show that finite data corrections can be accurately modeled through asymptotic random matrix theory calculations and provide the counterpart of generalized cross-validation in the energy based model context. Our analytical framework extends to binary-variable maximum-entropy pairwise models with minimal variations. These findings offer strategies to control overfitting in discrete-variable models through empirical shrinkage corrections, improving the management of overfitting in energy-based generative models.
2501.19159
GDO: Gradual Domain Osmosis
cs.CV
In this paper, we propose a new method called Gradual Domain Osmosis, which aims to solve the problem of smooth knowledge migration from source domain to target domain in Gradual Domain Adaptation (GDA). Traditional Gradual Domain Adaptation methods mitigate domain bias by introducing intermediate domains and self-training strategies, but often face the challenges of inefficient knowledge migration or missing data in intermediate domains. In this paper, we design an optimisation framework based on the hyperparameter $\lambda$ by dynamically balancing the loss weights of the source and target domains, which enables the model to progressively adjust the strength of knowledge migration ($\lambda$ incrementing from 0 to 1) during the training process, thus achieving cross-domain generalisation more efficiently. Specifically, the method incorporates self-training to generate pseudo-labels and iteratively updates the model by minimising a weighted loss function to ensure stability and robustness during progressive adaptation in the intermediate domain. The experimental part validates the effectiveness of the method on rotated MNIST, colour-shifted MNIST, portrait dataset and forest cover type dataset, and the results show that it outperforms existing baseline methods. The paper further analyses the impact of the dynamic tuning strategy of the hyperparameter $\lambda$ on the performance through ablation experiments, confirming the advantages of progressive domain penetration in mitigating the domain bias and enhancing the model generalisation capability. The study provides a theoretical support and practical framework for asymptotic domain adaptation and expands its application potential in dynamic environments.
2501.19160
RMDM: Radio Map Diffusion Model with Physics Informed
cs.CV
With the rapid development of wireless communication technology, the efficient utilization of spectrum resources, optimization of communication quality, and intelligent communication have become critical. Radio map reconstruction is essential for enabling advanced applications, yet challenges such as complex signal propagation and sparse data hinder accurate reconstruction. To address these issues, we propose the **Radio Map Diffusion Model (RMDM)**, a physics-informed framework that integrates **Physics-Informed Neural Networks (PINNs)** to incorporate constraints like the **Helmholtz equation**. RMDM employs a dual U-Net architecture: the first ensures physical consistency by minimizing PDE residuals, boundary conditions, and source constraints, while the second refines predictions via diffusion-based denoising. By leveraging physical laws, RMDM significantly enhances accuracy, robustness, and generalization. Experiments demonstrate that RMDM outperforms state-of-the-art methods, achieving **NMSE of 0.0031** and **RMSE of 0.0125** under the Static RM (SRM) setting, and **NMSE of 0.0047** and **RMSE of 0.0146** under the Dynamic RM (DRM) setting. These results establish a novel paradigm for integrating physics-informed and data-driven approaches in radio map reconstruction, particularly under sparse data conditions.
2501.19161
Locality-aware Surrogates for Gradient-based Black-box Optimization
cs.LG
In physics and engineering, many processes are modeled using non-differentiable black-box simulators, making the optimization of such functions particularly challenging. To address such cases, inspired by the Gradient Theorem, we propose locality-aware surrogate models for active model-based black-box optimization. We first establish a theoretical connection between gradient alignment and the minimization of a Gradient Path Integral Equation (GradPIE) loss, which enforces consistency of the surrogate's gradients in local regions of the design space. Leveraging this theoretical insight, we develop a scalable training algorithm that minimizes the GradPIE loss, enabling both offline and online learning while maintaining computational efficiency. We evaluate our approach on three real-world tasks - spanning automated in silico experiments such as coupled nonlinear oscillators, analog circuits, and optical systems - and demonstrate consistent improvements in optimization efficiency under limited query budgets. Our results offer dependable solutions for both offline and online optimization tasks where reliable gradient estimation is needed.
2501.19164
Poison as Cure: Visual Noise for Mitigating Object Hallucinations in LVMs
cs.CV
Large vision-language models (LVMs) extend large language models (LLMs) with visual perception capabilities, enabling them to process and interpret visual information. A major challenge compromising their reliability is object hallucination that LVMs may generate plausible but factually inaccurate information. We propose a novel visual adversarial perturbation (VAP) method to mitigate this hallucination issue. VAP alleviates LVM hallucination by applying strategically optimized visual noise without altering the base model. Our approach formulates hallucination suppression as an optimization problem, leveraging adversarial strategies to generate beneficial visual perturbations that enhance the model's factual grounding and reduce parametric knowledge bias. Extensive experimental results demonstrate that our method consistently reduces object hallucinations across 8 state-of-the-art LVMs, validating its efficacy across diverse evaluations.
2501.19168
Implications of zero-growth economics analysed with an agent-based model
econ.GN cs.MA q-fin.EC
The ever-approaching limits of the Earth's biosphere and the potentially catastrophic consequences caused by climate change have begun to call into question the endless growth of the economy. There is increasing interest in the prospects of zero economic growth from the degrowth and post-growth literature. In particular, the question arises as to whether a zero-growth trajectory in a capitalist system with interest-bearing debt can be economically stable. There have been several answers to this question using macroeconomic models; some find a zero-growth trajectory is stable, while other models show an economic breakdown. However, the capitalist system in a period of growth is not guaranteed to be stable. Hence, a more appropriate methodology is to compare the relative stability between a growth and zero-growth scenario on the same model. Such a question has not yet been answered at any disaggregated level. It's important to investigate the consequences of zero-growth on market share instability and concentration, bankruptcy rates, income distribution, and credit network risk. To answer such questions, we develop a macroeconomic agent-based model incorporating Minskyan financial dynamics. The growth and zero-growth scenarios are accomplished by changing an average productivity growth parameter for the firms in the model. The model results showed that real GDP growth rates were more stable in the zero-growth scenario, there were fewer economic crises, lower unemployment rates, a higher wage share of output for workers, and capital firm and bank market shares were relatively more stable. Some of the consequences of zero-growth were a higher rate of inflation than in the growth scenario, increased market concentration for both firms and banks, and a higher level of financial risk in the credit network.
2501.19172
PSyDUCK: Training-Free Steganography for Latent Diffusion
cs.LG cs.CR
Recent advances in AI-generated steganography highlight its potential for safeguarding the privacy of vulnerable democratic actors, including aid workers, journalists, and whistleblowers operating in oppressive regimes. In this work, we address current limitations and establish the foundations for large-throughput generative steganography. We introduce a novel approach that enables secure and efficient steganography within latent diffusion models. We show empirically that our methods perform well across a variety of open-source latent diffusion models, particularly in generative image and video tasks.
2501.19176
Augmented Intelligence for Multimodal Virtual Biopsy in Breast Cancer Using Generative Artificial Intelligence
eess.IV cs.AI cs.CV
Full-Field Digital Mammography (FFDM) is the primary imaging modality for routine breast cancer screening; however, its effectiveness is limited in patients with dense breast tissue or fibrocystic conditions. Contrast-Enhanced Spectral Mammography (CESM), a second-level imaging technique, offers enhanced accuracy in tumor detection. Nonetheless, its application is restricted due to higher radiation exposure, the use of contrast agents, and limited accessibility. As a result, CESM is typically reserved for select cases, leaving many patients to rely solely on FFDM despite the superior diagnostic performance of CESM. While biopsy remains the gold standard for definitive diagnosis, it is an invasive procedure that can cause discomfort for patients. We introduce a multimodal, multi-view deep learning approach for virtual biopsy, integrating FFDM and CESM modalities in craniocaudal and mediolateral oblique views to classify lesions as malignant or benign. To address the challenge of missing CESM data, we leverage generative artificial intelligence to impute CESM images from FFDM scans. Experimental results demonstrate that incorporating the CESM modality is crucial to enhance the performance of virtual biopsy. When real CESM data is missing, synthetic CESM images proved effective, outperforming the use of FFDM alone, particularly in multimodal configurations that combine FFDM and CESM modalities. The proposed approach has the potential to improve diagnostic workflows, providing clinicians with augmented intelligence tools to improve diagnostic accuracy and patient care. Additionally, as a contribution to the research community, we publicly release the dataset used in our experiments, facilitating further advancements in this field.
2501.19178
No Foundations without Foundations -- Why semi-mechanistic models are essential for regulatory biology
cs.LG
Despite substantial efforts, deep learning has not yet delivered a transformative impact on elucidating regulatory biology, particularly in the realm of predicting gene expression profiles. Here, we argue that genuine "foundation models" of regulatory biology will remain out of reach unless guided by frameworks that integrate mechanistic insight with principled experimental design. We present one such ground-up, semi-mechanistic framework that unifies perturbation-based experimental designs across both in vitro and in vivo CRISPR screens, accounting for differentiating and non-differentiating cellular systems. By revealing previously unrecognised assumptions in published machine learning methods, our approach clarifies links with popular techniques such as variational autoencoders and structural causal models. In practice, this framework suggests a modified loss function that we demonstrate can improve predictive performance, and further suggests an error analysis that informs batching strategies. Ultimately, since cellular regulation emerges from innumerable interactions amongst largely uncharted molecular components, we contend that systems-level understanding cannot be achieved through structural biology alone. Instead, we argue that real progress will require a first-principles perspective on how experiments capture biological phenomena, how data are generated, and how these processes can be reflected in more faithful modelling architectures.
2501.19179
Learning Non-Local Molecular Interactions via Equivariant Local Representations and Charge Equilibration
physics.chem-ph cs.LG physics.comp-ph
Graph Neural Network (GNN) potentials relying on chemical locality offer near-quantum mechanical accuracy at significantly reduced computational costs. By propagating local information to distance particles, Message-passing neural networks (MPNNs) extend the locality concept to model interactions beyond their local neighborhood. Still, this locality precludes modeling long-range effects, such as charge transfer, electrostatic interactions, and dispersion effects, which are critical to adequately describe many real-world systems. In this work, we propose the Charge Equilibration Layer for Long-range Interactions (CELLI) to address the challenging modeling of non-local interactions and the high computational cost of MPNNs. This novel architecture generalizes the fourth-generation high-dimensional neural network (4GHDNN) concept, integrating the charge equilibration (Qeq) method into a model-agnostic building block for modern equivariant GNN potentials. A series of benchmarks show that CELLI can extend the strictly local Allegro architecture to model highly non-local interactions and charge transfer. Our architecture generalizes to diverse datasets and large structures, achieving an accuracy comparable to MPNNs at about twice the computational efficiency.
2501.19180
Enhancing Model Defense Against Jailbreaks with Proactive Safety Reasoning
cs.CR cs.AI
Large language models (LLMs) are vital for a wide range of applications yet remain susceptible to jailbreak threats, which could lead to the generation of inappropriate responses. Conventional defenses, such as refusal and adversarial training, often fail to cover corner cases or rare domains, leaving LLMs still vulnerable to more sophisticated attacks. We propose a novel defense strategy, Safety Chain-of-Thought (SCoT), which harnesses the enhanced \textit{reasoning capabilities} of LLMs for proactive assessment of harmful inputs, rather than simply blocking them. SCoT augments any refusal training datasets to critically analyze the intent behind each request before generating answers. By employing proactive reasoning, SCoT enhances the generalization of LLMs across varied harmful queries and scenarios not covered in the safety alignment corpus. Additionally, it generates detailed refusals specifying the rules violated. Comparative evaluations show that SCoT significantly surpasses existing defenses, reducing vulnerability to out-of-distribution issues and adversarial manipulations while maintaining strong general capabilities.
2501.19182
A Communication Framework for Compositional Generation
cs.LG
Compositionality and compositional generalization--the ability to understand novel combinations of known concepts--are central characteristics of human language and are hypothesized to be essential for human cognition. In machine learning, the emergence of this property has been studied in a communication game setting, where independent agents (a sender and a receiver) converge to a shared encoding policy from a set of states to a space of discrete messages, where the receiver can correctly reconstruct the states observed by the sender using only the sender's messages. The use of communication games in generation tasks is still largely unexplored, with recent methods for compositional generation focusing mainly on the use of supervised guidance (either through class labels or text). In this work, we take the first steps to fill this gap, and we present a self-supervised generative communication game-based framework for creating compositional encodings in learned representations from pre-trained encoder-decoder models. In an Iterated Learning (IL) protocol involving a sender and a receiver, we apply alternating pressures for compression and diversity of encoded discrete messages, so that the protocol converges to an efficient but unambiguous encoding. Approximate message entropy regularization is used to favor compositional encodings. Our framework is based on rigorous justifications and proofs of defining and balancing the concepts of Efficiency, Unambiguity and Non-Holisticity in encoding. We test our method on the compositional image dataset Shapes3D, demonstrating robust performance in both reconstruction and compositionality metrics, surpassing other tested discrete message frameworks.
2501.19183
Position: Curvature Matrices Should Be Democratized via Linear Operators
cs.LG
Structured large matrices are prevalent in machine learning. A particularly important class is curvature matrices like the Hessian, which are central to understanding the loss landscape of neural nets (NNs), and enable second-order optimization, uncertainty quantification, model pruning, data attribution, and more. However, curvature computations can be challenging due to the complexity of automatic differentiation, and the variety and structural assumptions of curvature proxies, like sparsity and Kronecker factorization. In this position paper, we argue that linear operators -- an interface for performing matrix-vector products -- provide a general, scalable, and user-friendly abstraction to handle curvature matrices. To support this position, we developed $\textit{curvlinops}$, a library that provides curvature matrices through a unified linear operator interface. We demonstrate with $\textit{curvlinops}$ how this interface can hide complexity, simplify applications, be extensible and interoperable with other libraries, and scale to large NNs.
2501.19184
A Survey on Class-Agnostic Counting: Advancements from Reference-Based to Open-World Text-Guided Approaches
cs.CV
Visual object counting has recently shifted towards class-agnostic counting (CAC), which addresses the challenge of counting objects across arbitrary categories -- a crucial capability for flexible and generalizable counting systems. Unlike humans, who effortlessly identify and count objects from diverse categories without prior knowledge, most existing counting methods are restricted to enumerating instances of known classes, requiring extensive labeled datasets for training and struggling in open-vocabulary settings. In contrast, CAC aims to count objects belonging to classes never seen during training, operating in a few-shot setting. In this paper, we present the first comprehensive review of CAC methodologies. We propose a taxonomy to categorize CAC approaches into three paradigms based on how target object classes can be specified: reference-based, reference-less, and open-world text-guided. Reference-based approaches achieve state-of-the-art performance by relying on exemplar-guided mechanisms. Reference-less methods eliminate exemplar dependency by leveraging inherent image patterns. Finally, open-world text-guided methods use vision-language models, enabling object class descriptions via textual prompts, offering a flexible and promising solution. Based on this taxonomy, we provide an overview of the architectures of 29 CAC approaches and report their results on gold-standard benchmarks. We compare their performance and discuss their strengths and limitations. Specifically, we present results on the FSC-147 dataset, setting a leaderboard using gold-standard metrics, and on the CARPK dataset to assess generalization capabilities. Finally, we offer a critical discussion of persistent challenges, such as annotation dependency and generalization, alongside future directions. We believe this survey will be a valuable resource, showcasing CAC advancements and guiding future research.
2501.19191
Secured Communication Schemes for UAVs in 5G: CRYSTALS-Kyber and IDS
cs.CR cs.AI
This paper introduces a secure communication architecture for Unmanned Aerial Vehicles (UAVs) and ground stations in 5G networks, addressing critical challenges in network security. The proposed solution integrates the Advanced Encryption Standard (AES) with Elliptic Curve Cryptography (ECC) and CRYSTALS-Kyber for key encapsulation, offering a hybrid cryptographic approach. By incorporating CRYSTALS-Kyber, the framework mitigates vulnerabilities in ECC against quantum attacks, positioning it as a quantum-resistant alternative. The architecture is based on a server-client model, with UAVs functioning as clients and the ground station acting as the server. The system was rigorously evaluated in both VPN and 5G environments. Experimental results confirm that CRYSTALS-Kyber delivers strong protection against quantum threats with minimal performance overhead, making it highly suitable for UAVs with resource constraints. Moreover, the proposed architecture integrates an Artificial Intelligence (AI)-based Intrusion Detection System (IDS) to further enhance security. In performance evaluations, the IDS demonstrated strong results across multiple models with XGBoost, particularly in more demanding scenarios, outperforming other models with an accuracy of 97.33% and an AUC of 0.94. These findings underscore the potential of combining quantum-resistant encryption mechanisms with AI-driven IDS to create a robust, scalable, and secure communication framework for UAV networks, particularly within the high-performance requirements of 5G environments.
2501.19194
APEX: Automated Parameter Exploration for Low-Power Wireless Protocols
cs.NI cs.SY eess.SY
Careful parametrization of networking protocols is crucial to maximize the performance of low-power wireless systems and ensure that stringent application requirements can be met. This is a non-trivial task involving thorough characterization on testbeds and requiring expert knowledge. Unfortunately, the community still lacks a tool to facilitate parameter exploration while minimizing the necessary experimentation time on testbeds. Such a tool would be invaluable, as exhaustive parameter searches can be time-prohibitive or unfeasible given the limited availability of testbeds, whereas non-exhaustive unguided searches rarely deliver satisfactory results. In this paper, we present APEX, a framework enabling an automated and informed parameter exploration for low-power wireless protocols and allowing to converge to an optimal parameter set within a limited number of testbed trials. We design APEX using Gaussian processes to effectively handle noisy experimental data and estimate the optimality of a certain parameter combination. After developing a prototype of APEX, we demonstrate its effectiveness by parametrizing two IEEE 802.15.4 protocols for a wide range of application requirements. Our results show that APEX can return the best parameter set with up to 10.6x, 4.5x and 3.25x less testbed trials than traditional solutions based on exhaustive search, greedy approaches, and reinforcement learning, respectively.
2501.19195
Rethinking Early Stopping: Refine, Then Calibrate
cs.LG cs.AI
Machine learning classifiers often produce probabilistic predictions that are critical for accurate and interpretable decision-making in various domains. The quality of these predictions is generally evaluated with proper losses like cross-entropy, which decompose into two components: calibration error assesses general under/overconfidence, while refinement error measures the ability to distinguish different classes. In this paper, we provide theoretical and empirical evidence that these two errors are not minimized simultaneously during training. Selecting the best training epoch based on validation loss thus leads to a compromise point that is suboptimal for both calibration error and, most importantly, refinement error. To address this, we introduce a new metric for early stopping and hyperparameter tuning that makes it possible to minimize refinement error during training. The calibration error is minimized after training, using standard techniques. Our method integrates seamlessly with any architecture and consistently improves performance across diverse classification tasks.
2501.19196
RaySplats: Ray Tracing based Gaussian Splatting
cs.CV
3D Gaussian Splatting (3DGS) is a process that enables the direct creation of 3D objects from 2D images. This representation offers numerous advantages, including rapid training and rendering. However, a significant limitation of 3DGS is the challenge of incorporating light and shadow reflections, primarily due to the utilization of rasterization rather than ray tracing for rendering. This paper introduces RaySplats, a model that employs ray-tracing based Gaussian Splatting. Rather than utilizing the projection of Gaussians, our method employs a ray-tracing mechanism, operating directly on Gaussian primitives represented by confidence ellipses with RGB colors. In practice, we compute the intersection between ellipses and rays to construct ray-tracing algorithms, facilitating the incorporation of meshes with Gaussian Splatting models and the addition of lights, shadows, and other related effects.
2501.19200
A Variational Perspective on Generative Protein Fitness Optimization
cs.LG
The goal of protein fitness optimization is to discover new protein variants with enhanced fitness for a given use. The vast search space and the sparsely populated fitness landscape, along with the discrete nature of protein sequences, pose significant challenges when trying to determine the gradient towards configurations with higher fitness. We introduce Variational Latent Generative Protein Optimization (VLGPO), a variational perspective on fitness optimization. Our method embeds protein sequences in a continuous latent space to enable efficient sampling from the fitness distribution and combines a (learned) flow matching prior over sequence mutations with a fitness predictor to guide optimization towards sequences with high fitness. VLGPO achieves state-of-the-art results on two different protein benchmarks of varying complexity. Moreover, the variational design with explicit prior and likelihood functions offers a flexible plug-and-play framework that can be easily customized to suit various protein design tasks.
2501.19201
Efficient Reasoning with Hidden Thinking
cs.CL cs.AI cs.LG
Chain-of-Thought (CoT) reasoning has become a powerful framework for improving complex problem-solving capabilities in Multimodal Large Language Models (MLLMs). However, the verbose nature of textual reasoning introduces significant inefficiencies. In this work, we propose $\textbf{Heima}$ (as hidden llama), an efficient reasoning framework that leverages reasoning CoTs at hidden latent space. We design the Heima Encoder to condense each intermediate CoT into a compact, higher-level hidden representation using a single thinking token, effectively minimizing verbosity and reducing the overall number of tokens required during the reasoning process. Meanwhile, we design corresponding Heima Decoder with traditional Large Language Models (LLMs) to adaptively interpret the hidden representations into variable-length textual sequence, reconstructing reasoning processes that closely resemble the original CoTs. Experimental results across diverse reasoning MLLM benchmarks demonstrate that Heima model achieves higher generation efficiency while maintaining or even better zero-shot task accuracy. Moreover, the effective reconstruction of multimodal reasoning processes with Heima Decoder validates both the robustness and interpretability of our approach.
2501.19202
Improving the Robustness of Representation Misdirection for Large Language Model Unlearning
cs.CL
Representation Misdirection (RM) and variants are established large language model (LLM) unlearning methods with state-of-the-art performance. In this paper, we show that RM methods inherently reduce models' robustness, causing them to misbehave even when a single non-adversarial forget-token is in the retain-query. Toward understanding underlying causes, we reframe the unlearning process as backdoor attacks and defenses: forget-tokens act as backdoor triggers that, when activated in retain-queries, cause disruptions in RM models' behaviors, similar to successful backdoor attacks. To mitigate this vulnerability, we propose Random Noise Augmentation -- a model and method agnostic approach with theoretical guarantees for improving the robustness of RM methods. Extensive experiments demonstrate that RNA significantly improves the robustness of RM models while enhancing the unlearning performances.
2501.19203
Single cell resolution 3D imaging and segmentation within intact live tissues
q-bio.QM cs.AI cs.CV q-bio.CB q-bio.TO
Epithelial cells form diverse structures from squamous spherical organoids to densely packed pseudostratified tissues. Quantification of cellular properties in these contexts requires high-resolution deep imaging and computational techniques to achieve truthful three-dimensional (3D) structural features. Here, we describe a detailed step-by-step protocol for sample preparation, imaging and deep-learning-assisted cell segmentation to achieve accurate quantification of fluorescently labelled individual cells in 3D within live tissues. We share the lessons learned through troubleshooting 3D imaging of Drosophila wing discs, including considerations on the choice of microscopy modality and settings (objective, sample mounting) and available segmentation methods. In addition, we include a computational pipeline alongside custom code to assist replication of the protocol. While we focus on the segmentation of cell outlines from membrane labelling, this protocol applies to a wide variety of samples, and we believe it be valuable for studying other tissues that demand complex analysis in 3D.
2501.19205
RIGNO: A Graph-based framework for robust and accurate operator learning for PDEs on arbitrary domains
cs.LG
Learning the solution operators of PDEs on arbitrary domains is challenging due to the diversity of possible domain shapes, in addition to the often intricate underlying physics. We propose an end-to-end graph neural network (GNN) based neural operator to learn PDE solution operators from data on point clouds in arbitrary domains. Our multi-scale model maps data between input/output point clouds by passing it through a downsampled regional mesh. Many novel elements are also incorporated to ensure resolution invariance and temporal continuity. Our model, termed RIGNO, is tested on a challenging suite of benchmarks, composed of various time-dependent and steady PDEs defined on a diverse set of domains. We demonstrate that RIGNO is significantly more accurate than neural operator baselines and robustly generalizes to unseen spatial resolutions and time instances.
2501.19206
An Empirical Game-Theoretic Analysis of Autonomous Cyber-Defence Agents
cs.AI cs.CR cs.GT
The recent rise in increasingly sophisticated cyber-attacks raises the need for robust and resilient autonomous cyber-defence (ACD) agents. Given the variety of cyber-attack tactics, techniques and procedures (TTPs) employed, learning approaches that can return generalisable policies are desirable. Meanwhile, the assurance of ACD agents remains an open challenge. We address both challenges via an empirical game-theoretic analysis of deep reinforcement learning (DRL) approaches for ACD using the principled double oracle (DO) algorithm. This algorithm relies on adversaries iteratively learning (approximate) best responses against each others' policies; a computationally expensive endeavour for autonomous cyber operations agents. In this work we introduce and evaluate a theoretically-sound, potential-based reward shaping approach to expedite this process. In addition, given the increasing number of open-source ACD-DRL approaches, we extend the DO formulation to allow for multiple response oracles (MRO), providing a framework for a holistic evaluation of ACD approaches.
2501.19207
Learning Sheaf Laplacian Optimizing Restriction Maps
eess.SP cs.LG
The aim of this paper is to propose a novel framework to infer the sheaf Laplacian, including the topology of a graph and the restriction maps, from a set of data observed over the nodes of a graph. The proposed method is based on sheaf theory, which represents an important generalization of graph signal processing. The learning problem aims to find the sheaf Laplacian that minimizes the total variation of the observed data, where the variation over each edge is also locally minimized by optimizing the associated restriction maps. Compared to alternative methods based on semidefinite programming, our solution is significantly more numerically efficient, as all its fundamental steps are resolved in closed form. The method is numerically tested on data consisting of vectors defined over subspaces of varying dimensions at each node. We demonstrate how the resulting graph is influenced by two key factors: the cross-correlation and the dimensionality difference of the data residing on the graph's nodes.
2501.19208
Learning While Repositioning in On-Demand Vehicle Sharing Networks
stat.ML cs.LG math.OC
We consider a network inventory problem motivated by one-way, on-demand vehicle sharing services. Due to uncertainties in both demand and returns, as well as a fixed number of rental units across an $n$-location network, the service provider must periodically reposition vehicles to match supply with demand spatially while minimizing costs. The optimal repositioning policy under a general $n$-location network is intractable without knowing the optimal value function. We introduce the best base-stock repositioning policy as a generalization of the classical inventory control policy to $n$ dimensions, and establish its asymptotic optimality in two distinct limiting regimes under general network structures. We present reformulations to efficiently compute this best base-stock policy in an offline setting with pre-collected data. In the online setting, we show that a natural Lipschitz-bandit approach achieves a regret guarantee of $\widetilde{O}(T^{\frac{n}{n+1}})$, which suffers from the exponential dependence on $n$. We illustrate the challenges of learning with censored data in networked systems through a regret lower bound analysis and by demonstrating the suboptimality of alternative algorithmic approaches. Motivated by these challenges, we propose an Online Gradient Repositioning algorithm that relies solely on censored demand. Under a mild cost-structure assumption, we prove that it attains an optimal regret of $O(n^{2.5} \sqrt{T})$, which matches the regret lower bound in $T$ and achieves only polynomial dependence on $n$. The key algorithmic innovation involves proposing surrogate costs to disentangle intertemporal dependencies and leveraging dual solutions to find the gradient of policy change. Numerical experiments demonstrate the effectiveness of our proposed methods.
2501.19214
A single-loop SPIDER-type stochastic subgradient method for expectation-constrained nonconvex nonsmooth optimization
math.OC cs.CC cs.LG cs.NA math.NA
Many real-world problems, such as those with fairness constraints, involve complex expectation constraints and large datasets, necessitating the design of efficient stochastic methods to solve them. Most existing research focuses on cases with no {constraint} or easy-to-project constraints or deterministic constraints. In this paper, we consider nonconvex nonsmooth stochastic optimization problems with expectation constraints, for which we build a novel exact penalty model. We first show the relationship between the penalty model and the original problem. Then on solving the penalty problem, we present a single-loop SPIDER-type stochastic subgradient method, which utilizes the subgradients of both the objective and constraint functions, as well as the constraint function value at each iteration. Under certain regularity conditions (weaker than Slater-type constraint qualification or strong feasibility assumed in existing works), we establish an iteration complexity result of $O(\epsilon^{-4})$ to reach a near-$\epsilon$ stationary point of the penalized problem in expectation, matching the lower bound for such tasks. Building on the exact penalization, an $(\epsilon,\epsilon)$-KKT point of the original problem is obtained. For a few scenarios, our complexity of either the {objective} sample subgradient or the constraint sample function values can be lower than the state-of-the-art results by a factor of $\epsilon^{-2}$. Moreover, on solving two fairness-constrained problems, our method is significantly (up to 466 times) faster than the state-of-the-art algorithms, including switching subgradient method and inexact proximal point methods.
2501.19215
Strassen Attention: Unlocking Compositional Abilities in Transformers Based on a New Lower Bound Method
cs.LG cs.AI
We propose a novel method to evaluate the theoretical limits of Transformers, allowing us to prove the first lower bounds against one-layer softmax Transformers with infinite precision. We establish those bounds for three tasks that require advanced reasoning. The first task, Match3 (Sanford et al., 2023), requires looking at all triples of positions. The second and third tasks address compositionality-based reasoning: one is composition of functions (Peng et al., 2024) and the other is composition of binary relations. We formally prove the inability of one-layer softmax Transformers to solve any of these tasks. In an attempt to overcome these limitations, we introduce Strassen attention and prove that with this mechanism a one-layer Transformer can in principle solve all these tasks. We also show that it enjoys sub-cubic running-time complexity, making it more scalable than similar previously proposed mechanisms, such as higher-order attention (Sanford et al., 2023). To complement our theoretical findings, we experimentally studied Strassen attention and compared it against standard (Vaswani et al, 2017), higher-order attention (Sanford et al., 2023) and triangular attention (Bergen et al. 2021). Our results help to disentangle all these attention mechanisms, highlighting their strengths and limitations. In particular, Strassen attention outperforms standard attention significantly on all the tasks. Altogether, understanding the theoretical limitations can guide research towards scalable attention mechanisms that improve the reasoning abilities of Transformers.
2501.19216
E2Former: A Linear-time Efficient and Equivariant Transformer for Scalable Molecular Modeling
cs.LG cond-mat.mtrl-sci
Equivariant Graph Neural Networks (EGNNs) have demonstrated significant success in modeling microscale systems, including those in chemistry, biology and materials science. However, EGNNs face substantial computational challenges due to the high cost of constructing edge features via spherical tensor products, making them impractical for large-scale systems. To address this limitation, we introduce E2Former, an equivariant and efficient transformer architecture that incorporates the Wigner $6j$ convolution (Wigner $6j$ Conv). By shifting the computational burden from edges to nodes, the Wigner $6j$ Conv reduces the complexity from $O(|\mathcal{E}|)$ to $ O(| \mathcal{V}|)$ while preserving both the model's expressive power and rotational equivariance. We show that this approach achieves a 7x-30x speedup compared to conventional $\mathrm{SO}(3)$ convolutions. Furthermore, our empirical results demonstrate that the derived E2Former mitigates the computational challenges of existing approaches without compromising the ability to capture detailed geometric information. This development could suggest a promising direction for scalable and efficient molecular modeling.
2501.19218
A parallelizable variant of HCA*
eess.SY cs.SY
This paper presents a parallelizable variant of the well-known Hierarchical Cooperative A* algorithm (HCA*) for the multi-agent path finding (MAPF) problem. In this variant, all agents initially find their shortest paths disregarding the presence of others. This is done using A*. Then an intersection graph (IG) is constructed; each agent is a node and two nodes have an edge between them if the paths of corresponding agents collide. Thereafter, an independent set is extracted with the aid of an approximation algorithm for the maximum independent set problem. The paths for the agents belonging to independent set are fixed. The rest of agents now again find their shortest paths, this time ensuring no collision with the prior agents. Space-time A*, which is a crucial component of HCA*, is used here. These iterations continue until no agents are left. Since the tasks of finding shortest paths for the agents in any iteration are independent of each other, the proposed algorithm can be parallelized to a large extent. In addition to this, the task of determining the IG can also be done in parallel by dividing the map into sections and with each agent focusing on a particular section. The parallelism does come at a cost of communication between the agents and the server. This is accounted for in the simulations. As an added advantage, the user need not make a choice for the priority order. It is observed, empirically, that the proposed algorithm outperforms HCA* in terms of the computation time and the cost value in many cases. Simulations are provided for corroboration.
2501.19220
Analysis and predictability of centrality measures in competition networks
cs.SI
The Common Out-Neighbor (or CON) score quantifies shared influence through outgoing links in competitive contexts. A dynamic analysis of competition networks reveals the CON score as a powerful predictor of node rankings. Defined in first-order and second-order forms, the CON score captures both direct and indirect competitive interactions, offering a comprehensive metric for evaluating node influence. Using datasets from Survivor, Chess.com, and Dota~2 online gaming competitions, directed competition networks are constructed, and the dynamic CON score is integrated into supervised machine learning models. Empirical results show that the CON score consistently outperforms traditional centrality measures such as PageRank, closeness, and betweenness centrality in classification tasks. By integrating dynamic centrality measures with machine learning, our proposed methodology accurately predicts outcomes in competition networks. The findings underline the CON score's robustness as a feature in node classification, offering a significant advancement in understanding and analyzing competitive interactions.
2501.19223
Through the Looking Glass: LLM-Based Analysis of AR/VR Android Applications Privacy Policies
cs.CR cs.LG
\begin{abstract} This paper comprehensively analyzes privacy policies in AR/VR applications, leveraging BERT, a state-of-the-art text classification model, to evaluate the clarity and thoroughness of these policies. By comparing the privacy policies of AR/VR applications with those of free and premium websites, this study provides a broad perspective on the current state of privacy practices within the AR/VR industry. Our findings indicate that AR/VR applications generally offer a higher percentage of positive segments than free content but lower than premium websites. The analysis of highlighted segments and words revealed that AR/VR applications strategically emphasize critical privacy practices and key terms. This enhances privacy policies' clarity and effectiveness.
2501.19224
Fast exact recovery of noisy matrix from few entries: the infinity norm approach
math.ST cs.LG math.CO math.PR stat.AP stat.TH
The matrix recovery (completion) problem, a central problem in data science and theoretical computer science, is to recover a matrix $A$ from a relatively small sample of entries. While such a task is impossible in general, it has been shown that one can recover $A$ exactly in polynomial time, with high probability, from a random subset of entries, under three (basic and necessary) assumptions: (1) the rank of $A$ is very small compared to its dimensions (low rank), (2) $A$ has delocalized singular vectors (incoherence), and (3) the sample size is sufficiently large. There are many different algorithms for the task, including convex optimization by Candes, Tao and Recht (2009), alternating projection by Hardt and Wooters (2014) and low rank approximation with gradient descent by Keshavan, Montanari and Oh (2009, 2010). In applications, it is more realistic to assume that data is noisy. In this case, these approaches provide an approximate recovery with small root mean square error. However, it is hard to transform such approximate recovery to an exact one. Recently, results by Abbe et al. (2017) and Bhardwaj et al. (2023) concerning approximation in the infinity norm showed that we can achieve exact recovery even in the noisy case, given that the ground matrix has bounded precision. Beyond the three basic assumptions above, they required either the condition number of $A$ is small (Abbe et al.) or the gap between consecutive singular values is large (Bhardwaj et al.). In this paper, we remove these extra spectral assumptions. As a result, we obtain a simple algorithm for exact recovery in the noisy case, under only three basic assumptions. This is the first such algorithm. To analyse the algorithm, we introduce a contour integration argument which is totally different from all previous methods and may be of independent interest.
2501.19227
Integrating Semi-Supervised and Active Learning for Semantic Segmentation
cs.CV cs.AI
In this paper, we propose a novel active learning approach integrated with an improved semi-supervised learning framework to reduce the cost of manual annotation and enhance model performance. Our proposed approach effectively leverages both the labelled data selected through active learning and the unlabelled data excluded from the selection process. The proposed active learning approach pinpoints areas where the pseudo-labels are likely to be inaccurate. Then, an automatic and efficient pseudo-label auto-refinement (PLAR) module is proposed to correct pixels with potentially erroneous pseudo-labels by comparing their feature representations with those of labelled regions. This approach operates without increasing the labelling budget and is based on the cluster assumption, which states that pixels belonging to the same class should exhibit similar representations in feature space. Furthermore, manual labelling is only applied to the most difficult and uncertain areas in unlabelled data, where insufficient information prevents the PLAR module from making a decision. We evaluated the proposed hybrid semi-supervised active learning framework on two benchmark datasets, one from natural and the other from remote sensing imagery domains. In both cases, it outperformed state-of-the-art methods in the semantic segmentation task.
2501.19232
A Zero-Shot Generalization Framework for LLM-Driven Cross-Domain Sequential Recommendation
cs.IR cs.AI
Zero-shot cross-domain sequential recommendation (ZCDSR) enables predictions in unseen domains without the need for additional training or fine-tuning, making it particularly valuable in data-sparse environments where traditional models struggle. Recent advancements in large language models (LLMs) have greatly improved ZCDSR by leveraging rich pretrained representations to facilitate cross-domain knowledge transfer. However, a key challenge persists: domain semantic bias, which arises from variations in vocabulary and content focus across domains. This misalignment leads to inconsistencies in item embeddings and hinders generalization. To address this issue, we propose a novel framework designed to enhance LLM-based ZCDSR by improving cross-domain alignment at both the item and sequential levels. At the item level, we introduce a generalization loss that promotes inter-domain compactness by aligning embeddings of similar items across domains while maintaining intra-domain diversity to preserve unique item characteristics. This prevents embeddings from becoming overly generic while ensuring effective transferability. At the sequential level, we develop a method for transferring user behavioral patterns by clustering user sequences in the source domain and applying attention-based aggregation for target domain inference. This dynamic adaptation of user embeddings allows effective zero-shot recommendations without requiring target-domain interactions. Comprehensive experiments across multiple datasets and domains demonstrate that our framework significantly improves sequential recommendation performance in the ZCDSR setting. By mitigating domain bias and enhancing the transferability of sequential patterns, our method provides a scalable and robust approach for achieving more effective zero-shot recommendations across domains.
2501.19234
Hourly Short Term Load Forecasting for Residential Buildings and Energy Communities
cs.LG
Electricity load consumption may be extremely complex in terms of profile patterns, as it depends on a wide range of human factors, and it is often correlated with several exogenous factors, such as the availability of renewable energy and the weather conditions. The first goal of this paper is to investigate the performance of a large selection of different types of forecasting models in predicting the electricity load consumption within the short time horizon of a day or few hours ahead. Such forecasts may be rather useful for the energy management of individual residential buildings or small energy communities. In particular, we introduce persistence models, standard auto-regressive-based machine learning models, and more advanced deep learning models. The second goal of this paper is to introduce two alternative modeling approaches that are simpler in structure while they take into account domain specific knowledge, as compared to the previously mentioned black-box modeling techniques. In particular, we consider the persistence-based auto-regressive model (PAR) and the seasonal persistence-based regressive model (SPR), priorly introduced by the authors. In this paper, we specifically tailor these models to accommodate the generation of hourly forecasts. The introduced models and the induced comparative analysis extend prior work of the authors which was restricted to day-ahead forecasts. We observed a 15-30% increase in the prediction accuracy of the newly introduced hourly-based forecasting models over existing approaches.
2501.19237
DINAMO: Dynamic and INterpretable Anomaly MOnitoring for Large-Scale Particle Physics Experiments
hep-ex cs.LG
Ensuring reliable data collection in large-scale particle physics experiments demands Data Quality Monitoring (DQM) procedures to detect possible detector malfunctions and preserve data integrity. Traditionally, this resource-intensive task has been handled by human shifters that struggle with frequent changes in operational conditions. We present novel, interpretable, robust, and scalable DQM algorithms designed to automate anomaly detection in time-dependent settings. Our approach constructs evolving histogram templates with built-in uncertainties, featuring both a statistical variant - extending the classical Exponentially Weighted Moving Average (EWMA) - and a machine learning (ML)-enhanced version that leverages a transformer encoder for improved adaptability. Experimental validations on synthetic datasets demonstrate the high accuracy, adaptability, and interpretability of these methods, with the statistical variant being commissioned in the LHCb experiment at the Large Hadron Collider, underscoring its real-world impact. The code used in this study is available at https://github.com/ArseniiGav/DINAMO.
2501.19239
Multi-agent Multi-armed Bandit with Fully Heavy-tailed Dynamics
cs.LG stat.ML
We study decentralized multi-agent multi-armed bandits in fully heavy-tailed settings, where clients communicate over sparse random graphs with heavy-tailed degree distributions and observe heavy-tailed (homogeneous or heterogeneous) reward distributions with potentially infinite variance. The objective is to maximize system performance by pulling the globally optimal arm with the highest global reward mean across all clients. We are the first to address such fully heavy-tailed scenarios, which capture the dynamics and challenges in communication and inference among multiple clients in real-world systems. In homogeneous settings, our algorithmic framework exploits hub-like structures unique to heavy-tailed graphs, allowing clients to aggregate rewards and reduce noises via hub estimators when constructing UCB indices; under $M$ clients and degree distributions with power-law index $\alpha > 1$, our algorithm attains a regret bound (almost) of order $O(M^{1 -\frac{1}{\alpha}} \log{T})$. Under heterogeneous rewards, clients synchronize by communicating with neighbors, aggregating exchanged estimators in UCB indices; With our newly established information delay bounds on sparse random graphs, we prove a regret bound of $O(M \log{T})$. Our results improve upon existing work, which only address time-invariant connected graphs, or light-tailed dynamics in dense graphs and rewards.
2501.19241
Emancipatory Information Retrieval
cs.IR cs.HC
Our world today is facing a confluence of several mutually reinforcing crises each of which intersects with concerns of social justice and emancipation. This paper is a provocation for the role of computer-mediated information access in our emancipatory struggles. We define emancipatory information retrieval as the study and development of information access methods that challenge various forms of human oppression, and situates its activities within broader collective emancipatory praxis. The term "emancipatory" here signifies the moral concerns of universal humanization of all peoples and the elimination of oppression to create the conditions under which we can collectively flourish. To develop an emancipatory research agenda for information retrieval (IR), in this paper we speculate about the practices that the community can adopt, enumerate some of the projects that the field should undertake, and discuss provocations to spark new ideas and directions for research. We challenge the field of IR research to embrace humanistic values and commit to universal emancipation and social justice. We also invite scholars from fields such as human-computer interaction, information sciences, media studies, design, social sciences, humanities, democratic theory, and critical theory, as well as legal and policy experts, civil rights and social justice activists, and artists to join us in realizing this transformation. In this process, we must both imagine post-oppressive worlds, and reimagine the role of IR in that world and in the journey that leads us there.
2501.19243
Accelerating Diffusion Transformer via Error-Optimized Cache
cs.CV
Diffusion Transformer (DiT) is a crucial method for content generation. However, it needs a lot of time to sample. Many studies have attempted to use caching to reduce the time consumption of sampling. Existing caching methods accelerate generation by reusing DiT features from the previous time step and skipping calculations in the next, but they tend to locate and cache low-error modules without focusing on reducing caching-induced errors, resulting in a sharp decline in generated content quality when increasing caching intensity. To solve this problem, we propose the Error-Optimized Cache (EOC). This method introduces three key improvements: (1) Prior knowledge extraction: Extract and process the caching differences; (2) A judgment method for cache optimization: Determine whether certain caching steps need to be optimized; (3) Cache optimization: reduce caching errors. Experiments show that this algorithm significantly reduces the error accumulation caused by caching (especially over-caching). On the ImageNet dataset, without significantly increasing the computational burden, this method improves the quality of the generated images under the over-caching, rule-based, and training-based methods. Specifically, the Fr\'echet Inception Distance (FID) values are improved as follows: from 6.857 to 5.821, from 3.870 to 3.692 and form 3.539 to 3.451 respectively.
2501.19245
SHARPIE: A Modular Framework for Reinforcement Learning and Human-AI Interaction Experiments
cs.AI cs.HC
Reinforcement learning (RL) offers a general approach for modeling and training AI agents, including human-AI interaction scenarios. In this paper, we propose SHARPIE (Shared Human-AI Reinforcement Learning Platform for Interactive Experiments) to address the need for a generic framework to support experiments with RL agents and humans. Its modular design consists of a versatile wrapper for RL environments and algorithm libraries, a participant-facing web interface, logging utilities, deployment on popular cloud and participant recruitment platforms. It empowers researchers to study a wide variety of research questions related to the interaction between humans and RL agents, including those related to interactive reward specification and learning, learning from human feedback, action delegation, preference elicitation, user-modeling, and human-AI teaming. The platform is based on a generic interface for human-RL interactions that aims to standardize the field of study on RL in human contexts.
2501.19247
Clustering in hyperbolic balls
cs.LG
The idea of representations of the data in negatively curved manifolds recently attracted a lot of attention and gave a rise to the new research direction named {\it hyperbolic machine learning} (ML). In order to unveil the full potential of this new paradigm, efficient techniques for data analysis and statistical modeling in hyperbolic spaces are necessary. In the present paper rigorous mathematical framework for clustering in hyperbolic spaces is established. First, we introduce the $k$-means clustering in hyperbolic balls, based on the novel definition of barycenter. Second, we present the expectation-maximization (EM) algorithm for learning mixtures of novel probability distributions in hyperbolic balls. In such a way we lay the foundation of unsupervised learning in hyperbolic spaces.
2501.19252
Inference-Time Text-to-Video Alignment with Diffusion Latent Beam Search
cs.CV
The remarkable progress in text-to-video diffusion models enables photorealistic generations, although the contents of the generated video often include unnatural movement or deformation, reverse playback, and motionless scenes. Recently, an alignment problem has attracted huge attention, where we steer the output of diffusion models based on some quantity on the goodness of the content. Because there is a large room for improvement of perceptual quality along the frame direction, we should address which metrics we should optimize and how we can optimize them in the video generation. In this paper, we propose diffusion latent beam search with lookahead estimator, which can select better diffusion latent to maximize a given alignment reward, at inference time. We then point out that the improvement of perceptual video quality considering the alignment to prompts requires reward calibration by weighting existing metrics. When evaluating outputs by using vision language models as a proxy of humans, many previous metrics to quantify the naturalness of video do not always correlate with evaluation and also depend on the degree of dynamic descriptions in evaluation prompts. We demonstrate that our method improves the perceptual quality based on the calibrated reward, without model parameter update, and outputs the best generation compared to greedy search and best-of-N sampling. We provide practical guidelines on which axes, among search budget, lookahead steps for reward estimate, and denoising steps, in the reverse diffusion process, we should allocate the inference-time computation.
2501.19254
Linear $Q$-Learning Does Not Diverge: Convergence Rates to a Bounded Set
cs.LG cs.AI stat.ML
$Q$-learning is one of the most fundamental reinforcement learning algorithms. Previously, it is widely believed that $Q$-learning with linear function approximation (i.e., linear $Q$-learning) suffers from possible divergence. This paper instead establishes the first $L^2$ convergence rate of linear $Q$-learning to a bounded set. Notably, we do not make any modification to the original linear $Q$-learning algorithm, do not make any Bellman completeness assumption, and do not make any near-optimality assumption on the behavior policy. All we need is an $\epsilon$-softmax behavior policy with an adaptive temperature. The key to our analysis is the general result of stochastic approximations under Markovian noise with fast-changing transition functions. As a side product, we also use this general result to establish the $L^2$ convergence rate of tabular $Q$-learning with an $\epsilon$-softmax behavior policy, for which we rely on a novel pseudo-contraction property of the weighted Bellman optimality operator.
2501.19255
ContextFormer: Redefining Efficiency in Semantic Segmentation
cs.CV
Semantic segmentation assigns labels to pixels in images, a critical yet challenging task in computer vision. Convolutional methods, although capturing local dependencies well, struggle with long-range relationships. Vision Transformers (ViTs) excel in global context capture but are hindered by high computational demands, especially for high-resolution inputs. Most research optimizes the encoder architecture, leaving the bottleneck underexplored - a key area for enhancing performance and efficiency. We propose ContextFormer, a hybrid framework leveraging the strengths of CNNs and ViTs in the bottleneck to balance efficiency, accuracy, and robustness for real-time semantic segmentation. The framework's efficiency is driven by three synergistic modules: the Token Pyramid Extraction Module (TPEM) for hierarchical multi-scale representation, the Transformer and Modulating DepthwiseConv (Trans-MDC) block for dynamic scale-aware feature modeling, and the Feature Merging Module (FMM) for robust integration with enhanced spatial and contextual consistency. Extensive experiments on ADE20K, Pascal Context, CityScapes, and COCO-Stuff datasets show ContextFormer significantly outperforms existing models, achieving state-of-the-art mIoU scores, setting a new benchmark for efficiency and performance. The codes will be made publicly available.
2501.19256
Objective Metrics for Human-Subjects Evaluation in Explainable Reinforcement Learning
cs.AI cs.HC cs.RO
Explanation is a fundamentally human process. Understanding the goal and audience of the explanation is vital, yet existing work on explainable reinforcement learning (XRL) routinely does not consult humans in their evaluations. Even when they do, they routinely resort to subjective metrics, such as confidence or understanding, that can only inform researchers of users' opinions, not their practical effectiveness for a given problem. This paper calls on researchers to use objective human metrics for explanation evaluations based on observable and actionable behaviour to build more reproducible, comparable, and epistemically grounded research. To this end, we curate, describe, and compare several objective evaluation methodologies for applying explanations to debugging agent behaviour and supporting human-agent teaming, illustrating our proposed methods using a novel grid-based environment. We discuss how subjective and objective metrics complement each other to provide holistic validation and how future work needs to utilise standardised benchmarks for testing to enable greater comparisons between research.
2501.19258
VisualSpeech: Enhance Prosody with Visual Context in TTS
cs.CL
Text-to-Speech (TTS) synthesis faces the inherent challenge of producing multiple speech outputs with varying prosody from a single text input. While previous research has addressed this by predicting prosodic information from both text and speech, additional contextual information, such as visual features, remains underutilized. This paper investigates the potential of integrating visual context to enhance prosody prediction. We propose a novel model, VisualSpeech, which incorporates both visual and textual information for improved prosody generation. Empirical results demonstrate that visual features provide valuable prosodic cues beyond the textual input, significantly enhancing the naturalness and accuracy of the synthesized speech. Audio samples are available at https://ariameetgit.github.io/VISUALSPEECH-SAMPLES/.
2501.19259
Neuro-LIFT: A Neuromorphic, LLM-based Interactive Framework for Autonomous Drone FlighT at the Edge
cs.RO cs.CV cs.LG cs.NE cs.SY eess.SY
The integration of human-intuitive interactions into autonomous systems has been limited. Traditional Natural Language Processing (NLP) systems struggle with context and intent understanding, severely restricting human-robot interaction. Recent advancements in Large Language Models (LLMs) have transformed this dynamic, allowing for intuitive and high-level communication through speech and text, and bridging the gap between human commands and robotic actions. Additionally, autonomous navigation has emerged as a central focus in robotics research, with artificial intelligence (AI) increasingly being leveraged to enhance these systems. However, existing AI-based navigation algorithms face significant challenges in latency-critical tasks where rapid decision-making is critical. Traditional frame-based vision systems, while effective for high-level decision-making, suffer from high energy consumption and latency, limiting their applicability in real-time scenarios. Neuromorphic vision systems, combining event-based cameras and spiking neural networks (SNNs), offer a promising alternative by enabling energy-efficient, low-latency navigation. Despite their potential, real-world implementations of these systems, particularly on physical platforms such as drones, remain scarce. In this work, we present Neuro-LIFT, a real-time neuromorphic navigation framework implemented on a Parrot Bebop2 quadrotor. Leveraging an LLM for natural language processing, Neuro-LIFT translates human speech into high-level planning commands which are then autonomously executed using event-based neuromorphic vision and physics-driven planning. Our framework demonstrates its capabilities in navigating in a dynamic environment, avoiding obstacles, and adapting to human instructions in real-time.
2501.19264
mFollowIR: a Multilingual Benchmark for Instruction Following in Retrieval
cs.IR cs.CL cs.LG
Retrieval systems generally focus on web-style queries that are short and underspecified. However, advances in language models have facilitated the nascent rise of retrieval models that can understand more complex queries with diverse intents. However, these efforts have focused exclusively on English; therefore, we do not yet understand how they work across languages. We introduce mFollowIR, a multilingual benchmark for measuring instruction-following ability in retrieval models. mFollowIR builds upon the TREC NeuCLIR narratives (or instructions) that span three diverse languages (Russian, Chinese, Persian) giving both query and instruction to the retrieval models. We make small changes to the narratives and isolate how well retrieval models can follow these nuanced changes. We present results for both multilingual (XX-XX) and cross-lingual (En-XX) performance. We see strong cross-lingual performance with English-based retrievers that trained using instructions, but find a notable drop in performance in the multilingual setting, indicating that more work is needed in developing data for instruction-based multilingual retrievers.
2501.19265
Medical Semantic Segmentation with Diffusion Pretrain
cs.CV cs.LG
Recent advances in deep learning have shown that learning robust feature representations is critical for the success of many computer vision tasks, including medical image segmentation. In particular, both transformer and convolutional-based architectures have benefit from leveraging pretext tasks for pretraining. However, the adoption of pretext tasks in 3D medical imaging has been less explored and remains a challenge, especially in the context of learning generalizable feature representations. We propose a novel pretraining strategy using diffusion models with anatomical guidance, tailored to the intricacies of 3D medical image data. We introduce an auxiliary diffusion process to pretrain a model that produce generalizable feature representations, useful for a variety of downstream segmentation tasks. We employ an additional model that predicts 3D universal body-part coordinates, providing guidance during the diffusion process and improving spatial awareness in generated representations. This approach not only aids in resolving localization inaccuracies but also enriches the model's ability to understand complex anatomical structures. Empirical validation on a 13-class organ segmentation task demonstrate the effectiveness of our pretraining technique. It surpasses existing restorative pretraining methods in 3D medical image segmentation by $7.5\%$, and is competitive with the state-of-the-art contrastive pretraining approach, achieving an average Dice coefficient of 67.8 in a non-linear evaluation scenario.
2501.19266
Jackpot! Alignment as a Maximal Lottery
cs.AI cs.LG econ.TH
Reinforcement Learning from Human Feedback (RLHF), the standard for aligning Large Language Models (LLMs) with human values, is known to fail to satisfy properties that are intuitively desirable, such as respecting the preferences of the majority \cite{ge2024axioms}. To overcome these issues, we propose the use of a probabilistic Social Choice rule called \emph{maximal lotteries} as a replacement for RLHF. We show that a family of alignment techniques, namely Nash Learning from Human Feedback (NLHF) \cite{munos2023nash} and variants, approximate maximal lottery outcomes and thus inherit its beneficial properties. We confirm experimentally that our proposed methodology handles situations that arise when working with preferences more robustly than standard RLHF, including supporting the preferences of the majority, providing principled ways of handling non-transitivities in the preference data, and robustness to irrelevant alternatives. This results in systems that better incorporate human values and respect human intentions.
2501.19267
Transformer-Based Financial Fraud Detection with Cloud-Optimized Real-Time Streaming
cs.CE
As the financial industry becomes more interconnected and reliant on digital systems, fraud detection systems must evolve to meet growing threats. Cloud-enabled Transformer models present a transformative opportunity to address these challenges. By leveraging the scalability, flexibility, and advanced AI capabilities of cloud platforms, companies can deploy fraud detection solutions that adapt to real-time data patterns and proactively respond to evolving threats. Using the Graph self-attention Transformer neural network module, we can directly excavate gang fraud features from the transaction network without constructing complicated feature engineering. Finally, the fraud prediction network is combined to optimize the topological pattern and the temporal transaction pattern to realize the high-precision detection of fraudulent transactions. The results of antifraud experiments on credit card transaction data show that the proposed model outperforms the 7 baseline models on all evaluation indicators: In the transaction fraud detection task, the average accuracy (AP) increased by 20% and the area under the ROC curve (AUC) increased by 2.7% on average compared with the benchmark graph attention neural network (GAT), which verified the effectiveness of the proposed model in the detection of credit card fraud transactions.
2501.19270
Imagine with the Teacher: Complete Shape in a Multi-View Distillation Way
cs.CV
Point cloud completion aims to recover the completed 3D shape of an object from its partial observation caused by occlusion, sensor's limitation, noise, etc. When some key semantic information is lost in the incomplete point cloud, the neural network needs to infer the missing part based on the input information. Intuitively we would apply an autoencoder architecture to solve this kind of problem, which take the incomplete point cloud as input and is supervised by the ground truth. This process that develops model's imagination from incomplete shape to complete shape is done automatically in the latent space. But the knowledge for mapping from incomplete to complete still remains dark and could be further explored. Motivated by the knowledge distillation's teacher-student learning strategy, we design a knowledge transfer way for completing 3d shape. In this work, we propose a novel View Distillation Point Completion Network (VD-PCN), which solve the completion problem by a multi-view distillation way. The design methodology fully leverages the orderliness of 2d pixels, flexibleness of 2d processing and powerfulness of 2d network. Extensive evaluations on PCN, ShapeNet55/34, and MVP datasets confirm the effectiveness of our design and knowledge transfer strategy, both quantitatively and qualitatively. Committed to facilitate ongoing research, we will make our code publicly available.
2501.19271
Concept-Based Explainable Artificial Intelligence: Metrics and Benchmarks
cs.AI cs.LG
Concept-based explanation methods, such as concept bottleneck models (CBMs), aim to improve the interpretability of machine learning models by linking their decisions to human-understandable concepts, under the critical assumption that such concepts can be accurately attributed to the network's feature space. However, this foundational assumption has not been rigorously validated, mainly because the field lacks standardised metrics and benchmarks to assess the existence and spatial alignment of such concepts. To address this, we propose three metrics: the concept global importance metric, the concept existence metric, and the concept location metric, including a technique for visualising concept activations, i.e., concept activation mapping. We benchmark post-hoc CBMs to illustrate their capabilities and challenges. Through qualitative and quantitative experiments, we demonstrate that, in many cases, even the most important concepts determined by post-hoc CBMs are not present in input images; moreover, when they are present, their saliency maps fail to align with the expected regions by either activating across an entire object or misidentifying relevant concept-specific regions. We analyse the root causes of these limitations, such as the natural correlation of concepts. Our findings underscore the need for more careful application of concept-based explanation techniques especially in settings where spatial interpretability is critical.
2501.19273
Minimax discrete distribution estimation with self-consumption
cs.IT math.IT math.ST stat.TH
Learning distributions from i.i.d. samples is a well-understood problem. However, advances in generative machine learning prompt an interesting new, non-i.i.d. setting: after receiving a certain number of samples, an estimated distribution is fixed, and samples from this estimate are drawn and introduced into the sample corpus, undifferentiated from real samples. Subsequent generations of estimators now face contaminated environments, an effect referred to in the machine learning literature as self-consumption. In this paper, we study the effect of such contamination from previous estimates on the minimax loss of multi-stage discrete distribution estimation. In the data accumulation setting, where all batches of samples are available for estimation, we provide minimax bounds for the expected $\ell_2^2$ and $\ell_1$ losses at every stage. We show examples where our bounds match under mild conditions, and there is a strict gap with the corresponding oracle-assisted minimax loss where real and synthetic samples are differentiated. We also provide a lower bound on the minimax loss in the data replacement setting, where only the latest batch of samples is available, and use it to find a lower bound for the worst-case loss for bounded estimate trajectories.
2501.19274
GO: The Great Outdoors Multimodal Dataset
cs.RO
The Great Outdoors (GO) dataset is a multi-modal annotated data resource aimed at advancing ground robotics research in unstructured environments. This dataset provides the most comprehensive set of data modalities and annotations compared to existing off-road datasets. In total, the GO dataset includes six unique sensor types with high-quality semantic annotations and GPS traces to support tasks such as semantic segmentation, object detection, and SLAM. The diverse environmental conditions represented in the dataset present significant real-world challenges that provide opportunities to develop more robust solutions to support the continued advancement of field robotics, autonomous exploration, and perception systems in natural environments. The dataset can be downloaded at: https://www.unmannedlab.org/the-great-outdoors-dataset/
2501.19277
On Pareto Optimality for the Multinomial Logistic Bandit
stat.ML cs.LG
We provide a new online learning algorithm for tackling the Multinomial Logit Bandit (MNL-Bandit) problem. Despite the challenges posed by the combinatorial nature of the MNL model, we develop a novel Upper Confidence Bound (UCB)-based method that achieves Pareto optimality by balancing regret minimization and estimation error of the assortment revenues and the MNL parameters. We develop theoretical guarantees characterizing the tradeoff between regret and estimation error for the MNL-Bandit problem through information-theoretic bounds, and propose a modified UCB algorithm that incorporates forced exploration to improve parameter estimation accuracy while maintaining low regret. Our analysis sheds critical insights into how to optimally balance the collected revenues and the treatment estimation in dynamic assortment optimization.
2501.19278
Pheromone-based Learning of Optimal Reasoning Paths
cs.CL
Large Language Models (LLMs) have demonstrated remarkable reasoning capabilities through chain-of-thought prompting, yet discovering effective reasoning methods for complex problems remains challenging due to the vast space of possible intermediate steps. We introduce Ant Colony Optimization-guided Tree of Thought (ACO-ToT), a novel algorithm that combines ACO with LLMs to discover optimal reasoning paths for complex problems efficiently. Drawing inspiration from Hebbian learning in neurological systems, our method employs a collection of distinctly fine-tuned LLM "ants" to traverse and lay pheromone trails through a centralized tree of thought, with each ant's movement governed by a weighted combination of existing pheromone trails and its own specialized expertise. The algorithm evaluates complete reasoning paths using a mixture-of-experts-based scoring function, with pheromones reinforcing productive reasoning paths across iterations. Experiments on three challenging reasoning tasks (GSM8K, ARC-Challenge, and MATH) demonstrate that ACO-ToT performs significantly better than existing chain-of-thought optimization approaches, suggesting that incorporating biologically inspired collective search mechanisms into LLM inference can substantially enhance reasoning capabilities.
2501.19279
S-VOTE: Similarity-based Voting for Client Selection in Decentralized Federated Learning
cs.LG cs.DC
Decentralized Federated Learning (DFL) enables collaborative, privacy-preserving model training without relying on a central server. This decentralized approach reduces bottlenecks and eliminates single points of failure, enhancing scalability and resilience. However, DFL also introduces challenges such as suboptimal models with non-IID data distributions, increased communication overhead, and resource usage. Thus, this work proposes S-VOTE, a voting-based client selection mechanism that optimizes resource usage and enhances model performance in federations with non-IID data conditions. S-VOTE considers an adaptive strategy for spontaneous local training that addresses participation imbalance, allowing underutilized clients to contribute without significantly increasing resource costs. Extensive experiments on benchmark datasets demonstrate the S-VOTE effectiveness. More in detail, it achieves lower communication costs by up to 21%, 4-6% faster convergence, and improves local performance by 9-17% compared to baseline methods in some configurations, all while achieving a 14-24% energy consumption reduction. These results highlight the potential of S-VOTE to address DFL challenges in heterogeneous environments.
2501.19281
Statistical Physics of Deep Neural Networks: Generalization Capability, Beyond the Infinite Width, and Feature Learning
cond-mat.dis-nn cs.LG
Deep Neural Networks (DNNs) excel at many tasks, often rivaling or surpassing human performance. Yet their internal processes remain elusive, frequently described as "black boxes." While performance can be refined experimentally, achieving a fundamental grasp of their inner workings is still a challenge. Statistical Mechanics has long tackled computational problems, and this thesis applies physics-based insights to understand DNNs via three complementary approaches. First, by averaging over data, we derive an asymptotic bound on generalization that depends solely on the size of the last layer, rather than on the total number of parameters -- revealing how deep architectures process information differently across layers. Second, adopting a data-dependent viewpoint, we explore a finite-width thermodynamic limit beyond the infinite-width regime. This leads to: (i) a closed-form expression for the generalization error in a finite-width one-hidden-layer network (regression task); (ii) an approximate partition function for deeper architectures; and (iii) a link between deep networks in this thermodynamic limit and Student's t-processes. Finally, from a task-explicit perspective, we present a preliminary analysis of how DNNs interact with a controlled dataset, investigating whether they truly internalize its structure -- collapsing to the teacher -- or merely memorize it. By understanding when a network must learn data structure rather than just memorize, it sheds light on fostering meaningful internal representations. In essence, this thesis leverages the synergy between Statistical Physics and Machine Learning to illuminate the inner behavior of DNNs.
2501.19283
Application of Generative Adversarial Network (GAN) for Synthetic Training Data Creation to improve performance of ANN Classifier for extracting Built-Up pixels from Landsat Satellite Imagery
cs.CV cs.LG
Training a neural network for pixel based classification task using low resolution Landsat images is difficult as the size of the training data is usually small due to less number of available pixels that represent a single class without any mixing with other classes. Due to this scarcity of training data, neural network may not be able to attain expected level of accuracy. This limitation could be overcome using a generative network that aims to generate synthetic data having the same distribution as the sample data with which it is trained. In this work, we have proposed a methodology for improving the performance of ANN classifier to identify built-up pixels in the Landsat$7$ image with the help of developing a simple GAN architecture that could generate synthetic training pixels when trained using original set of sample built-up pixels. To ensure that the marginal and joint distributions of all the bands corresponding to the generated and original set of pixels are indistinguishable, non-parametric Kolmogorov Smirnov Test and Ball Divergence based Equality of Distributions Test have been performed respectively. It has been observed that the overall accuracy and kappa coefficient of the ANN model for built-up classification have continuously improved from $0.9331$ to $0.9983$ and $0.8277$ to $0.9958$ respectively, with the inclusion of generated sets of built-up pixels to the original one.
2501.19285
OneBatchPAM: A Fast and Frugal K-Medoids Algorithm
cs.LG
This paper proposes a novel k-medoids approximation algorithm to handle large-scale datasets with reasonable computational time and memory complexity. We develop a local-search algorithm that iteratively improves the medoid selection based on the estimation of the k-medoids objective. A single batch of size m << n provides the estimation, which reduces the required memory size and the number of pairwise dissimilarities computations to O(mn), instead of O(n^2) compared to most k-medoids baselines. We obtain theoretical results highlighting that a batch of size m = O(log(n)) is sufficient to guarantee, with strong probability, the same performance as the original local-search algorithm. Multiple experiments conducted on real datasets of various sizes and dimensions show that our algorithm provides similar performances as state-of-the-art methods such as FasterPAM and BanditPAM++ with a drastically reduced running time.
2501.19287
Differentially Private In-context Learning via Sampling Few-shot Mixed with Zero-shot Outputs
cs.LG
In-context learning (ICL) has shown promising improvement in downstream task adaptation of LLMs by augmenting prompts with relevant input-output examples (demonstrations). However, the ICL demonstrations can contain privacy-sensitive information, which can be leaked and/or regurgitated by the LLM output. Differential Privacy (DP), a widely adopted privacy safeguard, has emerged to mitigate this privacy leakage, with recent work demonstrating strong privacy-utility tradeoffs in classification tasks for ICL. However, generation tasks for ICL are challenging due to the high-dimensional output space of open-ended generation. To this end, we propose $\texttt{dps-mozo}$, Differentially Private Sampling by Mixing One-shot with Zero-shot Outputs, a decoding framework that generates DP text by sampling from the product of multiple one-shot outputs mixed with a zero-shot output. This mixing effectively reduces the amount of information that can be leaked by each demonstration. By utilizing the inherent randomness in sampling from the mixed distributions, we can achieve DP without adding noise, thereby improving the privacy-utility tradeoff. Our experimental evaluations show $\texttt{dps-mozo}$ can achieve a strong privacy guarantee, $\epsilon=2$, with minimal utility degradation compared to non-private few-shot learning, $\textbf{0.3}$% ROUGE-L F1 score decrease on the SAMSum dataset with Gemma 2 2B.
2501.19297
Analysis of LLMs vs Human Experts in Requirements Engineering
cs.SE cs.AI
The majority of research around Large Language Models (LLM) application to software development has been on the subject of code generation. There is little literature on LLMs' impact on requirements engineering (RE), which deals with the process of developing and verifying the system requirements. Within RE, there is a subdiscipline of requirements elicitation, which is the practice of discovering and documenting requirements for a system from users, customers, and other stakeholders. In this analysis, we compare LLM's ability to elicit requirements of a software system, as compared to that of a human expert in a time-boxed and prompt-boxed study. We found LLM-generated requirements were evaluated as more aligned (+1.12) than human-generated requirements with a trend of being more complete (+10.2%). Conversely, we found users tended to believe that solutions they perceived as more aligned had been generated by human experts. Furthermore, while LLM-generated documents scored higher and performed at 720x the speed, their cost was, on average, only 0.06% that of a human expert. Overall, these findings indicate that LLMs will play an increasingly important role in requirements engineering by improving requirements definitions, enabling more efficient resource allocation, and reducing overall project timelines.