id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2502.00064
Evaluating Large Language Models in Vulnerability Detection Under Variable Context Windows
cs.CR cs.LG
This study examines the impact of tokenized Java code length on the accuracy and explicitness of ten major LLMs in vulnerability detection. Using chi-square tests and known ground truth, we found inconsistencies across models: some, like GPT-4, Mistral, and Mixtral, showed robustness, while others exhibited a significant link between tokenized length and performance. We recommend future LLM development focus on minimizing the influence of input length for better vulnerability detection. Additionally, preprocessing techniques that reduce token count while preserving code structure could enhance LLM accuracy and explicitness in these tasks.
2502.00065
Blood Glucose Level Prediction in Type 1 Diabetes Using Machine Learning
q-bio.QM cs.LG
Type 1 Diabetes is a chronic autoimmune condition in which the immune system attacks and destroys insulin-producing beta cells in the pancreas, resulting in little to no insulin production. Insulin helps glucose in your blood enter your muscle, fat, and liver cells so they can use it for energy or store it for later use. If insulin is insufficient, it causes sugar to build up in the blood and leads to serious health problems. People with Type 1 Diabetes need synthetic insulin every day. In diabetes management, continuous glucose monitoring is an important feature that provides near real-time blood glucose data. It is useful in deciding the synthetic insulin dose. In this research work, we used machine learning tools, deep neural networks, deep reinforcement learning, and voting and stacking regressors to predict blood glucose levels at 30-min time intervals using the latest DiaTrend dataset. Predicting blood glucose levels is useful in better diabetes management systems. The trained models were compared using several evaluation metrics. Our evaluation results demonstrate the performance of various models across different glycemic conditions for blood glucose prediction. The source codes of this work can be found in: https://github.com/soon-jynn-chu/t1d_bg_prediction
2502.00068
Privacy Preserving Charge Location Prediction for Electric Vehicles
cs.CR cs.AI
By 2050, electric vehicles (EVs) are projected to account for 70% of global vehicle sales. While EVs provide environmental benefits, they also pose challenges for energy generation, grid infrastructure, and data privacy. Current research on EV routing and charge management often overlooks privacy when predicting energy demands, leaving sensitive mobility data vulnerable. To address this, we developed a Federated Learning Transformer Network (FLTN) to predict EVs' next charge location with enhanced privacy measures. Each EV operates as a client, training an onboard FLTN model that shares only model weights, not raw data with a community-based Distributed Energy Resource Management System (DERMS), which aggregates them into a community global model. To further enhance privacy, non-transitory EVs use peer-to-peer weight sharing and augmentation within their community, obfuscating individual contributions and improving model accuracy. Community DERMS global model weights are then redistributed to EVs for continuous training. Our FLTN approach achieved up to 92% accuracy while preserving data privacy, compared to our baseline centralised model, which achieved 98% accuracy with no data privacy. Simulations conducted across diverse charge levels confirm the FLTN's ability to forecast energy demands over extended periods. We present a privacy-focused solution for forecasting EV charge location prediction, effectively mitigating data leakage risks.
2502.00070
Can AI Solve the Peer Review Crisis? A Large Scale Experiment on LLM's Performance and Biases in Evaluating Economics Papers
cs.CY cs.AI econ.GN q-fin.EC
We investigate whether artificial intelligence can address the peer review crisis in economics by analyzing 27,090 evaluations of 9,030 unique submissions using a large language model (LLM). The experiment systematically varies author characteristics (e.g., affiliation, reputation, gender) and publication quality (e.g., top-tier, mid-tier, low-tier, AI generated papers). The results indicate that LLMs effectively distinguish paper quality but exhibit biases favoring prominent institutions, male authors, and renowned economists. Additionally, LLMs struggle to differentiate high-quality AI-generated papers from genuine top-tier submissions. While LLMs offer efficiency gains, their susceptibility to bias necessitates cautious integration and hybrid peer review models to balance equity and accuracy.
2502.00072
LLM Cyber Evaluations Don't Capture Real-World Risk
cs.CR cs.AI cs.CL cs.LG
Large language models (LLMs) are demonstrating increasing prowess in cybersecurity applications, creating creating inherent risks alongside their potential for strengthening defenses. In this position paper, we argue that current efforts to evaluate risks posed by these capabilities are misaligned with the goal of understanding real-world impact. Evaluating LLM cybersecurity risk requires more than just measuring model capabilities -- it demands a comprehensive risk assessment that incorporates analysis of threat actor adoption behavior and potential for impact. We propose a risk assessment framework for LLM cyber capabilities and apply it to a case study of language models used as cybersecurity assistants. Our evaluation of frontier models reveals high compliance rates but moderate accuracy on realistic cyber assistance tasks. However, our framework suggests that this particular use case presents only moderate risk due to limited operational advantages and impact potential. Based on these findings, we recommend several improvements to align research priorities with real-world impact assessment, including closer academia-industry collaboration, more realistic modeling of attacker behavior, and inclusion of economic metrics in evaluations. This work represents an important step toward more effective assessment and mitigation of LLM-enabled cybersecurity risks.
2502.00074
SpikingRTNH: Spiking Neural Network for 4D Radar Object Detection
cs.CV cs.AI cs.NE
Recently, 4D Radar has emerged as a crucial sensor for 3D object detection in autonomous vehicles, offering both stable perception in adverse weather and high-density point clouds for object shape recognition. However, processing such high-density data demands substantial computational resources and energy consumption. We propose SpikingRTNH, the first spiking neural network (SNN) for 3D object detection using 4D Radar data. By replacing conventional ReLU activation functions with leaky integrate-and-fire (LIF) spiking neurons, SpikingRTNH achieves significant energy efficiency gains. Furthermore, inspired by human cognitive processes, we introduce biological top-down inference (BTI), which processes point clouds sequentially from higher to lower densities. This approach effectively utilizes points with lower noise and higher importance for detection. Experiments on K-Radar dataset demonstrate that SpikingRTNH with BTI significantly reduces energy consumption by 78% while achieving comparable detection performance to its ANN counterpart (51.1% AP 3D, 57.0% AP BEV). These results establish the viability of SNNs for energy-efficient 4D Radar-based object detection in autonomous driving systems. All codes are available at https://github.com/kaist-avelab/k-radar.
2502.00075
BTS: Harmonizing Specialized Experts into a Generalist LLM
cs.CL cs.LG
We present Branch-Train-Stitch (BTS), an efficient and flexible training algorithm for combining independently trained large language model (LLM) experts into a single, capable generalist model. Following Li et al., we start with a single seed language model which is branched into domain-specific (e.g., coding or math) experts with continual pretraining. BTS combines experts into a generalist model using lightweight stitch layers, which are inserted between frozen experts and the seed LLM, and trained on a small datamix of the expert domains. Stitch layers enable the seed LLM to integrate representations from any number of experts during the forward pass, allowing it to generalize to new domains, despite remaining frozen. Because BTS does not alter the constituent LLMs, BTS provides a modular and flexible approach: experts can be easily removed and new experts can be added with only a small amount of training. Compared to alternative model merging approaches, BTS yields the best generalist performance on a variety of downstream tasks, retaining the specialized capabilities of each of the experts.
2502.00076
Influence of color correction on pathology detection in Capsule Endoscopy
cs.CV cs.AI cs.LG
Pathology detection in Wireless Capsule Endoscopy (WCE) using deep learning has been explored in the recent past. However, deep learning models can be influenced by the color quality of the dataset used to train them, impacting detection, segmentation and classification tasks. In this work, we evaluate the impact of color correction on pathology detection using two prominent object detection models: Retinanet and YOLOv5. We first generate two color corrected versions of a popular WCE dataset (i.e., SEE-AI dataset) using two different color correction functions. We then evaluate the performance of the Retinanet and YOLOv5 on the original and color corrected versions of the dataset. The results reveal that color correction makes the models generate larger bounding boxes and larger intersection areas with the ground truth annotations. Furthermore, color correction leads to an increased number of false positives for certain pathologies. However, these effects do not translate into a consistent improvement in performance metrics such as F1-scores, IoU, and AP50. The code is available at https://github.com/agossouema2011/WCE2024. Keywords: Wireless Capsule Endoscopy, Color correction, Retinanet, YOLOv5, Detection
2502.00077
Robot localization aided by quantum algorithms
cs.RO
Localization is a critical aspect of mobile robotics, enabling robots to navigate their environment efficiently and avoid obstacles. Current probabilistic localization methods, such as the Adaptive-Monte Carlo localization (AMCL) algorithm, are computationally intensive and may struggle with large maps or high-resolution sensor data. This paper explores the application of quantum computing in robotics, focusing on the use of Grover's search algorithm to improve the efficiency of localization in mobile robots. We propose a novel approach to utilize Grover's algorithm in a 2D map, enabling faster and more efficient localization. Despite the limitations of current physical quantum computers, our experimental results demonstrate a significant speedup over classical methods, highlighting the potential of quantum computing to improve robotic localization. This work bridges the gap between quantum computing and robotics, providing a practical solution for robotic localization and paving the way for future research in quantum robotics.
2502.00078
Deep Ensembling with Multimodal Image Fusion for Efficient Classification of Lung Cancer
eess.IV cs.CV
This study focuses on the classification of cancerous and healthy slices from multimodal lung images. The data used in the research comprises Computed Tomography (CT) and Positron Emission Tomography (PET) images. The proposed strategy achieves the fusion of PET and CT images by utilizing Principal Component Analysis (PCA) and an Autoencoder. Subsequently, a new ensemble-based classifier developed, Deep Ensembled Multimodal Fusion (DEMF), employing majority voting to classify the sample images under examination. Gradient-weighted Class Activation Mapping (Grad-CAM) employed to visualize the classification accuracy of cancer-affected images. Given the limited sample size, a random image augmentation strategy employed during the training phase. The DEMF network helps mitigate the challenges of scarce data in computer-aided medical image analysis. The proposed network compared with state-of-the-art networks across three publicly available datasets. The network outperforms others based on the metrics - Accuracy, F1-Score, Precision, and Recall. The investigation results highlight the effectiveness of the proposed network.
2502.00079
Advanced Assessment of Stroke in Retinal Fundus Imaging with Deep Multi-view Learning
eess.IV cs.CV
Stroke is globally a major cause of mortality and morbidity, and hence accurate and rapid diagnosis of stroke is valuable. Retinal fundus imaging reveals the known markers of elevated stroke risk in the eyes, which are retinal venular widening, arteriolar narrowing, and increased tortuosity. In contrast to other imaging techniques used for stroke diagnosis, the acquisition of fundus images is easy, non-invasive, fast, and inexpensive. Therefore, in this study, we propose a multi-view stroke network (MVS-Net) to detect stroke and transient ischemic attack (TIA) using retinal fundus images. Contrary to existing studies, our study proposes for the first time a solution to discriminate stroke and TIA with deep multi-view learning by proposing an end-to-end deep network, consisting of multi-view inputs of fundus images captured from both right and left eyes. Accordingly, the proposed MVS-Net defines representative features from fundus images of both eyes and determines the relation within their macula-centered and optic nerve head-centered views. Experiments performed on a dataset collected from stroke and TIA patients, in addition to healthy controls, show that the proposed framework achieves an AUC score of 0.84 for stroke and TIA detection.
2502.00083
CerraData-4MM: A multimodal benchmark dataset on Cerrado for land use and land cover classification
cs.CV eess.IV
The Cerrado faces increasing environmental pressures, necessitating accurate land use and land cover (LULC) mapping despite challenges such as class imbalance and visually similar categories. To address this, we present CerraData-4MM, a multimodal dataset combining Sentinel-1 Synthetic Aperture Radar (SAR) and Sentinel-2 MultiSpectral Imagery (MSI) with 10m spatial resolution. The dataset includes two hierarchical classification levels with 7 and 14 classes, respectively, focusing on the diverse Bico do Papagaio ecoregion. We highlight CerraData-4MM's capacity to benchmark advanced semantic segmentation techniques by evaluating a standard U-Net and a more sophisticated Vision Transformer (ViT) model. The ViT achieves superior performance in multimodal scenarios, with the highest macro F1-score of 57.60% and a mean Intersection over Union (mIoU) of 49.05% at the first hierarchical level. Both models struggle with minority classes, particularly at the second hierarchical level, where U-Net's performance drops to an F1-score of 18.16%. Class balancing improves representation for underrepresented classes but reduces overall accuracy, underscoring the trade-off in weighted training. CerraData-4MM offers a challenging benchmark for advancing deep learning models to handle class imbalance and multimodal data fusion. Code, trained models, and data are publicly available at https://github.com/ai4luc/CerraData-4MM.
2502.00085
Efficient Beam Search for Large Language Models Using Trie-Based Decoding
cs.CL
In Transformer-based sequence-to-sequence generation, beam search has proven effective in enhancing the quality of generated sequences compared to greedy decoding. Conventional beam search methods typically adopt either a sequential or batch-based approach. The sequential approach, while memory-efficient, requires multiple decoding passes to construct a complete search tree, leading to significantly slower inference. On the other hand, the batch-based approach enables parallel computation across beams, but at the expense of high memory consumption due to the need to maintain separate key-value (KV) caches for each beam. In this study, we introduce a novel trie (prefix-tree)-based parallel decoding method that addresses the memory inefficiency of batch-based beam search. By sharing a single KV cache among all beams that share the same prefix, the proposed method not only reduces memory consumption dramatically but also enables parallel decoding across all branches. This innovative use of a prefix tree offers an efficient alternative for beam search, achieving significant memory savings while preserving inference speed, making it particularly well-suited for memory-constrained environments or large-scale model deployments.
2502.00088
Re-Visiting Explainable AI Evaluation Metrics to Identify The Most Informative Features
cs.LG stat.ML
Functionality or proxy-based approach is one of the used approaches to evaluate the quality of explainable artificial intelligence methods. It uses statistical methods, definitions and new developed metrics for the evaluation without human intervention. Among them, Selectivity or RemOve And Retrain (ROAR), and Permutation Importance (PI) are the most commonly used metrics to evaluate the quality of explainable artificial intelligence methods to highlight the most significant features in machine learning models. They state that the model performance should experience a sharp reduction if the most informative feature is removed from the model or permuted. However, the efficiency of both metrics is significantly affected by multicollinearity, number of significant features in the model and the accuracy of the model. This paper shows with empirical examples that both metrics suffer from the aforementioned limitations. Accordingly, we propose expected accuracy interval (EAI), a metric to predict the upper and lower bounds of the the accuracy of the model when ROAR or IP is implemented. The proposed metric found to be very useful especially with collinear features.
2502.00089
Ensembles of Low-Rank Expert Adapters
cs.CL cs.AI cs.LG
The training and fine-tuning of large language models (LLMs) often involve diverse textual data from multiple sources, which poses challenges due to conflicting gradient directions, hindering optimization and specialization. These challenges can undermine model generalization across tasks, resulting in reduced downstream performance. Recent research suggests that fine-tuning LLMs on carefully selected, task-specific subsets of data can match or even surpass the performance of using the entire dataset. Building on these insights, we propose the Ensembles of Low-Rank Expert Adapters (ELREA) framework to improve the model's capability to handle diverse tasks. ELREA clusters the training instructions based on their gradient directions, representing different areas of expertise and thereby reducing conflicts during optimization. Expert adapters are then trained on these clusters, utilizing the low-rank adaptation (LoRA) technique to ensure training efficiency and model scalability. During inference, ELREA combines predictions from the most relevant expert adapters based on the input data's gradient similarity to the training clusters, ensuring optimal adapter selection for each task. Experiments show that our method outperforms baseline LoRA adapters trained on the full dataset and other ensemble approaches with similar training and inference complexity across a range of domain-specific tasks.
2502.00090
Disambiguating Numeral Sequences to Decipher Ancient Accounting Corpora
cs.CL
A numeration system encodes abstract numeric quantities as concrete strings of written characters. The numeration systems used by modern scripts tend to be precise and unambiguous, but this was not so for the ancient and partially-deciphered proto-Elamite (PE) script, where written numerals can have up to four distinct readings depending on the system that is used to read them. We consider the task of disambiguating between these readings in order to determine the values of the numeric quantities recorded in this corpus. We algorithmically extract a list of possible readings for each PE numeral notation, and contribute two disambiguation techniques based on structural properties of the original documents and classifiers learned with the bootstrapping algorithm. We also contribute a test set for evaluating disambiguation techniques, as well as a novel approach to cautious rule selection for bootstrapped classifiers. Our analysis confirms existing intuitions about this script and reveals previously-unknown correlations between tablet content and numeral magnitude. This work is crucial to understanding and deciphering PE, as the corpus is heavily accounting-focused and contains many more numeric tokens than tokens of text.
2502.00094
AIN: The Arabic INclusive Large Multimodal Model
cs.CV cs.AI cs.CL cs.HC cs.LG
Amid the swift progress of large language models (LLMs) and their evolution into large multimodal models (LMMs), significant strides have been made in high-resource languages such as English and Chinese. While Arabic LLMs have seen notable progress, Arabic LMMs remain largely unexplored, often narrowly focusing on a few specific aspects of the language and visual understanding. To bridge this gap, we introduce AIN-the Arabic Inclusive Multimodal Model-designed to excel across diverse domains. AIN is an English-Arabic bilingual LMM designed to excel in English and Arabic, leveraging carefully constructed 3.6 million high-quality Arabic-English multimodal data samples. AIN demonstrates state-of-the-art Arabic performance, while also possessing strong English-language visual capabilities. On the recent CAMEL-Bench benchmark comprising 38 sub-domains including, multi-image understanding, complex visual perception, handwritten document understanding, video understanding, medical imaging, plant diseases, and remote sensing-based land use understanding, our AIN demonstrates strong performance with the 7B model outperforming GPT-4o by an absolute gain of 3.4% averaged over eight domains and 38 sub-domains. AIN's superior capabilities position it as a significant step toward empowering Arabic speakers with advanced multimodal generative AI tools across diverse applications.
2502.00108
Tracking Most Significant Shifts in Infinite-Armed Bandits
cs.LG stat.ML
We study an infinite-armed bandit problem where actions' mean rewards are initially sampled from a reservoir distribution. Most prior works in this setting focused on stationary rewards (Berry et al., 1997; Wang et al., 2008; Bonald and Proutiere, 2013; Carpentier and Valko, 2015) with the more challenging adversarial/non-stationary variant only recently studied in the context of rotting/decreasing rewards (Kim et al., 2022; 2024). Furthermore, optimal regret upper bounds were only achieved using parameter knowledge of non-stationarity and only known for certain regimes of regularity of the reservoir. This work shows the first parameter-free optimal regret bounds for all regimes while also relaxing distributional assumptions on the reservoir. We first introduce a blackbox scheme to convert a finite-armed MAB algorithm designed for near-stationary environments into a parameter-free algorithm for the infinite-armed non-stationary problem with optimal regret guarantees. We next study a natural notion of significant shift for this problem inspired by recent developments in finite-armed MAB (Suk & Kpotufe, 2022). We show that tighter regret bounds in terms of significant shifts can be adaptively attained by employing a randomized variant of elimination within our blackbox scheme. Our enhanced rates only depend on the rotting non-stationarity and thus exhibit an interesting phenomenon for this problem where rising rewards do not factor into the difficulty of non-stationarity.
2502.00112
SAGRAD: A Program for Neural Network Training with Simulated Annealing and the Conjugate Gradient Method
cs.LG cs.NE
SAGRAD (Simulated Annealing GRADient), a Fortran 77 program for computing neural networks for classification using batch learning, is discussed. Neural network training in SAGRAD is based on a combination of simulated annealing and M{\o}ller's scaled conjugate gradient algorithm, the latter a variation of the traditional conjugate gradient method, better suited for the nonquadratic nature of neural networks. Different aspects of the implementation of the training process in SAGRAD are discussed, such as the efficient computation of gradients and multiplication of vectors by Hessian matrices that are required by M{\o}ller's algorithm; the (re)initialization of weights with simulated annealing required to (re)start M{\o}ller's algorithm the first time and each time thereafter that it shows insufficient progress in reaching a possibly local minimum; and the use of simulated annealing when M{\o}ller's algorithm, after possibly making considerable progress, becomes stuck at a local minimum or flat area of weight space. Outlines of the scaled conjugate gradient algorithm, the simulated annealing procedure and the training process used in SAGRAD are presented together with results from running SAGRAD on two examples of training data.
2502.00114
Mobile Robot Navigation Using Hand-Drawn Maps: A Vision Language Model Approach
cs.RO cs.CV
Hand-drawn maps can be used to convey navigation instructions between humans and robots in a natural and efficient manner. However, these maps can often contain inaccuracies such as scale distortions and missing landmarks which present challenges for mobile robot navigation. This paper introduces a novel Hand-drawn Map Navigation (HAM-Nav) architecture that leverages pre-trained vision language models (VLMs) for robot navigation across diverse environments, hand-drawing styles, and robot embodiments, even in the presence of map inaccuracies. HAM-Nav integrates a unique Selective Visual Association Prompting approach for topological map-based position estimation and navigation planning as well as a Predictive Navigation Plan Parser to infer missing landmarks. Extensive experiments were conducted in photorealistic simulated environments, using both wheeled and legged robots, demonstrating the effectiveness of HAM-Nav in terms of navigation success rates and Success weighted by Path Length. Furthermore, a user study in real-world environments highlighted the practical utility of hand-drawn maps for robot navigation as well as successful navigation outcomes.
2502.00115
A Direct Semi-Exhaustive Search Method for Robust, Partial-to-Full Point Cloud Registration
cs.RO cs.CV
Point cloud registration refers to the problem of finding the rigid transformation that aligns two given point clouds, and is crucial for many applications in robotics and computer vision. The main insight of this paper is that we can directly optimize the point cloud registration problem without correspondences by utilizing an algorithmically simple, yet computationally complex, semi-exhaustive search approach that is very well-suited for parallelization on modern GPUs. Our proposed algorithm, Direct Semi-Exhaustive Search (DSES), iterates over potential rotation matrices and efficiently computes the inlier-maximizing translation associated with each rotation. It then computes the optimal rigid transformation based on any desired distance metric by directly computing the error associated with each transformation candidate $\{R, t\}$. By leveraging the parallelism of modern GPUs, DSES outperforms state-of-the-art methods for partial-to-full point cloud registration on the simulated ModelNet40 benchmark and demonstrates high performance and robustness for pose estimation on a real-world robotics problem (https://youtu.be/q0q2-s2KSuA).
2502.00127
Sparse Autoencoder Insights on Voice Embeddings
cs.CL
Recent advances in explainable machine learning have highlighted the potential of sparse autoencoders in uncovering mono-semantic features in densely encoded embeddings. While most research has focused on Large Language Model (LLM) embeddings, the applicability of this technique to other domains remains largely unexplored. This study applies sparse autoencoders to speaker embeddings generated from a Titanet model, demonstrating the effectiveness of this technique in extracting mono-semantic features from non-textual embedded data. The results show that the extracted features exhibit characteristics similar to those found in LLM embeddings, including feature splitting and steering. The analysis reveals that the autoencoder can identify and manipulate features such as language and music, which are not evident in the original embedding. The findings suggest that sparse autoencoders can be a valuable tool for understanding and interpreting embedded data in many domains, including audio-based speaker recognition.
2502.00129
ProtoSnap: Prototype Alignment for Cuneiform Signs
cs.CV cs.LG
The cuneiform writing system served as the medium for transmitting knowledge in the ancient Near East for a period of over three thousand years. Cuneiform signs have a complex internal structure which is the subject of expert paleographic analysis, as variations in sign shapes bear witness to historical developments and transmission of writing and culture over time. However, prior automated techniques mostly treat sign types as categorical and do not explicitly model their highly varied internal configurations. In this work, we present an unsupervised approach for recovering the fine-grained internal configuration of cuneiform signs by leveraging powerful generative models and the appearance and structure of prototype font images as priors. Our approach, ProtoSnap, enforces structural consistency on matches found with deep image features to estimate the diverse configurations of cuneiform characters, snapping a skeleton-based template to photographed cuneiform signs. We provide a new benchmark of expert annotations and evaluate our method on this task. Our evaluation shows that our approach succeeds in aligning prototype skeletons to a wide variety of cuneiform signs. Moreover, we show that conditioning on structures produced by our method allows for generating synthetic data with correct structural configurations, significantly boosting the performance of cuneiform sign recognition beyond existing techniques, in particular over rare signs. Our code, data, and trained models are available at the project page: https://tau-vailab.github.io/ProtoSnap/
2502.00131
Middleman Bias in Advertising: Aligning Relevance of Keyphrase Recommendations with Search
cs.IR
E-commerce sellers are recommended keyphrases based on their inventory on which they advertise to increase buyer engagement (clicks/sales). Keyphrases must be pertinent to items; otherwise, it can result in seller dissatisfaction and poor targeting -- towards that end relevance filters are employed. In this work, we describe the shortcomings of training relevance filter models on biased click/sales signals. We re-conceptualize advertiser keyphrase relevance as interaction between two dynamical systems -- Advertising which produces the keyphrases and Search which acts as a middleman to reach buyers. We discuss the bias of search relevance systems (middleman bias) and the need to align advertiser keyphrases with search relevance signals. We also compare the performance of cross encoders and bi-encoders in modeling this alignment and the scalability of such a solution for sellers at eBay.
2502.00133
Exploring Transfer Learning for Deep Learning Polyp Detection in Colonoscopy Images Using YOLOv8
cs.CV cs.AI
Deep learning methods have demonstrated strong performance in objection tasks; however, their ability to learn domain-specific applications with limited training data remains a significant challenge. Transfer learning techniques address this issue by leveraging knowledge from pre-training on related datasets, enabling faster and more efficient learning for new tasks. Finding the right dataset for pre-training can play a critical role in determining the success of transfer learning and overall model performance. In this paper, we investigate the impact of pre-training a YOLOv8n model on seven distinct datasets, evaluating their effectiveness when transferred to the task of polyp detection. We compare whether large, general-purpose datasets with diverse objects outperform niche datasets with characteristics similar to polyps. In addition, we assess the influence of the size of the dataset on the efficacy of transfer learning. Experiments on the polyp datasets show that models pre-trained on relevant datasets consistently outperform those trained from scratch, highlighting the benefit of pre-training on datasets with shared domain-specific features.
2502.00136
A Three-Branch Checks-and-Balances Frameworkfor Context-Aware Ethical Alignment of Large Language Models
cs.CL cs.AI
This paper introduces a three-branch checks-and-balances framework for ethical alignment of Large Language Models (LLMs), inspired by governmental systems. It implements three independent yet interacting components: LLMs as the executive branch for knowledge generation, DIKE as the legislative branch establishing ethical guardrails, and ERIS as the judicial branch for contextual interpretation. The adversarial DIKE-ERIS duality enables adaptation to diverse cultural contexts while upholding consistent ethical principles. This architecture addresses limitations of reinforcement learning with human feedback (RLHF) by providing interpretable, adaptable, and culturally-aware ethical reasoning. Through self-supervised learning and adversarial testing, our framework demonstrates how emotional modeling can guide linguistic behaviors toward ethical outcomes while preserving independence across knowledge generation, ethical oversight, and contextual interpretation.
2502.00138
JustAct+: Justified and Accountable Actions in Policy-Regulated, Multi-Domain Data Processing
cs.LO cs.DC cs.MA cs.PL
Inter-organisational data exchange is regulated by norms originating from sources ranging from (inter)national laws, to processing agreements, and individual consent. Verifying norm compliance is complex because laws (e.g., GDPR) distribute responsibility and require accountability. Moreover, in some application domains (e.g., healthcare), privacy requirements extend the norms (e.g., patient consent). In contrast, existing solutions such as smart contracts, access- and usage-control assume policies to be public, or otherwise, statically partition policy information at the cost of accountability and flexibility. Instead, our framework prescribes how decentralised agents justify their actions with policy fragments that the agents autonomously create, gossip, and assemble. Crucially, the permission of actions is always reproducible by any observer, even with a partial view of all the dynamic policies. Actors can be sure that future auditors will confirm their permissions. Systems centralise control by (re)configuring externally synchronised agreements, the bases of all justifications. As a result, control is centralised only to the extent desired by the agents. In this paper, we define the JustAct framework, detail its implementation in a particular data-processing system, and design a suitable policy language based on logic programming. A case study reproduces Brane - an existing policy-regulated, inter-domain, medical data processing system - and serves to demonstrate and assess the qualities of the framework.
2502.00139
Beamforming with Joint Phase and Time Array: System Design, Prototyping and Performance
cs.IT eess.SP math.IT
Joint phase-time arrays (JPTA) is a new mmWave radio frequency front-end architecture constructed with appending time-delay elements to phase shifters for analog beamforming. JPTA allows the mmWave base station (BS) to form multiple frequency-dependent beams with a single RF chain, exploiting the extra degrees of freedom the time-delay elements offer. Without requiring extra power-hungry RF chains, a BS with JPTA can schedule multiple users in different directions in a frequency-division multiplexing (FDM) manner. A BS with JPTA achieves various advantages over the traditional analog beamforming system. Simulation results show that JPTA can bring significant system-level benefits, e.g., extending uplink throughput coverage by 100%. To realize these system benefits of JPTA, high-resolution delay elements with a wide delay dynamic range are essential. With newly developed delay elements, we demonstrate that a single TRX RF chain can serve four users in four different directions in the mmWave band.
2502.00140
Demystifying MPNNs: Message Passing as Merely Efficient Matrix Multiplication
cs.LG cs.AI cs.NE cs.SI
While Graph Neural Networks (GNNs) have achieved remarkable success, their design largely relies on empirical intuition rather than theoretical understanding. In this paper, we present a comprehensive analysis of GNN behavior through three fundamental aspects: (1) we establish that \textbf{$k$-layer} Message Passing Neural Networks efficiently aggregate \textbf{$k$-hop} neighborhood information through iterative computation, (2) analyze how different loop structures influence neighborhood computation, and (3) examine behavior across structure-feature hybrid and structure-only tasks. For deeper GNNs, we demonstrate that gradient-related issues, rather than just over-smoothing, can significantly impact performance in sparse graphs. We also analyze how different normalization schemes affect model performance and how GNNs make predictions with uniform node features, providing a theoretical framework that bridges the gap between empirical success and theoretical understanding.
2502.00145
Counting and Reasoning with Plans
cs.AI
Classical planning asks for a sequence of operators reaching a given goal. While the most common case is to compute a plan, many scenarios require more than that. However, quantitative reasoning on the plan space remains mostly unexplored. A fundamental problem is to count plans, which relates to the conditional probability on the plan space. Indeed, qualitative and quantitative approaches are well-established in various other areas of automated reasoning. We present the first study to quantitative and qualitative reasoning on the plan space. In particular, we focus on polynomially bounded plans. On the theoretical side, we study its complexity, which gives rise to rich reasoning modes. Since counting is hard in general, we introduce the easier notion of facets, which enables understanding the significance of operators. On the practical side, we implement quantitative reasoning for planning. Thereby, we transform a planning task into a propositional formula and use knowledge compilation to count different plans. This framework scales well to large plan spaces, while enabling rich reasoning capabilities such as learning pruning functions and explainable planning.
2502.00146
Multimodal MRI-Ultrasound AI for Prostate Cancer Detection Outperforms Radiologist MRI Interpretation: A Multi-Center Study
eess.IV cs.AI cs.CV
Pre-biopsy magnetic resonance imaging (MRI) is increasingly used to target suspicious prostate lesions. This has led to artificial intelligence (AI) applications improving MRI-based detection of clinically significant prostate cancer (CsPCa). However, MRI-detected lesions must still be mapped to transrectal ultrasound (TRUS) images during biopsy, which results in missing CsPCa. This study systematically evaluates a multimodal AI framework integrating MRI and TRUS image sequences to enhance CsPCa identification. The study included 3110 patients from three cohorts across two institutions who underwent prostate biopsy. The proposed framework, based on the 3D UNet architecture, was evaluated on 1700 test cases, comparing performance to unimodal AI models that use either MRI or TRUS alone. Additionally, the proposed model was compared to radiologists in a cohort of 110 patients. The multimodal AI approach achieved superior sensitivity (80%) and Lesion Dice (42%) compared to unimodal MRI (73%, 30%) and TRUS models (49%, 27%). Compared to radiologists, the multimodal model showed higher specificity (88% vs. 78%) and Lesion Dice (38% vs. 33%), with equivalent sensitivity (79%). Our findings demonstrate the potential of multimodal AI to improve CsPCa lesion targeting during biopsy and treatment planning, surpassing current unimodal models and radiologists; ultimately improving outcomes for prostate cancer patients.
2502.00151
A Comprehensive Review: Applicability of Deep Neural Networks in Business Decision Making and Market Prediction Investment
econ.GN cs.AI q-fin.EC
Big data, both in its structured and unstructured formats, have brought in unforeseen challenges in economics and business. How to organize, classify, and then analyze such data to obtain meaningful insights are the ever-going research topics for business leaders and academic researchers. This paper studies recent applications of deep neural networks in decision making in economical business and investment; especially in risk management, portfolio optimization, and algorithmic trading. Set aside limitation in data privacy and cross-market analysis, the article establishes that deep neural networks have performed remarkably in financial classification and prediction. Moreover, the study suggests that by compositing multiple neural networks, spanning different data type modalities, a more robust, efficient, and scalable financial prediction framework can be constructed.
2502.00156
ALBAR: Adversarial Learning approach to mitigate Biases in Action Recognition
cs.CV cs.CR
Bias in machine learning models can lead to unfair decision making, and while it has been well-studied in the image and text domains, it remains underexplored in action recognition. Action recognition models often suffer from background bias (i.e., inferring actions based on background cues) and foreground bias (i.e., relying on subject appearance), which can be detrimental to real-life applications such as autonomous vehicles or assisted living monitoring. While prior approaches have mainly focused on mitigating background bias using specialized augmentations, we thoroughly study both biases. We propose ALBAR, a novel adversarial training method that mitigates foreground and background biases without requiring specialized knowledge of the bias attributes. Our framework applies an adversarial cross-entropy loss to the sampled static clip (where all the frames are the same) and aims to make its class probabilities uniform using a proposed entropy maximization loss. Additionally, we introduce a gradient penalty loss for regularization against the debiasing process. We evaluate our method on established background and foreground bias protocols, setting a new state-of-the-art and strongly improving combined debiasing performance by over 12% on HMDB51. Furthermore, we identify an issue of background leakage in the existing UCF101 protocol for bias evaluation which provides a shortcut to predict actions and does not provide an accurate measure of the debiasing capability of a model. We address this issue by proposing more fine-grained segmentation boundaries for the actor, where our method also outperforms existing approaches. Project Page: https://joefioresi718.github.io/ALBAR_webpage/
2502.00158
Resolving Editing-Unlearning Conflicts: A Knowledge Codebook Framework for Large Language Model Updating
cs.CL
Large Language Models (LLMs) excel in natural language processing by encoding extensive human knowledge, but their utility relies on timely updates as knowledge evolves. Updating LLMs involves two key tasks simultaneously: unlearning to remove unwanted knowledge and editing to incorporate new information. Existing methods face two major challenges: ineffective knowledge storage (either too sparse or too dense) and task conflicts between editing and unlearning, as validated through our theoretical and experimental results. To address these issues, we propose LOKA, a conflict-free framework for LLM updating based on a knowledge codebook. During training, updated knowledge is stored in multiple codebook memories. To optimize knowledge storage, a similarity-aware knowledge mapping ensures that related knowledge pieces are clustered and allocated to the same memory. Additionally, LOKA resolves task conflicts by employing task-specific and multi-task memories guided by a conflict score. In the inference stage, LOKA retrieves the most relevant memory from the codebook and plugs it into the original LLM to apply the updated knowledge. A learning-based router controls codebook activation to further improve knowledge utilization. Extensive experiments demonstrate the effectiveness of LOKA in LLM knowledge updating tasks.
2502.00160
Improving Quality Control Of MRI Images Using Synthetic Motion Data
eess.IV cs.CV
MRI quality control (QC) is challenging due to unbalanced and limited datasets, as well as subjective scoring, which hinder the development of reliable automated QC systems. To address these issues, we introduce an approach that pretrains a model on synthetically generated motion artifacts before applying transfer learning for QC classification. This method not only improves the accuracy in identifying poor-quality scans but also reduces training time and resource requirements compared to training from scratch. By leveraging synthetic data, we provide a more robust and resource-efficient solution for QC automation in MRI, paving the way for broader adoption in diverse research settings.
2502.00162
Physics-informed Split Koopman Operators for Data-efficient Soft Robotic Simulation
cs.RO
Koopman operator theory provides a powerful data-driven technique for modeling nonlinear dynamical systems in a linear framework, in comparison to computationally expensive and highly nonlinear physics-based simulations. However, Koopman operator-based models for soft robots are very high dimensional and require considerable amounts of data to properly resolve. Inspired by physics-informed techniques from machine learning, we present a novel physics-informed Koopman operator identification method that improves simulation accuracy for small dataset sizes. Through Strang splitting, the method takes advantage of both continuous and discrete Koopman operator approximation to obtain information both from trajectory and phase space data. The method is validated on a tendon-driven soft robotic arm, showing orders of magnitude improvement over standard methods in terms of the shape error. We envision this method can significantly reduce the data requirement of Koopman operators for systems with partially known physical models, and thus reduce the cost of obtaining data.
2502.00168
Supervised Quadratic Feature Analysis: An Information Geometry Approach to Dimensionality Reduction
stat.ML cs.LG math.DG math.ST stat.TH
Supervised dimensionality reduction aims to map labeled data to a low-dimensional feature space while maximizing class discriminability. Despite the availability of methods for learning complex non-linear features (e.g. Deep Learning), there is an enduring demand for dimensionality reduction methods that learn linear features due to their interpretability, low computational cost, and broad applicability. However, there is a gap between methods that optimize linear separability (e.g. LDA), and more flexible but computationally expensive methods that optimize over arbitrary class boundaries (e.g. metric-learning methods). Here, we present Supervised Quadratic Feature Analysis (SQFA), a dimensionality reduction method for learning linear features that maximize the differences between class-conditional first- and second-order statistics, which allow for quadratic discrimination. SQFA exploits the information geometry of second-order statistics in the symmetric positive definite manifold. We show that SQFA features support quadratic discriminability in real-world problems. We also provide a theoretical link, based on information geometry, between SQFA and the Quadratic Discriminant Analysis (QDA) classifier.
2502.00172
Distribution-Specific Agnostic Conditional Classification With Halfspaces
cs.LG cs.CC stat.ML
We study ``selective'' or ``conditional'' classification problems under an agnostic setting. Classification tasks commonly focus on modeling the relationship between features and categories that captures the vast majority of data. In contrast to common machine learning frameworks, conditional classification intends to model such relationships only on a subset of the data defined by some selection rule. Most work on conditional classification either solves the problem in a realizable setting or does not guarantee the error is bounded compared to an optimal solution. In this work, we consider selective/conditional classification by sparse linear classifiers for subsets defined by halfspaces, and give both positive as well as negative results for Gaussian feature distributions. On the positive side, we present the first PAC-learning algorithm for homogeneous halfspace selectors with error guarantee $\bigO*{\sqrt{\mathrm{opt}}}$, where $\mathrm{opt}$ is the smallest conditional classification error over the given class of classifiers and homogeneous halfspaces. On the negative side, we find that, under cryptographic assumptions, approximating the conditional classification loss within a small additive error is computationally hard even under Gaussian distribution. We prove that approximating conditional classification is at least as hard as approximating agnostic classification in both additive and multiplicative form.
2502.00173
Lifting by Gaussians: A Simple, Fast and Flexible Method for 3D Instance Segmentation
cs.CV
We introduce Lifting By Gaussians (LBG), a novel approach for open-world instance segmentation of 3D Gaussian Splatted Radiance Fields (3DGS). Recently, 3DGS Fields have emerged as a highly efficient and explicit alternative to Neural Field-based methods for high-quality Novel View Synthesis. Our 3D instance segmentation method directly lifts 2D segmentation masks from SAM (alternately FastSAM, etc.), together with features from CLIP and DINOv2, directly fusing them onto 3DGS (or similar Gaussian radiance fields such as 2DGS). Unlike previous approaches, LBG requires no per-scene training, allowing it to operate seamlessly on any existing 3DGS reconstruction. Our approach is not only an order of magnitude faster and simpler than existing approaches; it is also highly modular, enabling 3D semantic segmentation of existing 3DGS fields without requiring a specific parametrization of the 3D Gaussians. Furthermore, our technique achieves superior semantic segmentation for 2D semantic novel view synthesis and 3D asset extraction results while maintaining flexibility and efficiency. We further introduce a novel approach to evaluate individually segmented 3D assets from 3D radiance field segmentation methods.
2502.00174
The role of positional encodings in the ARC benchmark
cs.AI cs.LG
The Abstraction and Reasoning Corpus challenges AI systems to perform abstract reasoning with minimal training data, a task intuitive for humans but demanding for machine learning models. Using CodeT5+ as a case study, we demonstrate how limitations in positional encoding hinder reasoning and impact performance. This work further examines the role of positional encoding across transformer architectures, highlighting its critical influence on models of varying sizes and configurations. Comparing several strategies, we find that while 2D positional encoding and Rotary Position Embedding offer competitive performance, 2D encoding excels in data-constrained scenarios, emphasizing its effectiveness for ARC tasks
2502.00177
Evaluating Deep Human-in-the-Loop Optimization for Retinal Implants Using Sighted Participants
cs.LG cs.CV cs.HC
Human-in-the-loop optimization (HILO) is a promising approach for personalizing visual prostheses by iteratively refining stimulus parameters based on user feedback. Previous work demonstrated HILO's efficacy in simulation, but its performance with human participants remains untested. Here we evaluate HILO using sighted participants viewing simulated prosthetic vision to assess its ability to optimize stimulation strategies under realistic conditions. Participants selected between phosphenes generated by competing encoders to iteratively refine a deep stimulus encoder (DSE). We tested HILO in three conditions: standard optimization, threshold misspecifications, and out-of-distribution parameter sampling. Participants consistently preferred HILO-generated stimuli over both a na\"ive encoder and the DSE alone, with log odds favoring HILO across all conditions. We also observed key differences between human and simulated decision-making, highlighting the importance of validating optimization strategies with human participants. These findings support HILO as a viable approach for adapting visual prostheses to individuals.
2502.00180
Designing Scheduling for Diffusion Models via Spectral Analysis
cs.LG stat.ML
Diffusion models (DMs) have emerged as powerful tools for modeling complex data distributions and generating realistic new samples. Over the years, advanced architectures and sampling methods have been developed to make these models practically usable. However, certain synthesis process decisions still rely on heuristics without a solid theoretical foundation. In our work, we offer a novel analysis of the DM's inference process, introducing a comprehensive frequency response perspective. Specifically, by relying on Gaussianity and shift-invariance assumptions, we present the inference process as a closed-form spectral transfer function, capturing how the generated signal evolves in response to the initial noise. We demonstrate how the proposed analysis can be leveraged for optimizing the noise schedule, ensuring the best alignment with the original dataset's characteristics. Our results lead to scheduling curves that are dependent on the frequency content of the data, offering a theoretical justification for some of the heuristics taken by practitioners.
2502.00182
Understanding Federated Learning from IID to Non-IID dataset: An Experimental Study
cs.LG cs.AI stat.ML
As privacy concerns and data regulations grow, federated learning (FL) has emerged as a promising approach for training machine learning models across decentralized data sources without sharing raw data. However, a significant challenge in FL is that client data are often non-IID (non-independent and identically distributed), leading to reduced performance compared to centralized learning. While many methods have been proposed to address this issue, their underlying mechanisms are often viewed from different perspectives. Through a comprehensive investigation from gradient descent to FL, and from IID to non-IID data settings, we find that inconsistencies in client loss landscapes primarily cause performance degradation in non-IID scenarios. From this understanding, we observe that existing methods can be grouped into two main strategies: (i) adjusting parameter update paths and (ii) modifying client loss landscapes. These findings offer a clear perspective on addressing non-IID challenges in FL and help guide future research in the field.
2502.00185
Optimal Coupled Sensor Placement and Path-Planning in Unknown Time-Varying Environments
eess.SY cs.SY
We address path-planning for a mobile agent to navigate in an unknown environment with minimum exposure to a spatially and temporally varying threat field. The threat field is estimated using pointwise noisy measurements from a mobile sensor network. For this problem, we present a new information gain measure for optimal sensor placement that quantifies reduction in uncertainty in the path cost rather than the environment state. This measure, which we call the context-relevant mutual information (CRMI), couples the sensor placement and path-planning problem. We propose an iterative coupled sensor configuration and path-planning (CSCP) algorithm. At each iteration, this algorithm places sensors to maximize CRMI, updates the threat estimate using new measurements, and recalculates the path with minimum expected exposure to the threat. The iterations converge when the path cost variance, which is an indicator of risk, reduces below a desired threshold. We show that CRMI is submodular, and therefore, greedy optimization provides near-optimal sensor placements while maintaining computational efficiency of the CSCP algorithm. Distance-based sensor reconfiguration costs are introduced in a modified CRMI measure, which we also show to be submodular. Through numerical simulations, we demonstrate that the principal advantage of this algorithm is that near-optimal low-variance paths are achieved using far fewer sensor measurements as compared to a standard sensor placement method.
2502.00186
Formalising Propositional Information via Implication Hypergraphs
math.LO cs.IT math.IT
This work introduces a framework for quantifying the information content of logical propositions through the use of implication hypergraphs. We posit that a proposition's informativeness is primarily determined by its relationships with other propositions -- specifically, the extent to which it implies or derives other propositions. To formalize this notion, we develop a framework based on implication hypergraphs, that seeks to capture these relationships. Within this framework, we define propositional information, derive some key properties, and illustrate the concept through examples. While the approach is broadly applicable, mathematical propositions emerge as an ideal domain for its application due to their inherently rich and interconnected structure. We provide several examples to illustrate this and subsequently discuss the limitations of the framework, along with suggestions for potential refinements.
2502.00190
On the Effectiveness of Random Weights in Graph Neural Networks
cs.LG stat.ML
Graph Neural Networks (GNNs) have achieved remarkable success across diverse tasks on graph-structured data, primarily through the use of learned weights in message passing layers. In this paper, we demonstrate that random weights can be surprisingly effective, achieving performance comparable to end-to-end training counterparts, across various tasks and datasets. Specifically, we show that by replacing learnable weights with random weights, GNNs can retain strong predictive power, while significantly reducing training time by up to 6$\times$ and memory usage by up to 3$\times$. Moreover, the random weights combined with our construction yield random graph propagation operators, which we show to reduce the problem of feature rank collapse in GNNs. These understandings and empirical results highlight random weights as a lightweight and efficient alternative, offering a compelling perspective on the design and training of GNN architectures.
2502.00193
Byzantine-Resilient Zero-Order Optimization for Communication-Efficient Heterogeneous Federated Learning
cs.LG cs.CR cs.DC stat.ML
We introduce CyBeR-0, a Byzantine-resilient federated zero-order optimization method that is robust under Byzantine attacks and provides significant savings in uplink and downlink communication costs. We introduce transformed robust aggregation to give convergence guarantees for general non-convex objectives under client data heterogeneity. Empirical evaluations for standard learning tasks and fine-tuning large language models show that CyBeR-0 exhibits stable performance with only a few scalars per-round communication cost and reduced memory requirements.
2502.00194
Physics-Informed Neural Network based Damage Identification for Truss Railroad Bridges
cs.LG cs.AI physics.comp-ph
Railroad bridges are a crucial component of the U.S. freight rail system, which moves over 40 percent of the nation's freight and plays a critical role in the economy. However, aging bridge infrastructure and increasing train traffic pose significant safety hazards and risk service disruptions. The U.S. rail network includes over 100,000 railroad bridges, averaging one every 1.4 miles of track, with steel bridges comprising over 50% of the network's total bridge length. Early identification and assessment of damage in these bridges remain challenging tasks. This study proposes a physics-informed neural network (PINN) based approach for damage identification in steel truss railroad bridges. The proposed approach employs an unsupervised learning approach, eliminating the need for large datasets typically required by supervised methods. The approach utilizes train wheel load data and bridge response during train crossing events as inputs for damage identification. The PINN model explicitly incorporates the governing differential equations of the linear time-varying (LTV) bridge-train system. Herein, this model employs a recurrent neural network (RNN) based architecture incorporating a custom Runge-Kutta (RK) integrator cell, designed for gradient-based learning. The proposed approach updates the bridge finite element model while also quantifying damage severity and localizing the affected structural members. A case study on the Calumet Bridge in Chicago, Illinois, with simulated damage scenarios, is used to demonstrate the model's effectiveness in identifying damage while maintaining low false-positive rates. Furthermore, the damage identification pipeline is designed to seamlessly integrate prior knowledge from inspections and drone surveys, also enabling context-aware updating and assessment of bridge's condition.
2502.00196
DermaSynth: Rich Synthetic Image-Text Pairs Using Open Access Dermatology Datasets
cs.CV cs.AI cs.CL
A major barrier to developing vision large language models (LLMs) in dermatology is the lack of large image--text pairs dataset. We introduce DermaSynth, a dataset comprising of 92,020 synthetic image--text pairs curated from 45,205 images (13,568 clinical and 35,561 dermatoscopic) for dermatology-related clinical tasks. Leveraging state-of-the-art LLMs, using Gemini 2.0, we used clinically related prompts and self-instruct method to generate diverse and rich synthetic texts. Metadata of the datasets were incorporated into the input prompts by targeting to reduce potential hallucinations. The resulting dataset builds upon open access dermatological image repositories (DERM12345, BCN20000, PAD-UFES-20, SCIN, and HIBA) that have permissive CC-BY-4.0 licenses. We also fine-tuned a preliminary Llama-3.2-11B-Vision-Instruct model, DermatoLlama 1.0, on 5,000 samples. We anticipate this dataset to support and accelerate AI research in dermatology. Data and code underlying this work are accessible at https://github.com/abdurrahimyilmaz/DermaSynth.
2502.00197
Model Successor Functions
cs.LG stat.ML
The notion of generalization has moved away from the classical one defined in statistical learning theory towards an emphasis on out-of-domain generalization (OODG). Recently, there is a growing focus on inductive generalization, where a progression of difficulty implicitly governs the direction of domain shifts. In inductive generalization, it is often assumed that the training data lie in the easier side, while the testing data lie in the harder side. The challenge is that training data are always finite, but a learner is expected to infer an inductive principle that could be applied in an unbounded manner. This emerging regime has appeared in the literature under different names, such as length/logical/algorithmic extrapolation, but a formal definition is lacking. This work provides such a formalization that centers on the concept of model successors. Then we outline directions to adapt well-established techniques towards the learning of model successors. This work calls for restructuring of the research discussion around inductive generalization from fragmented task-centric communities to a more unified effort, focused on universal properties of learning and computation.
2502.00198
Fairshare Data Pricing for Large Language Models
cs.GT cs.CL
Training data is a pivotal resource for building large language models (LLMs), but unfair pricing in data markets poses a serious challenge for both data buyers (e.g., LLM builders) and sellers (e.g., human annotators), which discourages market participation, reducing data quantity and quality. In this paper, we propose a fairshare pricing framework that sets training data prices using data valuation methods to quantify their contribution to LLMs. In our framework, buyers make purchasing decisions using data valuation and sellers set prices to maximize their profits based on the anticipated buyer purchases. We theoretically show that pricing derived from our framework is tightly linked to data valuation and buyers' budget, optimal for both buyers and sellers. Through market simulations using current LLMs and datasets (math problems, medical diagnosis, and physical reasoning), we show that our framework is fairshare for buyers by ensuring their purchased data is reflective of model training value, leading to higher LLM task performances per-dollar spent on data, and fairshare for sellers by ensuring they sell their data at optimal prices. Our framework lays the foundation for future research on equitable and sustainable data markets for large-scale AI.
2502.00199
Optimal Construction of Data Injection Attacks on Process Systems
eess.SY cs.SY
An information-theoretic framework for constructing data injection attacks on process systems, from the attacker's standpoint, is studied. The attack construction aims to distract the stationary distributions of the process variables and stay stealthy, simultaneously. The problem is formulated as designing a multivariate Gaussian distribution to maximize the Kullback-Leibler divergence between the stationary distributions of states and state estimates under attacks and without attacks, while minimizing that between the distributions of sensor measurements. When the attacker has limited access to sensors, sparse attacks are proposed by incorporating a sparsity constraint on the attack. We conduct a theoretical analysis on the convexity of the attack construction problem and present a greedy algorithm, which allows for a systematic quantification of measurements' vulnerability of process systems. We numerically evaluate the performance of proposed constructions on a two-reactor process.
2502.00201
Year-over-Year Developments in Financial Fraud Detection via Deep Learning: A Systematic Literature Review
cs.LG cs.AI q-fin.ST
This paper systematically reviews advancements in deep learning (DL) techniques for financial fraud detection, a critical issue in the financial sector. Using the Kitchenham systematic literature review approach, 57 studies published between 2019 and 2024 were analyzed. The review highlights the effectiveness of various deep learning models such as Convolutional Neural Networks, Long Short-Term Memory, and transformers across domains such as credit card transactions, insurance claims, and financial statement audits. Performance metrics such as precision, recall, F1-score, and AUC-ROC were evaluated. Key themes explored include the impact of data privacy frameworks and advancements in feature engineering and data preprocessing. The study emphasizes challenges such as imbalanced datasets, model interpretability, and ethical considerations, alongside opportunities for automation and privacy-preserving techniques such as blockchain integration and Principal Component Analysis. By examining trends over the past five years, this review identifies critical gaps and promising directions for advancing DL applications in financial fraud detection, offering actionable insights for researchers and practitioners.
2502.00203
Reward-aware Preference Optimization: A Unified Mathematical Framework for Model Alignment
cs.LG cs.CL
The rapid development of large language model (LLM) alignment algorithms has resulted in a complex and fragmented landscape, with limited clarity on the effectiveness of different methods and their inter-connections. This paper introduces Reward-Aware Preference Optimization (RPO), a mathematical framework that unifies popular preference optimization techniques in LLM alignment, including DPO, IPO, SimPO, and REINFORCE (LOO), among others. RPO provides a structured approach to disentangle and systematically study the impact of various design choices, such as the optimization objective, the number of responses per prompt, and the use of implicit versus explicit reward models, on LLM preference optimization. We additionally propose a new experimental setup that enables the clean and direct ablation of such design choices. Through an extensive series of ablation studies within the RPO framework, we gain insights into the critical factors shaping model alignment, offering practical guidance on the most effective strategies for improving LLM alignment.
2502.00204
Nearly-Optimal Bandit Learning in Stackelberg Games with Side Information
cs.LG cs.GT
We study the problem of online learning in Stackelberg games with side information between a leader and a sequence of followers. In every round the leader observes contextual information and commits to a mixed strategy, after which the follower best-responds. We provide learning algorithms for the leader which achieve $O(T^{1/2})$ regret under bandit feedback, an improvement from the previously best-known rates of $O(T^{2/3})$. Our algorithms rely on a reduction to linear contextual bandits in the utility space: In each round, a linear contextual bandit algorithm recommends a utility vector, which our algorithm inverts to determine the leader's mixed strategy. We extend our algorithms to the setting in which the leader's utility function is unknown, and also apply it to the problems of bidding in second-price auctions with side information and online Bayesian persuasion with public and private states. Finally, we observe that our algorithms empirically outperform previous results on numerical simulations.
2502.00205
EcoWeedNet: A Lightweight and Automated Weed Detection Method for Sustainable Next-Generation Agricultural Consumer Electronics
cs.CV cs.AI
Sustainable agriculture plays a crucial role in ensuring world food security for consumers. A critical challenge faced by sustainable precision agriculture is weed growth, as weeds share essential resources with the crops, such as water, soil nutrients, and sunlight, which notably affect crop yields. The traditional methods employed to combat weeds include the usage of chemical herbicides and manual weed removal methods. However, these could damage the environment and pose health hazards. The adoption of automated computer vision technologies and ground agricultural consumer electronic vehicles in precision agriculture offers sustainable, low-carbon solutions. However, prior works suffer from issues such as low accuracy and precision and high computational expense. This work proposes EcoWeedNet, a novel model with enhanced weed detection performance without adding significant computational complexity, aligning with the goals of low-carbon agricultural practices. Additionally, our model is lightweight and optimal for deployment on ground-based consumer electronic agricultural vehicles and robots. The effectiveness of the proposed model is demonstrated through comprehensive experiments on the CottonWeedDet12 benchmark dataset reflecting real-world scenarios. EcoWeedNet achieves performance close to that of large models yet with much fewer parameters. (approximately 4.21% of the parameters and 6.59% of the GFLOPs of YOLOv4). This work contributes effectively to the development of automated weed detection methods for next-generation agricultural consumer electronics featuring lower energy consumption and lower carbon footprint. This work paves the way forward for sustainable agricultural consumer technologies.
2502.00206
BICompFL: Stochastic Federated Learning with Bi-Directional Compression
cs.LG cs.DC cs.IT math.IT stat.ML
We address the prominent communication bottleneck in federated learning (FL). We specifically consider stochastic FL, in which models or compressed model updates are specified by distributions rather than deterministic parameters. Stochastic FL offers a principled approach to compression, and has been shown to reduce the communication load under perfect downlink transmission from the federator to the clients. However, in practice, both the uplink and downlink communications are constrained. We show that bi-directional compression for stochastic FL has inherent challenges, which we address by introducing BICompFL. Our BICompFL is experimentally shown to reduce the communication cost by an order of magnitude compared to multiple benchmarks, while maintaining state-of-the-art accuracies. Theoretically, we study the communication cost of BICompFL through a new analysis of an importance-sampling based technique, which exposes the interplay between uplink and downlink communication costs.
2502.00208
Discovering Dataset Nature through Algorithmic Clustering based on String Compression
cs.IT math.IT
Text datasets can be represented using models that do not preserve text structure, or using models that preserve text structure. Our hypothesis is that depending on the dataset nature, there can be advantages using a model that preserves text structure over one that does not, and viceversa. The key is to determine the best way of representing a particular dataset, based on the dataset itself. In this work, we propose to investigate this problem by combining text distortion and algorithmic clustering based on string compression. Specifically, a distortion technique previously developed by the authors is applied to destroy text structure progressively. Following this, a clustering algorithm based on string compression is used to analyze the effects of the distortion on the information contained in the texts. Several experiments are carried out on text datasets and artificially-generated datasets. The results show that in strongly structural datasets the clustering results worsen as text structure is progressively destroyed. Besides, they show that using a compressor which enables the choice of the size of the left-context symbols helps to determine the nature of the datasets. Finally, the results are contrasted with a method based on multidimensional projections and analogous conclusions are obtained.
2502.00212
STP: Self-play LLM Theorem Provers with Iterative Conjecturing and Proving
cs.LG cs.AI cs.LO
A fundamental challenge in formal theorem proving by LLMs is the lack of high-quality training data. Although reinforcement learning or expert iteration partially mitigates this issue by alternating between LLM generating proofs and finetuning them on correctly generated ones, performance quickly plateaus due to the scarcity of correct proofs (sparse rewards). To keep improving the models with limited data, we draw inspiration from mathematicians, who continuously develop new results, partly by proposing novel conjectures or exercises (which are often variants of known results) and attempting to solve them. We design the Self-play Theorem Prover (STP) that simultaneously takes on two roles, conjecturer and prover, each providing training signals to the other. The conjecturer is trained iteratively on previously generated conjectures that are barely provable by the current prover, which incentivizes it to generate increasingly challenging conjectures over time. The prover attempts to prove the conjectures with standard expert iteration. We evaluate STP with both Lean and Isabelle formal versifiers. With 19.8 billion tokens generated during the training in Lean, STP proves 26.3% of the statements in the LeanWorkbook dataset, doubling the previous best result of 13.2% achieved through expert iteration. The final model achieves state-of-the-art performance among whole-proof generation methods on miniF2F-test (61.7%, pass@3200), Proofnet-test (23.1%, pass@3200) and PutnamBench (8/644, pass@3200).
2502.00213
Understanding Why Adam Outperforms SGD: Gradient Heterogeneity in Transformers
cs.LG cs.AI cs.NE
Transformer models are challenging to optimize with SGD and typically require adaptive optimizers such as Adam. However, the reasons behind the superior performance of Adam over SGD remain unclear. In this study, we investigate the optimization of transformer models by focusing on \emph{gradient heterogeneity}, defined as the disparity in gradient norms among parameters. Our analysis shows that gradient heterogeneity hinders gradient-based optimization, including SGD, while sign-based optimization, a simplified variant of Adam, is less affected. We further examine gradient heterogeneity in transformer models and show that it is influenced by the placement of layer normalization. Additionally, we show that the momentum term in sign-based optimization is important for preventing the excessive growth of linear-head parameters in tasks with many classes. Experimental results from fine-tuning transformer models in both NLP and vision domains validate our theoretical analyses. This study provides insights into the optimization challenges of transformer models and offers guidance for designing future optimization algorithms. Code is available at \url{https://github.com/tom4649/gradient-heterogeneity}.
2502.00215
Impulsive Relative Motion Control with Continuous-Time Constraint Satisfaction for Cislunar Space Missions
eess.SY cs.SY
Recent investments in cislunar applications open new frontiers for space missions within highly nonlinear dynamical regimes. In this paper, we propose a method based on Sequential Convex Programming (SCP) to loiter around a given target with impulsive actuation while satisfying path constraints continuously over the finite time-horizon, i.e., independently of the number of nodes in which domain is discretized. Location, timing, magnitude, and direction of a fixed number of impulses are optimized in a model predictive framework, exploiting the exact nonlinear dynamics of non-stationary orbital regimes. The proposed approach is validated on a relative orbiting problem with respect to a selenocentric Near Rectilinear Halo Orbit.
2502.00217
Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone
cs.LG cs.AI cs.CV
Balancing competing objectives remains a fundamental challenge in multi-task learning (MTL), primarily due to conflicting gradients across individual tasks. A common solution relies on computing a dynamic gradient update vector that balances competing tasks as optimization progresses. Building on this idea, we propose ConicGrad, a principled, scalable, and robust MTL approach formulated as a constrained optimization problem. Our method introduces an angular constraint to dynamically regulate gradient update directions, confining them within a cone centered on the reference gradient of the overall objective. By balancing task-specific gradients without over-constraining their direction or magnitude, ConicGrad effectively resolves inter-task gradient conflicts. Moreover, our framework ensures computational efficiency and scalability to high-dimensional parameter spaces. We conduct extensive experiments on standard supervised learning and reinforcement learning MTL benchmarks, and demonstrate that ConicGrad achieves state-of-the-art performance across diverse tasks.
2502.00219
Team Size and Its Negative Impact on the Disruption Index
cs.SI
As science transitions from the age of lone geniuses to an era of collaborative teams, the question of whether large teams can sustain the creativity of individuals and continue driving innovation has become increasingly important. Our previous research first revealed a negative relationship between team size and the Disruption Index-a network-based metric of innovation-by analyzing 65 million projects across papers, patents, and software over half a century. This work has sparked lively debates within the scientific community about the robustness of the Disruption Index in capturing the impact of team size on innovation. Here, we present additional evidence that the negative link between team size and disruption holds, even when accounting for factors such as reference length, citation impact, and historical time. We further show how a narrow 5-year window for measuring disruption can misrepresent this relationship as positive, underestimating the long-term disruptive potential of small teams. Like "sleeping beauties," small teams need a decade or more to see their transformative contributions to science.
2502.00220
Algorithmic Clustering based on String Compression to Extract P300 Structure in EEG Signals
cs.LG cs.IT eess.SP math.IT
P300 is an Event-Related Potential widely used in Brain-Computer Interfaces, but its detection is challenging due to inter-subject and temporal variability. This work introduces a clustering methodology based on Normalized Compression Distance (NCD) to extract the P300 structure, ensuring robustness against variability. We propose a novel signal-to-ASCII transformation to generate compression-friendly objects, which are then clustered using a hierarchical tree-based method and a multidimensional projection approach. Experimental results on two datasets demonstrate the method's ability to reveal relevant P300 structures, showing clustering performance comparable to state-of-the-art approaches. Furthermore, analysis at the electrode level suggests that the method could assist in electrode selection for P300 detection. This compression-driven clustering methodology offers a complementary tool for EEG analysis and P300 identification.
2502.00221
Social Robots as Social Proxies for Fostering Connection and Empathy Towards Humanity
cs.HC cs.RO
Despite living in an increasingly connected world, social isolation is a prevalent issue today. While social robots have been explored as tools to enhance social connection through companionship, their potential as asynchronous social platforms for fostering connection towards humanity has received less attention. In this work, we introduce the design of a social support companion that facilitates the exchange of emotionally relevant stories and scaffolds reflection to enhance feelings of connection via five design dimensions. We investigate how social robots can serve as "social proxies" facilitating human stories, passing stories from other human narrators to the user. To this end, we conduct a real-world deployment of 40 robot stations in users' homes over the course of two weeks. Through thematic analysis of user interviews, we find that social proxy robots can foster connection towards other people's experiences via mechanisms such as identifying connections across stories or offering diverse perspectives. We present design guidelines from our study insights on the use of social robot systems that serve as social platforms to enhance human empathy and connection.
2502.00222
The Free Termination Property of Queries Over Time
cs.DB cs.DC cs.PL
Building on prior work on distributed databases and the CALM Theorem, we define and study the question of free termination: in the absence of distributed coordination, what query properties allow nodes in a distributed (database) system to unilaterally terminate execution even though they may receive additional data or messages in the future? This completeness question is complementary to the soundness questions studied in the CALM literature. We also develop a new model based on semiautomata that allows us to bridge from the relational transducer model of the CALM papers to algebraic models that are popular among software engineers (e.g. CRDTs) and of increasing interest to database theory for datalog extensions and incremental view maintenance.
2502.00225
Should You Use Your Large Language Model to Explore or Exploit?
cs.LG cs.AI cs.CL
We evaluate the ability of the current generation of large language models (LLMs) to help a decision-making agent facing an exploration-exploitation tradeoff. We use LLMs to explore and exploit in silos in various (contextual) bandit tasks. We find that while the current LLMs often struggle to exploit, in-context mitigations may be used to substantially improve performance for small-scale tasks. However even then, LLMs perform worse than a simple linear regression. On the other hand, we find that LLMs do help at exploring large action spaces with inherent semantics, by suggesting suitable candidates to explore.
2502.00226
HackerRank-ASTRA: Evaluating Correctness & Consistency of Large Language Models on cross-domain multi-file project problems
cs.LG cs.SE
Evaluating the real-world applicability of large language models (LLMs) provides valuable insights for their development and use in software development tasks. Existing benchmarks often focus on standalone coding problems or specific libraries, overlooking multi-file, project-based scenarios and lacking a rigorous evaluation of consistency. The HackerRank-ASTRA Benchmark introduces project-based coding problems that mirror real-world scenarios. It evaluates model consistency through 32 runs (k = 32) and median standard deviation while incorporating taxonomy-level analysis to assess sub-skill capabilities. Initial evaluations on 65 problems show that the top three models -- o1, o1-preview, and Claude-3.5-Sonnet-1022 -- achieved comparable average scores of 75%, with no statistically significant differences in performance. Notably, Claude-3.5-Sonnet-1022 demonstrated the highest consistency across problems, with low variability (SD = 0.0497), which was statistically significant compared to other models, highlighting its reliability for real-world software development tasks.
2502.00227
AK-SLRL: Adaptive Krylov Subspace Exploration Using Single-Life Reinforcement Learning for Sparse Linear System
cs.CE
This paper presents a single-life reinforcement learning (SLRL) approach to adaptively select the dimension of the Krylov subspace during the generalized minimal residual (GMRES) iteration. GMRES is an iterative algorithm for solving large and sparse linear systems of equations in the form of \(Ax = b\) which are mainly derived from partial differential equations (PDEs). The proposed framework uses RL to adjust the Krylov subspace dimension (m) in the GMRES (m) algorithm. This research demonstrates that altering the dimension of the Krylov subspace in an online setup using SLRL can accelerate the convergence of the GMRES algorithm by more than an order of magnitude. A comparison of different matrix sizes and sparsity levels is performed to demonstrate the effectiveness of adaptive Krylov subspace exploration using single-life RL (AK-SLRL). We compare AK-SLRL with constant-restart GMRES by applying the highest restart value used in AK-SLRL to the GMRES method. The results show that using an adjustable restart parameter with single-life soft-actor critic (SLSAC) and an experience replay buffer sized to half the matrix dimension converges significantly faster than the constant restart GMRES with higher values. Higher values of the restart parameter are equivalent to a higher number of Arnoldi iterations to construct an orthonormal basis for the Krylov subspace $ K_m(A, r_0) $. This process includes constructing $m$ orthonormal vectors and updating the Hessenberg matrix $H$. Therefore, lower values of $m$ result in reduced computation needed in GMRES minimization to solve the least-squares problem in the smaller Hessenberg matrix. The robustness of the result is validated through a wide range of matrix dimensions and sparsity. This paper contributes to the series of RL combinations with numerical solvers to achieve accelerated scientific computing.
2502.00232
A Hybrid Random Forest and CNN Framework for Tile-Wise Oil-Water Classification in Hyperspectral Images
cs.CV cs.AI
A novel hybrid Random Forest and Convolutional Neural Network (CNN) framework is presented for oil-water classification in hyperspectral images (HSI). To address the challenge of preserving spatial context, the images were divided into smaller, non-overlapping tiles, which served as the basis for training, validation, and testing. Random Forest demonstrated strong performance in pixel-wise classification, outperforming models such as XGBoost, Attention-Based U-Net, and HybridSN. However, Random Forest loses spatial context, limiting its ability to fully exploit the spatial relationships in hyperspectral data. To improve performance, a CNN was trained on the probability maps generated by the Random Forest, leveraging the CNN's capacity to incorporate spatial context. The hybrid approach achieved 7.6% improvement in recall (to 0.85), 2.4% improvement in F1 score (to 0.84), and 0.54% improvement in AUC (to 0.99) compared to the baseline. These results highlight the effectiveness of combining probabilistic outputs with spatial feature learning for context-aware analysis of hyperspectral images.
2502.00233
Vision-Based Fuzzy Control System for Smart Walkers: Enhancing Usability for Stroke Survivors with Unilateral Upper Limb Impairments
cs.RO cs.SY eess.SY
Mobility impairments, particularly those caused by stroke-induced hemiparesis, significantly impact independence and quality of life. Current smart walker controllers operate by using input forces from the user to control linear motion and input torques to dictate rotational movement; however, because they predominantly rely on user-applied torque exerted on the device handle as an indicator of user intent to turn, they fail to adequately accommodate users with unilateral upper limb impairments. This leads to increased physical strain and cognitive load. This paper introduces a novel smart walker equipped with a fuzzy control algorithm that leverages shoulder abduction angles to intuitively interpret user intentions using just one functional hand. By integrating a force sensor and stereo camera, the system enhances walker responsiveness and usability. Experimental evaluations with five participants showed that the fuzzy controller outperformed the traditional admittance controller, reducing wrist torque while using the right hand to operate the walker by 12.65% for left turns, 80.36% for straight paths, and 81.16% for right turns. Additionally, average user comfort ratings on a Likert scale increased from 1 to 4. Results confirmed a strong correlation between shoulder abduction angles and directional intent, with users reporting decreased effort and enhanced ease of use. This study contributes to assistive robotics by providing an adaptable control mechanism for smart walkers, suggesting a pathway towards enhancing mobility and independence for individuals with mobility impairments.
2502.00234
Fast Solvers for Discrete Diffusion Models: Theory and Applications of High-Order Algorithms
cs.LG cs.CV cs.NA math.NA physics.comp-ph stat.ML
Discrete diffusion models have emerged as a powerful generative modeling framework for discrete data with successful applications spanning from text generation to image synthesis. However, their deployment faces challenges due to the high dimensionality of the state space, necessitating the development of efficient inference algorithms. Current inference approaches mainly fall into two categories: exact simulation and approximate methods such as $\tau$-leaping. While exact methods suffer from unpredictable inference time and redundant function evaluations, $\tau$-leaping is limited by its first-order accuracy. In this work, we advance the latter category by tailoring the first extension of high-order numerical inference schemes to discrete diffusion models, enabling larger step sizes while reducing error. We rigorously analyze the proposed schemes and establish the second-order accuracy of the $\theta$-trapezoidal method in KL divergence. Empirical evaluations on GPT-2 level text and ImageNet-level image generation tasks demonstrate that our method achieves superior sample quality compared to existing approaches under equivalent computational constraints.
2502.00240
Learning Difference-of-Convex Regularizers for Inverse Problems: A Flexible Framework with Theoretical Guarantees
stat.ML cs.LG eess.IV math.OC
Learning effective regularization is crucial for solving ill-posed inverse problems, which arise in a wide range of scientific and engineering applications. While data-driven methods that parameterize regularizers using deep neural networks have demonstrated strong empirical performance, they often result in highly nonconvex formulations that lack theoretical guarantees. Recent work has shown that incorporating structured nonconvexity into neural network-based regularizers, such as weak convexity, can strike a balance between empirical performance and theoretical tractability. In this paper, we demonstrate that a broader class of nonconvex functions, difference-of-convex (DC) functions, can yield improved empirical performance while retaining strong convergence guarantees. The DC structure enables the use of well-established optimization algorithms, such as the Difference-of-Convex Algorithm (DCA) and a Proximal Subgradient Method (PSM), which extend beyond standard gradient descent. Furthermore, we provide theoretical insights into the conditions under which optimal regularizers can be expressed as DC functions. Extensive experiments on computed tomography (CT) reconstruction tasks show that our approach achieves strong performance across sparse and limited-view settings, consistently outperforming other weakly supervised learned regularizers. Our code is available at \url{https://github.com/YasminZhang/ADCR}.
2502.00241
Mordal: Automated Pretrained Model Selection for Vision Language Models
cs.LG cs.AI cs.CL cs.CV
Incorporating multiple modalities into large language models (LLMs) is a powerful way to enhance their understanding of non-textual data, enabling them to perform multimodal tasks. Vision language models (VLMs) form the fastest growing category of multimodal models because of their many practical use cases, including in healthcare, robotics, and accessibility. Unfortunately, even though different VLMs in the literature demonstrate impressive visual capabilities in different benchmarks, they are handcrafted by human experts; there is no automated framework to create task-specific multimodal models. We introduce Mordal, an automated multimodal model search framework that efficiently finds the best VLM for a user-defined task without manual intervention. Mordal achieves this both by reducing the number of candidates to consider during the search process and by minimizing the time required to evaluate each remaining candidate. Our evaluation shows that Mordal can find the best VLM for a given problem using up to $8.9\times$--$11.6\times$ lower GPU hours than grid search. In the process of our evaluation, we have also discovered new VLMs that outperform their state-of-the-art counterparts.
2502.00242
Digital-Twin assisted Network Energy Optimization during Low Traffic Hours
cs.NI cs.SY eess.SY
As wireless network technology advances towards the sixth generation (6G), increasing network energy consumption has become a critical concern due to the growing demand for diverse services, radio deployments at various frequencies, larger bandwidths, and more antennas. Network operators must manage energy usage not only to reduce operational cost and improve revenue but also to minimize environmental impact by reducing the carbon footprint. The 3rd Generation Partnership Project (3GPP) has introduced several network energy savings (NES) features. However, the implementation details and system-level aspects of these features have not been thoroughly investigated. In this paper, we explore system-level resource optimization for network energy savings in low-traffic scenarios. We introduce multiple NES optimization formulations and strategies, and further analyze their performance using a detailed network digital twin. Our results demonstrate promising NES gains of up to 44%. Additionally, we provide practical considerations for implementing the proposed schemes and examine their impacts on user equipment (UE) operation.
2502.00245
Contrastive Private Data Synthesis via Weighted Multi-PLM Fusion
cs.LG
Substantial quantity and high quality are the golden rules of making a good training dataset with sample privacy protection equally important. Generating synthetic samples that resemble high-quality private data while ensuring Differential Privacy (DP), a formal privacy guarantee, promises scalability and practicality. However, existing methods relying on pre-trained models for data synthesis %that avoid fine-tuning large pre-trained generative models often struggle in data-deficient scenarios, suffering from limited sample size, inevitable generation noise and existing pre-trained model bias. To address these challenges, we propose a novel contrAstive private data Synthesis via Weighted multiple Pre-trained language models (PLM) framework, named as WASP. WASP utilizes limited private samples for more accurate private data distribution estimation via a Top-Q voting mechanism, and leverages low-quality synthetic samples for contrastive generation via collaboration among dynamically weighted multiple pre-trained models.Extensive experiments on 6 well-developed datasets with 6 open-source and 3 closed-source PLMs demonstrate the superiority of WASP in improving model performance over diverse downstream tasks. Code is available at https://anonymous.4open.science/r/WASP.
2502.00246
Context-Preserving Tensorial Reconfiguration in Large Language Model Training
cs.CL
Handling long-range dependencies in neural architectures has remained a persistent challenge due to computational limitations and inefficient contextual retention mechanisms. Tensorial operations have provided a foundation for restructuring model representations, yet conventional architectures have struggled to incorporate such techniques without introducing excessive complexity. A novel approach, Context-Preserving Tensorial Reconfiguration (CPTR), enables dynamic reorganization of weight tensors through structured factorization and adaptive contraction, allowing for enhanced contextual integration without substantial computational overhead. Empirical evaluations demonstrate that CPTR improves coherence retention across extended sequences, leading to measurable reductions in perplexity and improved recall accuracy for long-context tasks. Performance comparisons reveal that CPTR-enhanced models exhibit greater computational efficiency and reduced memory consumption while maintaining competitive language generation fluency and accuracy. Gradient stability metrics further validate the improved training efficiency, revealing more controlled variance in weight updates. Comparative studies across baseline and CPTR-enhanced models confirm that tensorial reconfiguration contributes to more stable and computationally efficient language modeling. The findings support the potential of CPTR in refining contemporary neural architectures for tasks requiring long-range contextual understanding and efficient memory utilization.
2502.00248
Provably-Stable Neural Network-Based Control of Nonlinear Systems
math.OC cs.LG cs.SY eess.SY
In recent years, Neural Networks (NNs) have been employed to control nonlinear systems due to their potential capability in dealing with situations that might be difficult for conventional nonlinear control schemes. However, to the best of our knowledge, the current literature on NN-based control lacks theoretical guarantees for stability and tracking performance. This precludes the application of NN-based control schemes to systems where stringent stability and performance guarantees are required. To address this gap, this paper proposes a systematic and comprehensive methodology to design provably-stable NN-based control schemes for affine nonlinear systems. Rigorous analysis is provided to show that the proposed approach guarantees stability of the closed-loop system with the NN in the loop. Also, it is shown that the resulting NN-based control scheme ensures that system states asymptotically converge to a neighborhood around the desired equilibrium point, with a tunable proximity threshold. The proposed methodology is validated and evaluated via simulation studies on an inverted pendulum and experimental studies on a Parrot Bebop 2 drone.
2502.00250
Transformer-Based Vector Font Classification Using Different Font Formats: TrueType versus PostScript
cs.CV
Modern fonts adopt vector-based formats, which ensure scalability without loss of quality. While many deep learning studies on fonts focus on bitmap formats, deep learning for vector fonts remains underexplored. In studies involving deep learning for vector fonts, the choice of font representation has often been made conventionally. However, the font representation format is one of the factors that can influence the computational performance of machine learning models in font-related tasks. Here we show that font representations based on PostScript outlines outperform those based on TrueType outlines in Transformer-based vector font classification. TrueType outlines represent character shapes as sequences of points and their associated flags, whereas PostScript outlines represent them as sequences of commands. In previous research, PostScript outlines have been predominantly used when fonts are treated as part of vector graphics, while TrueType outlines are mainly employed when focusing on fonts alone. Whether to use PostScript or TrueType outlines has been mainly determined by file format specifications and precedent settings in previous studies, rather than performance considerations. To date, few studies have compared which outline format provides better embedding representations. Our findings suggest that information aggregation is crucial in Transformer-based deep learning for vector graphics, as in tokenization in language models and patch division in bitmap-based image recognition models. This insight provides valuable guidance for selecting outline formats in future research on vector graphics.
2502.00253
Patch Triplet Similarity Purification for Guided Real-World Low-Dose CT Image Denoising
eess.IV cs.CV
Image denoising of low-dose computed tomography (LDCT) is an important problem for clinical diagnosis with reduced radiation exposure. Previous methods are mostly trained with pairs of synthetic or misaligned LDCT and normal-dose CT (NDCT) images. However, trained with synthetic noise or misaligned LDCT/NDCT image pairs, the denoising networks would suffer from blurry structure or motion artifacts. Since non-contrast CT (NCCT) images share the content characteristics to the corresponding NDCT images in a three-phase scan, they can potentially provide useful information for real-world LDCT image denoising. To exploit this aspect, in this paper, we propose to incorporate clean NCCT images as useful guidance for the learning of real-world LDCT image denoising networks. To alleviate the issue of spatial misalignment in training data, we design a new Patch Triplet Similarity Purification (PTSP) strategy to select highly similar patch (instead of image) triplets of LDCT, NDCT, and NCCT images for network training. Furthermore, we modify two image denoising transformers of SwinIR and HAT to accommodate the NCCT image guidance, by replacing vanilla self-attention with cross-attention. On our collected clinical dataset, the modified transformers trained with the data selected by our PTSP strategy show better performance than 15 comparison methods on real-world LDCT image denoising. Ablation studies validate the effectiveness of our NCCT image guidance and PTSP strategy. We will publicly release our data and code.
2502.00258
ProxSparse: Regularized Learning of Semi-Structured Sparsity Masks for Pretrained LLMs
cs.LG cs.CL
Large Language Models (LLMs) have demonstrated exceptional performance in natural language processing tasks, yet their massive size makes serving them inefficient and costly. Semi-structured pruning has emerged as an effective method for model acceleration, but existing approaches are suboptimal because they focus on local, layer-wise optimizations using heuristic rules, failing to leverage global feedback. We present ProxSparse, a learning-based framework for mask selection enabled by regularized optimization. ProxSparse transforms the rigid, non-differentiable mask selection process into a smoother optimization procedure, allowing gradual mask exploration with flexibility. ProxSparse does not involve additional weight updates once the mask is determined. Our extensive evaluations on 7 widely used models show that ProxSparse consistently outperforms previously proposed semi-structured mask selection methods with significant improvement, demonstrating the effectiveness of our learned approach towards semi-structured pruning.
2502.00262
INSIGHT: Enhancing Autonomous Driving Safety through Vision-Language Models on Context-Aware Hazard Detection and Edge Case Evaluation
cs.CV cs.AI
Autonomous driving systems face significant challenges in handling unpredictable edge-case scenarios, such as adversarial pedestrian movements, dangerous vehicle maneuvers, and sudden environmental changes. Current end-to-end driving models struggle with generalization to these rare events due to limitations in traditional detection and prediction approaches. To address this, we propose INSIGHT (Integration of Semantic and Visual Inputs for Generalized Hazard Tracking), a hierarchical vision-language model (VLM) framework designed to enhance hazard detection and edge-case evaluation. By using multimodal data fusion, our approach integrates semantic and visual representations, enabling precise interpretation of driving scenarios and accurate forecasting of potential dangers. Through supervised fine-tuning of VLMs, we optimize spatial hazard localization using attention-based mechanisms and coordinate regression techniques. Experimental results on the BDD100K dataset demonstrate a substantial improvement in hazard prediction straightforwardness and accuracy over existing models, achieving a notable increase in generalization performance. This advancement enhances the robustness and safety of autonomous driving systems, ensuring improved situational awareness and potential decision-making in complex real-world scenarios.
2502.00264
Beyond the Permutation Symmetry of Transformers: The Role of Rotation for Model Fusion
cs.LG cs.CV
Symmetry in the parameter space of deep neural networks (DNNs) has proven beneficial for various deep learning applications. A well-known example is the permutation symmetry in Multi-Layer Perceptrons (MLPs), where permuting the rows of weight matrices in one layer and applying the inverse permutation to adjacent layers yields a functionally equivalent model. While permutation symmetry fully characterizes the equivalence set for MLPs, its discrete nature limits its utility for transformers. In this paper, we introduce rotation symmetry, a novel form of parameter space symmetry for transformers that generalizes permutation symmetry by rotating parameter matrices in self-attention layers. Unlike permutation symmetry, rotation symmetry operates in a continuous domain, thereby significantly expanding the equivalence set for transformers. Based on this property, we propose a theoretically optimal parameter matching algorithm as a plug-and-play module to enhance model fusion. We evaluate our approach using pre-trained transformers across diverse natural language and vision tasks. Experimental results demonstrate that our rotation symmetry-based matching algorithm substantially improves model fusion, highlighting the potential of parameter space symmetry to facilitate model fusion. Our code is available on https://github.com/zhengzaiyi/RotationSymmetry.
2502.00265
RADx Data Hub: A Cloud Platform for FAIR, Harmonized COVID-19 Data
cs.DB
The COVID-19 pandemic highlighted the urgent need for robust systems to enable rapid data collection, integration, and analysis for public health responses. Existing approaches often relied on disparate, non-interoperable systems, creating bottlenecks in comprehensive analyses and timely decision-making. To address these challenges, the U.S. National Institutes of Health (NIH) launched the Rapid Acceleration of Diagnostics (RADx) initiative in 2020, with the RADx Data Hub, a centralized repository for de-identified and curated COVID-19 data, as its cornerstone. The RADx Data Hub hosts diverse study data, including clinical data, testing results, smart sensor outputs, self-reported symptoms, and information on social determinants of health. Built on cloud infrastructure, the RADx Data Hub integrates metadata standards, interoperable formats, and ontology-based tools to adhere to the FAIR (Findable, Accessible, Interoperable, Reusable) principles for data sharing. Initially developed for COVID-19 research, its architecture and processes are adaptable to other scientific disciplines. This paper provides an overview of the data hosted by the RADx Data Hub and describes the platform's capabilities and architecture.
2502.00266
MCM: Multi-layer Concept Map for Efficient Concept Learning from Masked Images
cs.CV cs.LG
Masking strategies commonly employed in natural language processing are still underexplored in vision tasks such as concept learning, where conventional methods typically rely on full images. However, using masked images diversifies perceptual inputs, potentially offering significant advantages in concept learning with large-scale Transformer models. To this end, we propose Multi-layer Concept Map (MCM), the first work to devise an efficient concept learning method based on masked images. In particular, we introduce an asymmetric concept learning architecture by establishing correlations between different encoder and decoder layers, updating concept tokens using backward gradients from reconstruction tasks. The learned concept tokens at various levels of granularity help either reconstruct the masked image patches by filling in gaps or guide the reconstruction results in a direction that reflects specific concepts. Moreover, we present both quantitative and qualitative results across a wide range of metrics, demonstrating that MCM significantly reduces computational costs by training on fewer than 75% of the total image patches while enhancing concept prediction performance. Additionally, editing specific concept tokens in the latent space enables targeted image generation from masked images, aligning both the visible contextual patches and the provided concepts. By further adjusting the testing time mask ratio, we could produce a range of reconstructions that blend the visible patches with the provided concepts, proportional to the chosen ratios.
2502.00270
DUET: Optimizing Training Data Mixtures via Feedback from Unseen Evaluation Tasks
cs.LG cs.AI stat.ML
The performance of a machine learning (ML) model depends heavily on the relevance of its training data to the domain of the downstream evaluation task. However, in practice, the data involved in an unseen evaluation task is often not known to us (e.g., conversations between an LLM and a user are end-to-end encrypted). So, it is not obvious what data would be relevant for training/fine-tuning the ML model to maximize its task performance. Instead, one can only deploy the ML model in the unseen evaluation task to gather multiple rounds of coarse feedback on how well the model has performed. This paper presents a novel global-to-local algorithm called DUET that can exploit the feedback loop by interleaving a data selection method with Bayesian optimization. As a result, DUET can efficiently refine the training data mixture from a pool of data domains to maximize the model's performance on the unseen evaluation task and its convergence to the optimal data mixture can be theoretically guaranteed by analyzing its cumulative regret. Empirical evaluation on image and LLM evaluation tasks shows that DUET finds better training data mixtures than conventional baselines.
2502.00271
Scaling Flaws of Verifier-Guided Search in Mathematical Reasoning
cs.CL
Large language models (LLMs) struggle with multi-step reasoning, where inference-time scaling has emerged as a promising strategy for performance improvement. Verifier-guided search outperforms repeated sampling when sample size is limited by selecting and prioritizing valid reasoning paths. However, we identify a critical limitation: scaling flaws, prevalent across different models (Mistral 7B and DeepSeekMath 7B), benchmarks (GSM8K and MATH), and verifiers (outcome value models and process reward models). As sample size increases, verifier-guided search exhibits diminishing advantages and eventually underperforms repeated sampling. Our analysis attributes this to verifier failures, where imperfect verifiers misrank candidates and erroneously prune all valid paths. These issues are further exacerbated in challenging and out-of-distribution problems, restricting search effectiveness. To mitigate verifier failures, we explore reducing reliance on verifiers and conduct preliminary investigations using two simple methods. Our findings reveal fundamental limitations in verifier-guided search and suggest future directions.
2502.00274
AoI in M/G/1/1 Queues with Probabilistic Preemption
cs.IT math.IT
We consider a status update system consisting of one source, one server, and one sink. The source generates packets according to a Poisson process and the packets are served according to a generally distributed service time. We consider a system with a capacity of one packet, i.e., there is no waiting buffer in the system, and model it as an M/G/1/1 queueing system. We introduce a probabilistically preemptive packet management policy and calculate the moment generating functions (MGFs) of the age of information (AoI) and peak AoI (PAoI) under the policy. According to the probabilistically preemptive policy, when a packet arrives, the possible packet in the system is replaced by the arriving packet with a fixed probability. Numerical results show the effectiveness of the packet management policy.
2502.00275
Simultaneous Estimation of Manipulation Skill and Hand Grasp Force from Forearm Ultrasound Images
cs.RO cs.CV cs.ET cs.HC
Accurate estimation of human hand configuration and the forces they exert is critical for effective teleoperation and skill transfer in robotic manipulation. A deeper understanding of human interactions with objects can further enhance teleoperation performance. To address this need, researchers have explored methods to capture and translate human manipulation skills and applied forces to robotic systems. Among these, biosignal-based approaches, particularly those using forearm ultrasound data, have shown significant potential for estimating hand movements and finger forces. In this study, we present a method for simultaneously estimating manipulation skills and applied hand force using forearm ultrasound data. Data collected from seven participants were used to train deep learning models for classifying manipulation skills and estimating grasp force. Our models achieved an average classification accuracy of 94.87 percent plus or minus 10.16 percent for manipulation skills and an average root mean square error (RMSE) of 0.51 plus or minus 0.19 Newtons for force estimation, as evaluated using five-fold cross-validation. These results highlight the effectiveness of forearm ultrasound in advancing human-machine interfacing and robotic teleoperation for complex manipulation tasks. This work enables new and effective possibilities for human-robot skill transfer and tele-manipulation, bridging the gap between human dexterity and robotic control.
2502.00277
Regularized Langevin Dynamics for Combinatorial Optimization
cs.LG stat.ML
This work proposes a simple yet effective sampling framework for combinatorial optimization (CO). Our method builds on discrete Langevin dynamics (LD), an efficient gradient-guided generative algorithm. However, we observed that directly applying LD often leads to limited exploration. To overcome this limitation, we propose the Regularized Langevin Dynamics (RLD), which enforces an expected distance between the sampled and current solutions, effectively avoiding local minima. We develop two CO solvers on top of RLD, one based on simulated annealing (SA) and the other one based on neural network (NN). Empirical results on three classical CO problems demonstrate that both of our methods can achieve comparable or better performance against the previous state-of-the-art (SOTA) SA and NN-based solvers. In particular, our SA algorithm reduces the running time of the previous SOTA SA method by up to 80\%, while achieving equal or superior performance. In summary, RLD offers a promising framework for enhancing both traditional heuristics and NN models to solve CO problems.
2502.00279
Improving realistic semi-supervised learning with doubly robust estimation
cs.LG stat.ML
A major challenge in Semi-Supervised Learning (SSL) is the limited information available about the class distribution in the unlabeled data. In many real-world applications this arises from the prevalence of long-tailed distributions, where the standard pseudo-label approach to SSL is biased towards the labeled class distribution and thus performs poorly on unlabeled data. Existing methods typically assume that the unlabeled class distribution is either known a priori, which is unrealistic in most situations, or estimate it on-the-fly using the pseudo-labels themselves. We propose to explicitly estimate the unlabeled class distribution, which is a finite-dimensional parameter, \emph{as an initial step}, using a doubly robust estimator with a strong theoretical guarantee; this estimate can then be integrated into existing methods to pseudo-label the unlabeled data during training more accurately. Experimental results demonstrate that incorporating our techniques into common pseudo-labeling approaches improves their performance.
2502.00280
On the study of frequency control and spectral bias in Wavelet-Based Kolmogorov Arnold networks: A path to physics-informed KANs
cs.LG cs.NA math.NA
Spectral bias, the tendency of neural networks to prioritize learning low-frequency components of functions during the initial training stages, poses a significant challenge when approximating solutions with high-frequency details. This issue is particularly pronounced in physics-informed neural networks (PINNs), widely used to solve differential equations that describe physical phenomena. In the literature, contributions such as Wavelet Kolmogorov Arnold Networks (Wav-KANs) have demonstrated promising results in capturing both low- and high-frequency components. Similarly, Fourier features (FF) are often employed to address this challenge. However, the theoretical foundations of Wav-KANs, particularly the relationship between the frequency of the mother wavelet and spectral bias, remain underexplored. A more in-depth understanding of how Wav-KANs manage high-frequency terms could offer valuable insights for addressing oscillatory phenomena encountered in parabolic, elliptic, and hyperbolic differential equations. In this work, we analyze the eigenvalues of the neural tangent kernel (NTK) of Wav-KANs to enhance their ability to converge on high-frequency components, effectively mitigating spectral bias. Our theoretical findings are validated through numerical experiments, where we also discuss the limitations of traditional approaches, such as standard PINNs and Fourier features, in addressing multi-frequency problems.
2502.00281
Sigmoid Self-Attention is Better than Softmax Self-Attention: A Mixture-of-Experts Perspective
cs.LG cs.AI
At the core of the popular Transformer architecture is the self-attention mechanism, which dynamically assigns softmax weights to each input token so that the model can focus on the most salient information. However, the softmax structure slows down the attention computation due to its row-wise nature, and inherently introduces competition among tokens: as the weight assigned to one token increases, the weights of others decrease. This competitive dynamic may narrow the focus of self-attention to a limited set of features, potentially overlooking other informative characteristics. Recent experimental studies have shown that using the element-wise sigmoid function helps eliminate token competition and reduce the computational overhead. Despite these promising empirical results, a rigorous comparison between sigmoid and softmax self-attention mechanisms remains absent in the literature. This paper closes this gap by theoretically demonstrating that sigmoid self-attention is more sample-efficient than its softmax counterpart. Toward that goal, we illustrate that each row of the self-attention matrix can be represented as a mixture of experts. Our analysis shows that ''experts'' in sigmoid self-attention require significantly less data to achieve the same approximation error as those in softmax self-attention. We corroborate our theoretical findings through extensive experiments on both synthetic and real-world datasets.
2502.00282
GraphMinNet: Learning Dependencies in Graphs with Light Complexity Minimal Architecture
cs.LG
Graph Neural Networks (GNNs) have demonstrated remarkable success in various applications, yet they often struggle to capture long-range dependencies (LRD) effectively. This paper introduces GraphMinNet, a novel GNN architecture that generalizes the idea of minimal Gated Recurrent Units to graph-structured data. Our approach achieves efficient LRD modeling with linear computational complexity while maintaining permutation equivariance and stability. The model incorporates both structural and positional information through a unique combination of feature and positional encodings, leading to provably stronger expressiveness than the 1-WL test. Theoretical analysis establishes that GraphMinNet maintains non-decaying gradients over long distances, ensuring effective long-range information propagation. Extensive experiments on ten diverse datasets, including molecular graphs, image graphs, and synthetic networks, demonstrate that GraphMinNet achieves state-of-the-art performance while being computationally efficient. Our results show superior performance on 6 out of 10 datasets and competitive results on the others, validating the effectiveness of our approach in capturing both local and global graph structures.
2502.00284
Bounded-Confidence Models of Multi-Dimensional Opinions with Topic-Weighted Discordance
physics.soc-ph cs.SI math.DS
People's opinions on a wide range of topics often evolve over time through their interactions with others. Models of opinion dynamics primarily focus on one-dimensional opinions which represent opinions on one topic. However, opinions on various topics are rarely isolated; instead, they can be interdependent and exhibit correlations. In a bounded-confidence model (BCM) of opinion dynamics, agents influence each other's opinions only if their opinions are sufficiently similar. We extend classical agent-based BCMs -- namely, the Hegeselmann--Krause BCM, which has synchronous interactions, and the Deffuant--Weisbuch BCM, which has asynchronous interactions -- to a multidimensional setting, in which the opinions are multidimensional vectors representing opinions of different topics and opinions on different topics are interdependent. To measure opinion differences between agents, we introduce topic-weighted discordance functions that account for opinion differences in all topics. We use the regions of receptiveness to characterize the steady-state opinion clusters and provide an analytical approach to compute these regions. In addition, we numerically simulate our models on various networks with initial opinions drawn from a variety of distributions. When initial opinions are correlated across different topics, our topic-weighted BCMs yield significantly different results in both transient and steady states compared to baseline models, where the dynamics of each opinion topic are independent.
2502.00285
K Nearest Neighbor-Guided Trajectory Similarity Learning
cs.LG cs.CV cs.DB
Trajectory similarity is fundamental to many spatio-temporal data mining applications. Recent studies propose deep learning models to approximate conventional trajectory similarity measures, exploiting their fast inference time once trained. Although efficient inference has been reported, challenges remain in similarity approximation accuracy due to difficulties in trajectory granularity modeling and in exploiting similarity signals in the training data. To fill this gap, we propose TSMini, a highly effective trajectory similarity model with a sub-view modeling mechanism capable of learning multi-granularity trajectory patterns and a k nearest neighbor-based loss that guides TSMini to learn not only absolute similarity values between trajectories but also their relative similarity ranks. Together, these two innovations enable highly accurate trajectory similarity approximation. Experiments show that TSMini can outperform the state-of-the-art models by 22% in accuracy on average when learning trajectory similarity measures.
2502.00288
Learning from Suboptimal Data in Continuous Control via Auto-Regressive Soft Q-Network
cs.LG cs.RO
Reinforcement learning (RL) for continuous control often requires large amounts of online interaction data. Value-based RL methods can mitigate this burden by offering relatively high sample efficiency. Some studies further enhance sample efficiency by incorporating offline demonstration data to "kick-start" training, achieving promising results in continuous control. However, they typically compute the Q-function independently for each action dimension, neglecting interdependencies and making it harder to identify optimal actions when learning from suboptimal data, such as non-expert demonstration and online-collected data during the training process. To address these issues, we propose Auto-Regressive Soft Q-learning (ARSQ), a value-based RL algorithm that models Q-values in a coarse-to-fine, auto-regressive manner. First, ARSQ decomposes the continuous action space into discrete spaces in a coarse-to-fine hierarchy, enhancing sample efficiency for fine-grained continuous control tasks. Next, it auto-regressively predicts dimensional action advantages within each decision step, enabling more effective decision-making in continuous control tasks. We evaluate ARSQ on two continuous control benchmarks, RLBench and D4RL, integrating demonstration data into online training. On D4RL, which includes non-expert demonstrations, ARSQ achieves an average $1.62\times$ performance improvement over SOTA value-based baseline. On RLBench, which incorporates expert demonstrations, ARSQ surpasses various baselines, demonstrating its effectiveness in learning from suboptimal online-collected data.
2502.00290
Estimating LLM Uncertainty with Logits
cs.CL cs.AI
In recent years, Large Language Models (LLMs) have seen remarkable advancements and have been extensively integrated across various fields. Despite their progress, LLMs are prone to hallucinations, producing responses that may not be dependable if the models lack sufficient grounding knowledge. To mitigate this issue, methods for estimating uncertainty have been adopted, with a focus on critical tokens as indicators of reliability. Nevertheless, probability-based approaches have shown limitations in assessing token-level reliability due to the erosion of evidence strength information acquired during training. In this paper, we introduce Logits-induced Token Uncertainty (LogU), a novel framework designed to estimate token-specific uncertainty in LLMs in real time, without the need for multiple sampling rounds. By leveraging evidence modeling for the implementation of LogU, we utilize the derived uncertainty measures to steer downstream tasks. Our experimental findings highlight the substantial effectiveness and potential of LogU, marking a significant advancement in addressing the challenge of model hallucinations.
2502.00294
On the Source Model Key Agreement Problem
cs.IT math.IT
We consider the source model key agreement problem involving two legitimate parties and an eavesdropper who observe n i.i.d. samples of X and Y and Z respectively. The best-known upper bound on the key capacity is characterized by an inf-max optimization problem that generally lacks a closed-form solution. In this paper, we solve the optimization for some class of sources, thereby providing simple expressions for the upper bound. We provide general conditions under which the upper bound reduces to I(X;Y). As an example, we consider the XOR setting in which X and Y are binary, and Z is the XOR of X and Y . The upper bound reduces to I(X;Y) for this source. Next, we conjecture that the rate I(X;Y) is not achievable for the XOR source, and provide some ideas that might be useful for developing a new upper bound on the source model problem.
2502.00298
The Price of Linear Time: Error Analysis of Structured Kernel Interpolation
cs.LG stat.ML
Structured Kernel Interpolation (SKI) (Wilson et al. 2015) helps scale Gaussian Processes (GPs) by approximating the kernel matrix via interpolation at inducing points, achieving linear computational complexity. However, it lacks rigorous theoretical error analysis. This paper bridges the gap: we prove error bounds for the SKI Gram matrix and examine the error's effect on hyperparameter estimation and posterior inference. We further provide a practical guide to selecting the number of inducing points under convolutional cubic interpolation: they should grow as $n^{d/3}$ for error control. Crucially, we identify two dimensionality regimes governing the trade-off between SKI Gram matrix spectral norm error and computational complexity. For $d \leq 3$, any error tolerance can achieve linear time for sufficiently large sample size. For $d > 3$, the error must increase with sample size to maintain linear time. Our analysis provides key insights into SKI's scalability-accuracy trade-offs, establishing precise conditions for achieving linear-time GP inference with controlled approximation error.