id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2502.04143
A data-driven two-microphone method for in-situ sound absorption measurements
cs.SD cs.LG eess.AS
This work presents a data-driven approach to estimating the sound absorption coefficient of an infinite porous slab using a neural network and a two-microphone measurement on a finite porous sample. A 1D-convolutional network predicts the sound absorption coefficient from the complex-valued transfer function between the sound pressure measured at the two microphone positions. The network is trained and validated with numerical data generated by a boundary element model using the Delany-Bazley-Miki model, demonstrating accurate predictions for various numerical samples. The method is experimentally validated with baffled rectangular samples of a fibrous material, where sample size and source height are varied. The results show that the neural network offers the possibility to reliably predict the in-situ sound absorption of a porous material using the traditional two-microphone method as if the sample were infinite. The normal-incidence sound absorption coefficient obtained by the network compares well with that obtained theoretically and in an impedance tube. The proposed method has promising perspectives for estimating the sound absorption coefficient of acoustic materials after installation and in realistic operational conditions.
2502.04144
HD-EPIC: A Highly-Detailed Egocentric Video Dataset
cs.CV
We present a validation dataset of newly-collected kitchen-based egocentric videos, manually annotated with highly detailed and interconnected ground-truth labels covering: recipe steps, fine-grained actions, ingredients with nutritional values, moving objects, and audio annotations. Importantly, all annotations are grounded in 3D through digital twinning of the scene, fixtures, object locations, and primed with gaze. Footage is collected from unscripted recordings in diverse home environments, making HDEPIC the first dataset collected in-the-wild but with detailed annotations matching those in controlled lab environments. We show the potential of our highly-detailed annotations through a challenging VQA benchmark of 26K questions assessing the capability to recognise recipes, ingredients, nutrition, fine-grained actions, 3D perception, object motion, and gaze direction. The powerful long-context Gemini Pro only achieves 38.5% on this benchmark, showcasing its difficulty and highlighting shortcomings in current VLMs. We additionally assess action recognition, sound recognition, and long-term video-object segmentation on HD-EPIC. HD-EPIC is 41 hours of video in 9 kitchens with digital twins of 413 kitchen fixtures, capturing 69 recipes, 59K fine-grained actions, 51K audio events, 20K object movements and 37K object masks lifted to 3D. On average, we have 263 annotations per minute of our unscripted videos.
2502.04153
UltraIF: Advancing Instruction Following from the Wild
cs.CL cs.AI
Instruction-following made modern large language models (LLMs) helpful assistants. However, the key to taming LLMs on complex instructions remains mysterious, for that there are huge gaps between models trained by open-source community and those trained by leading companies. To bridge the gap, we propose a simple and scalable approach UltraIF for building LLMs that can follow complex instructions with open-source data. UltraIF first decomposes real-world user prompts into simpler queries, constraints, and corresponding evaluation questions for the constraints. Then, we train an UltraComposer to compose constraint-associated prompts with evaluation questions. This prompt composer allows us to synthesize complicated instructions as well as filter responses with evaluation questions. In our experiment, for the first time, we successfully align LLaMA-3.1-8B-Base to catch up with its instruct version on 5 instruction-following benchmarks without any benchmark information, using only 8B model as response generator and evaluator. The aligned model also achieved competitive scores on other benchmarks. Moreover, we also show that UltraIF could further improve LLaMA-3.1-8B-Instruct through self-alignment, motivating broader use cases for the method. Our code will be available at https://github.com/kkk-an/UltraIF.
2502.04161
YOLOv4: A Breakthrough in Real-Time Object Detection
cs.CV
YOLOv4 achieved the best performance on the COCO dataset by combining advanced techniques for regression (bounding box positioning) and classification (object class identification) using the Darknet framework. To enhance accuracy and adaptability, it employs Cross mini-Batch Normalization, Cross-Stage-Partial-connections, Self-Adversarial-Training, and Weighted-Residual-Connections, as well as CIoU loss, Mosaic data augmentation, and DropBlock regularization. With Mosaic augmentation and multi-resolution training, YOLOv4 achieves superior detection in diverse scenarios, attaining 43.5\% AP (in contrast, 65.7\% AP50) on a Tesla V100 at ~65 frames per second, ensuring efficiency, affordability, and adaptability for real-world environments.
2502.04162
A Pseudo Markov-Chain Model and Time-Elapsed Measures of Mobility from Collective Data
stat.AP cs.LG cs.SI stat.ML
In this paper we develop a pseudo Markov-chain model to understand time-elapsed flows, over multiple intervals, from time and space aggregated collective inter-location trip data, given as a time-series. Building on the model, we develop measures of mobility that parallel those known for individual mobility data, such as the radius of gyration. We apply these measures to the NetMob 2024 Data Challenge data, and obtain interesting results that are consistent with published statistics and commuting patterns in cities. Besides building a new framework, we foresee applications of this approach to an improved understanding of human mobility in the context of environmental changes and sustainable development.
2502.04163
Multi-task Online Learning for Probabilistic Load Forecasting
stat.ML cs.LG
Load forecasting is essential for the efficient, reliable, and cost-effective management of power systems. Load forecasting performance can be improved by learning the similarities among multiple entities (e.g., regions, buildings). Techniques based on multi-task learning obtain predictions by leveraging consumption patterns from the historical load demand of multiple entities and their relationships. However, existing techniques cannot effectively assess inherent uncertainties in load demand or account for dynamic changes in consumption patterns. This paper proposes a multi-task learning technique for online and probabilistic load forecasting. This technique provides accurate probabilistic predictions for the loads of multiple entities by leveraging their dynamic similarities. The method's performance is evaluated using datasets that register the load demand of multiple entities and contain diverse and dynamic consumption patterns. The experimental results show that the proposed method can significantly enhance the effectiveness of current multi-task learning approaches across a wide variety of load consumption scenarios.
2502.04164
Efficient Distributed Optimization under Heavy-Tailed Noise
cs.LG
Distributed optimization has become the default training paradigm in modern machine learning due to the growing scale of models and datasets. To mitigate communication overhead, local updates are often applied before global aggregation, resulting in a nested optimization approach with inner and outer steps. However, heavy-tailed stochastic gradient noise remains a significant challenge, particularly in attention-based models, hindering effective training. In this work, we propose TailOPT, an efficient framework designed to address heavy-tailed noise by leveraging adaptive optimization or clipping techniques. We establish convergence guarantees for the TailOPT framework under heavy-tailed noise with potentially unbounded gradient variance and local updates. Among its variants, we highlight a memory and communication efficient instantiation which we call $Bi^2Clip$, which performs coordinate-wise clipping at both the inner and outer optimizers, achieving adaptive-like performance (e.g., Adam) without the cost of maintaining or transmitting additional gradient statistics. Empirically, TailOPT, including $Bi^2Clip$, demonstrates superior performance on several language tasks and models, outperforming state-of-the-art methods.
2502.04167
Making Sense of Touch: Unsupervised Shapelet Learning in Bag-of-words Sense
cs.LG cs.RO
This paper introduces NN-STNE, a neural network using t-distributed stochastic neighbor embedding (t-SNE) as a hidden layer to reduce input dimensions by mapping long time-series data into shapelet membership probabilities. A Gaussian kernel-based mean square error preserves local data structure, while K-means initializes shapelet candidates due to the non-convex optimization challenge. Unlike existing methods, our approach uses t-SNE to address crowding in low-dimensional space and applies L1-norm regularization to optimize shapelet length. Evaluations on the UCR dataset and an electrical component manipulation task, like switching on, demonstrate improved clustering accuracy over state-of-the-art feature-learning methods in robotics.
2502.04170
From Configuration-Space Clearance to Feature-Space Margin: Sample Complexity in Learning-Based Collision Detection
cs.RO
Motion planning is a central challenge in robotics, with learning-based approaches gaining significant attention in recent years. Our work focuses on a specific aspect of these approaches: using machine-learning techniques, particularly Support Vector Machines (SVM), to evaluate whether robot configurations are collision free, an operation termed ``collision detection''. Despite the growing popularity of these methods, there is a lack of theory supporting their efficiency and prediction accuracy. This is in stark contrast to the rich theoretical results of machine-learning methods in general and of SVMs in particular. Our work bridges this gap by analyzing the sample complexity of an SVM classifier for learning-based collision detection in motion planning. We bound the number of samples needed to achieve a specified accuracy at a given confidence level. This result is stated in terms relevant to robot motion-planning such as the system's clearance. Building on these theoretical results, we propose a collision-detection algorithm that can also provide statistical guarantees on the algorithm's error in classifying robot configurations as collision-free or not.
2502.04172
Archetypal Analysis for Binary Data
cs.LG cs.AI stat.ML
Archetypal analysis (AA) is a matrix decomposition method that identifies distinct patterns using convex combinations of the data points denoted archetypes with each data point in turn reconstructed as convex combinations of the archetypes. AA thereby forms a polytope representing trade-offs of the distinct aspects in the data. Most existing methods for AA are designed for continuous data and do not exploit the structure of the data distribution. In this paper, we propose two new optimization frameworks for archetypal analysis for binary data. i) A second order approximation of the AA likelihood based on the Bernoulli distribution with efficient closed-form updates using an active set procedure for learning the convex combinations defining the archetypes, and a sequential minimal optimization strategy for learning the observation specific reconstructions. ii) A Bernoulli likelihood based version of the principal convex hull analysis (PCHA) algorithm originally developed for least squares optimization. We compare these approaches with the only existing binary AA procedure relying on multiplicative updates and demonstrate their superiority on both synthetic and real binary data. Notably, the proposed optimization frameworks for AA can easily be extended to other data distributions providing generic efficient optimization frameworks for AA based on tailored likelihood functions reflecting the underlying data distribution.
2502.04173
Lexical Substitution is not Synonym Substitution: On the Importance of Producing Contextually Relevant Word Substitutes
cs.CL
Lexical Substitution is the task of replacing a single word in a sentence with a similar one. This should ideally be one that is not necessarily only synonymous, but also fits well into the surrounding context of the target word, while preserving the sentence's grammatical structure. Recent advances in Lexical Substitution have leveraged the masked token prediction task of Pre-trained Language Models to generate replacements for a given word in a sentence. With this technique, we introduce ConCat, a simple augmented approach which utilizes the original sentence to bolster contextual information sent to the model. Compared to existing approaches, it proves to be very effective in guiding the model to make contextually relevant predictions for the target word. Our study includes a quantitative evaluation, measured via sentence similarity and task performance. In addition, we conduct a qualitative human analysis to validate that users prefer the substitutions proposed by our method, as opposed to previous methods. Finally, we test our approach on the prevailing benchmark for Lexical Substitution, CoInCo, revealing potential pitfalls of the benchmark. These insights serve as the foundation for a critical discussion on the way in which Lexical Substitution is evaluated.
2502.04174
Dense Fixed-Wing Swarming using Receding-Horizon NMPC
cs.RO cs.SY eess.SY
In this paper, we present an approach for controlling a team of agile fixed-wing aerial vehicles in close proximity to one another. Our approach relies on receding-horizon nonlinear model predictive control (NMPC) to plan maneuvers across an expanded flight envelope to enable inter-agent collision avoidance. To facilitate robust collision avoidance and characterize the likelihood of inter-agent collisions, we compute a statistical bound on the probability of the system leaving a tube around the planned nominal trajectory. Finally, we propose a metric for evaluating highly dynamic swarms and use this metric to evaluate our approach. We successfully demonstrated our approach through both simulation and hardware experiments, and to our knowledge, this the first time close-quarters swarming has been achieved with physical aerobatic fixed-wing vehicles.
2502.04176
MRAMG-Bench: A BeyondText Benchmark for Multimodal Retrieval-Augmented Multimodal Generation
cs.LG cs.IR
Recent advancements in Retrieval-Augmented Generation (RAG) have shown remarkable performance in enhancing response accuracy and relevance by integrating external knowledge into generative models. However, existing RAG methods primarily focus on providing text-only answers, even in multimodal retrieval-augmented generation scenarios. In this work, we introduce the Multimodal Retrieval-Augmented Multimodal Generation (MRAMG) task, which aims to generate answers that combine both text and images, fully leveraging the multimodal data within a corpus. Despite the importance of this task, there is a notable absence of a comprehensive benchmark to effectively evaluate MRAMG performance. To bridge this gap, we introduce the MRAMG-Bench, a carefully curated, human-annotated dataset comprising 4,346 documents, 14,190 images, and 4,800 QA pairs, sourced from three categories: Web Data, Academic Papers, and Lifestyle. The dataset incorporates diverse difficulty levels and complex multi-image scenarios, providing a robust foundation for evaluating multimodal generation tasks. To facilitate rigorous evaluation, our MRAMG-Bench incorporates a comprehensive suite of both statistical and LLM-based metrics, enabling a thorough analysis of the performance of popular generative models in the MRAMG task. Besides, we propose an efficient multimodal answer generation framework that leverages both LLMs and MLLMs to generate multimodal responses. Our datasets are available at: https://huggingface.co/MRAMG.
2502.04180
Multi-agent Architecture Search via Agentic Supernet
cs.LG cs.CL cs.MA
Large Language Model (LLM)-empowered multi-agent systems extend the cognitive boundaries of individual agents through disciplined collaboration and interaction, while constructing these systems often requires labor-intensive manual designs. Despite the availability of methods to automate the design of agentic workflows, they typically seek to identify a static, complex, one-size-fits-all system, which, however, fails to dynamically allocate inference resources based on the difficulty and domain of each query. To address this challenge, we shift away from the pursuit of a monolithic agentic system, instead optimizing the \textbf{agentic supernet}, a probabilistic and continuous distribution of agentic architectures. We introduce MaAS, an automated framework that samples query-dependent agentic systems from the supernet, delivering high-quality solutions and tailored resource allocation (\textit{e.g.}, LLM calls, tool calls, token cost). Comprehensive evaluation across six benchmarks demonstrates that MaAS \textbf{(I)} requires only $6\sim45\%$ of the inference costs of existing handcrafted or automated multi-agent systems, \textbf{(II)} surpasses them by $0.54\%\sim11.82\%$, and \textbf{(III)} enjoys superior cross-dataset and cross-LLM-backbone transferability.
2502.04190
Compliant Beaded-String Jamming For Variable Stiffness Anthropomorphic Fingers
cs.RO
Achieving human-like dexterity in robotic grippers remains an open challenge, particularly in ensuring robust manipulation in uncertain environments. Soft robotic hands try to address this by leveraging passive compliance, a characteristic that is crucial to the adaptability of the human hand, to achieve more robust manipulation while reducing reliance on high-resolution sensing and complex control. Further improvements in terms of precision and postural stability in manipulation tasks are achieved through the integration of variable stiffness mechanisms, but these tend to lack residual compliance, be bulky and have slow response times. To address these limitations, this work introduces a Compliant Joint Jamming mechanism for anthropomorphic fingers that exhibits passive residual compliance and adjustable stiffness, while achieving a range of motion in line with that of human interphalangeal joints. The stiffness range provided by the mechanism is controllable from 0.48 Nm/rad to 1.95 Nm/rad (a 4x increase). Repeatability, hysteresis and stiffness were also characterized as a function of the jamming force. To demonstrate the importance of the passive residual compliance afforded by the proposed system, a peg-in-hole task was conducted, which showed a 60% higher success rate for a gripper integrating our joint design when compared to a rigid one.
2502.04192
PixFoundation: Are We Heading in the Right Direction with Pixel-level Vision Foundation Models?
cs.CV
Multiple works have emerged to push the boundaries on multi-modal large language models (MLLMs) towards pixel-level understanding. Such approaches have shown strong performance on benchmarks for referring expression segmentation and grounded conversation generation. The current trend in pixel-level MLLMs is to train with pixel-level grounding supervision on large-scale labelled data. However, we show that such MLLMs when evaluated on recent challenging vision centric benchmarks, exhibit a weak ability in visual question answering. Surprisingly, some of these methods even downgrade the grounding ability of MLLMs that were never trained with such supervision. In this work, we propose two novel challenging benchmarks and show that MLLMs without pixel-level grounding supervision can outperform the state of the art in such tasks when evaluating both the pixel-level grounding and visual question answering. We propose simple baselines to extract the grounding information that can be plugged into any MLLM, which we call as PixFoundation. More importantly, we study the research question of ``When does grounding emerge in MLLMs that are not trained with pixel-level grounding supervision?'' We show that grounding can coincide with object parts or location/appearance information. Code repository is at https://github.com/MSiam/PixFoundation/.
2502.04194
The Best Instruction-Tuning Data are Those That Fit
cs.CL cs.AI cs.LG
High-quality supervised fine-tuning (SFT) data are crucial for eliciting strong capabilities from pretrained large language models (LLMs). Typically, instructions are paired with multiple responses sampled from other LLMs, which are often out of the distribution of the target model to be fine-tuned. This, at scale, can lead to diminishing returns and even hurt the models' performance and robustness. We propose **GRAPE**, a novel SFT framework that accounts for the unique characteristics of the target model. For each instruction, it gathers responses from various LLMs and selects the one with the highest probability measured by the target model, indicating that it aligns most closely with the target model's pretrained distribution; it then proceeds with standard SFT training. We first evaluate GRAPE with a controlled experiment, where we sample various solutions for each question in UltraInteract from multiple models and fine-tune commonly used LMs like LLaMA3.1-8B, Mistral-7B, and Qwen2.5-7B on GRAPE-selected data. GRAPE significantly outperforms strong baselines, including distilling from the strongest model with an absolute gain of up to 13.8%, averaged across benchmarks, and training on 3x more data with a maximum performance improvement of 17.3%. GRAPE's strong performance generalizes to realistic settings. We experiment with the post-training data used for Tulu3 and Olmo-2. GRAPE outperforms strong baselines trained on 4.5 times more data by 6.1% and a state-of-the-art data selection approach by 3% on average performance. Remarkably, using 1/3 of the data and half the number of epochs, GRAPE enables LLaMA3.1-8B to surpass the performance of Tulu3-SFT by 3.5%.
2502.04195
Integration of Prior Knowledge into Direct Learning for Safe Control of Linear Systems
eess.SY cs.SY
This paper integrates prior knowledge into direct learning of safe controllers for linear uncertain systems under disturbances. To this end, we characterize the set of all closed-loop systems that can be explained by available prior knowledge of the system model and the disturbances. We leverage matrix zonotopes for data-based characterization of closed-loop systems and show that the explainability of closed-loop systems by prior knowledge can be formalized by adding an equality conformity constraint to the matrix zonotope. We then leverage the resulting constraint matrix zonotope and design safe controllers that conform with both data and prior knowledge. This is achieved by ensuring the inclusion of a constrained zonotope of all possible next states in a {\lambda}-scaled level set of the safe set. We consider both polytope and zonotope safe sets and provide set inclusion conditions using linear programming.
2502.04199
Expanding Training Data for Endoscopic Phenotyping of Eosinophilic Esophagitis
eess.IV cs.CV
Eosinophilic esophagitis (EoE) is a chronic esophageal disorder marked by eosinophil-dominated inflammation. Diagnosing EoE usually involves endoscopic inspection of the esophageal mucosa and obtaining esophageal biopsies for histologic confirmation. Recent advances have seen AI-assisted endoscopic imaging, guided by the EREFS system, emerge as a potential alternative to reduce reliance on invasive histological assessments. Despite these advancements, significant challenges persist due to the limited availability of data for training AI models - a common issue even in the development of AI for more prevalent diseases. This study seeks to improve the performance of deep learning-based EoE phenotype classification by augmenting our training data with a diverse set of images from online platforms, public datasets, and electronic textbooks increasing our dataset from 435 to 7050 images. We utilized the Data-efficient Image Transformer for image classification and incorporated attention map visualizations to boost interpretability. The findings show that our expanded dataset and model enhancements improved diagnostic accuracy, robustness, and comprehensive analysis, enhancing patient outcomes.
2502.04201
Safeguarding connected autonomous vehicle communication: Protocols, intra- and inter-vehicular attacks and defenses
cs.CR cs.CV cs.NI
The advancements in autonomous driving technology, coupled with the growing interest from automotive manufacturers and tech companies, suggest a rising adoption of Connected Autonomous Vehicles (CAVs) in the near future. Despite some evidence of higher accident rates in AVs, these incidents tend to result in less severe injuries compared to traditional vehicles due to cooperative safety measures. However, the increased complexity of CAV systems exposes them to significant security vulnerabilities, potentially compromising their performance and communication integrity. This paper contributes by presenting a detailed analysis of existing security frameworks and protocols, focusing on intra- and inter-vehicle communications. We systematically evaluate the effectiveness of these frameworks in addressing known vulnerabilities and propose a set of best practices for enhancing CAV communication security. The paper also provides a comprehensive taxonomy of attack vectors in CAV ecosystems and suggests future research directions for designing more robust security mechanisms. Our key contributions include the development of a new classification system for CAV security threats, the proposal of practical security protocols, and the introduction of use cases that demonstrate how these protocols can be integrated into real-world CAV applications. These insights are crucial for advancing secure CAV adoption and ensuring the safe integration of autonomous vehicles into intelligent transportation systems.
2502.04204
"Short-length" Adversarial Training Helps LLMs Defend "Long-length" Jailbreak Attacks: Theoretical and Empirical Evidence
cs.LG cs.CR stat.ML
Jailbreak attacks against large language models (LLMs) aim to induce harmful behaviors in LLMs through carefully crafted adversarial prompts. To mitigate attacks, one way is to perform adversarial training (AT)-based alignment, i.e., training LLMs on some of the most adversarial prompts to help them learn how to behave safely under attacks. During AT, the length of adversarial prompts plays a critical role in the robustness of aligned LLMs. This paper focuses on adversarial suffix jailbreak attacks and unveils that to defend against a jailbreak attack with an adversarial suffix of length $\Theta(M)$, it is enough to align LLMs on prompts with adversarial suffixes of length $\Theta(\sqrt{M})$. Theoretically, we analyze the adversarial in-context learning of linear transformers on linear regression tasks and prove a robust generalization bound for trained transformers. The bound depends on the term $\Theta(\sqrt{M_{\text{test}}}/M_{\text{train}})$, where $M_{\text{train}}$ and $M_{\text{test}}$ are the number of adversarially perturbed in-context samples during training and testing. Empirically, we conduct AT on popular open-source LLMs and evaluate their robustness against jailbreak attacks of different adversarial suffix lengths. Results confirm a positive correlation between the attack success rate and the ratio of the square root of the adversarial suffix during jailbreaking to the length during AT. Our findings show that it is practical to defend "long-length" jailbreak attacks via efficient "short-length" AT. The code is available at https://github.com/fshp971/adv-icl.
2502.04206
Ensuring Reliability via Hyperparameter Selection: Review and Advances
cs.LG cs.IT math.IT
Hyperparameter selection is a critical step in the deployment of artificial intelligence (AI) models, particularly in the current era of foundational, pre-trained, models. By framing hyperparameter selection as a multiple hypothesis testing problem, recent research has shown that it is possible to provide statistical guarantees on population risk measures attained by the selected hyperparameter. This paper reviews the Learn-Then-Test (LTT) framework, which formalizes this approach, and explores several extensions tailored to engineering-relevant scenarios. These extensions encompass different risk measures and statistical guarantees, multi-objective optimization, the incorporation of prior knowledge and dependency structures into the hyperparameter selection process, as well as adaptivity. The paper also includes illustrative applications for communication systems.
2502.04207
Enhanced Feature-based Image Stitching for Endoscopic Videos in Pediatric Eosinophilic Esophagitis
cs.CV
Video endoscopy represents a major advance in the investigation of gastrointestinal diseases. Reviewing endoscopy videos often involves frequent adjustments and reorientations to piece together a complete view, which can be both time-consuming and prone to errors. Image stitching techniques address this issue by providing a continuous and complete visualization of the examined area. However, endoscopic images, particularly those of the esophagus, present unique challenges. The smooth surface, lack of distinct feature points, and non-horizontal orientation complicate the stitching process, rendering traditional feature-based methods often ineffective for these types of images. In this paper, we propose a novel preprocessing pipeline designed to enhance endoscopic image stitching through advanced computational techniques. Our approach converts endoscopic video data into continuous 2D images by following four key steps: (1) keyframe selection, (2) image rotation adjustment to correct distortions, (3) surface unwrapping using polar coordinate transformation to generate a flat image, and (4) feature point matching enhanced by Adaptive Histogram Equalization for improved feature detection. We evaluate stitching quality through the assessment of valid feature point match pairs. Experiments conducted on 20 pediatric endoscopy videos demonstrate that our method significantly improves image alignment and stitching quality compared to traditional techniques, laying a robust foundation for more effective panoramic image creation.
2502.04210
Algorithmic causal structure emerging through compression
cs.LG cs.AI cs.CC cs.IT math.IT
We explore the relationship between causality, symmetry, and compression. We build on and generalize the known connection between learning and compression to a setting where causal models are not identifiable. We propose a framework where causality emerges as a consequence of compressing data across multiple environments. We define algorithmic causality as an alternative definition of causality when traditional assumptions for causal identifiability do not hold. We demonstrate how algorithmic causal and symmetric structures can emerge from minimizing upper bounds on Kolmogorov complexity, without knowledge of intervention targets. We hypothesize that these insights may also provide a novel perspective on the emergence of causality in machine learning models, such as large language models, where causal relationships may not be explicitly identifiable.
2502.04218
Sports and Women's Sports: Gender Bias in Text Generation with Olympic Data
cs.CL
Large Language Models (LLMs) have been shown to be biased in prior work, as they generate text that is in line with stereotypical views of the world or that is not representative of the viewpoints and values of historically marginalized demographic groups. In this work, we propose using data from parallel men's and women's events at the Olympic Games to investigate different forms of gender bias in language models. We define three metrics to measure bias, and find that models are consistently biased against women when the gender is ambiguous in the prompt. In this case, the model frequently retrieves only the results of the men's event with or without acknowledging them as such, revealing pervasive gender bias in LLMs in the context of athletics.
2502.04219
NLP-Based .NET CLR Event Logs Analyzer
cs.SE cs.AI
In this paper, we present a tool for analyzing .NET CLR event logs based on a novel method inspired by Natural Language Processing (NLP) approach. Our research addresses the growing need for effective monitoring and optimization of software systems through detailed event log analysis. We utilize a BERT-based architecture with an enhanced tokenization process customized to event logs. The tool, developed using Python, its libraries, and an SQLite database, allows both conducting experiments for academic purposes and efficiently solving industry-emerging tasks. Our experiments demonstrate the efficacy of our approach in compressing event sequences, detecting recurring patterns, and identifying anomalies. The trained model shows promising results, with a high accuracy rate in anomaly detection, which demonstrates the potential of NLP methods to improve the reliability and stability of software systems.
2502.04223
\'Eclair -- Extracting Content and Layout with Integrated Reading Order for Documents
cs.CV
Optical Character Recognition (OCR) technology is widely used to extract text from images of documents, facilitating efficient digitization and data retrieval. However, merely extracting text is insufficient when dealing with complex documents. Fully comprehending such documents requires an understanding of their structure -- including formatting, formulas, tables, and the reading order of multiple blocks and columns across multiple pages -- as well as semantic information for detecting elements like footnotes and image captions. This comprehensive understanding is crucial for downstream tasks such as retrieval, document question answering, and data curation for training Large Language Models (LLMs) and Vision Language Models (VLMs). To address this, we introduce \'Eclair, a general-purpose text-extraction tool specifically designed to process a wide range of document types. Given an image, \'Eclair is able to extract formatted text in reading order, along with bounding boxes and their corresponding semantic classes. To thoroughly evaluate these novel capabilities, we introduce our diverse human-annotated benchmark for document-level OCR and semantic classification. \'Eclair achieves state-of-the-art accuracy on this benchmark, outperforming other methods across key metrics. Additionally, we evaluate \'Eclair on established benchmarks, demonstrating its versatility and strength across several evaluation standards.
2502.04226
Keep It Light! Simplifying Image Clustering Via Text-Free Adapters
cs.CV cs.LG cs.NE stat.CO stat.ML
Many competitive clustering pipelines have a multi-modal design, leveraging large language models (LLMs) or other text encoders, and text-image pairs, which are often unavailable in real-world downstream applications. Additionally, such frameworks are generally complicated to train and require substantial computational resources, making widespread adoption challenging. In this work, we show that in deep clustering, competitive performance with more complex state-of-the-art methods can be achieved using a text-free and highly simplified training pipeline. In particular, our approach, Simple Clustering via Pre-trained models (SCP), trains only a small cluster head while leveraging pre-trained vision model feature representations and positive data pairs. Experiments on benchmark datasets including CIFAR-10, CIFAR-20, CIFAR-100, STL-10, ImageNet-10, and ImageNet-Dogs, demonstrate that SCP achieves highly competitive performance. Furthermore, we provide a theoretical result explaining why, at least under ideal conditions, additional text-based embeddings may not be necessary to achieve strong clustering performance in vision.
2502.04229
Dark Distillation: Backdooring Distilled Datasets without Accessing Raw Data
cs.CR cs.AI
Dataset distillation (DD) enhances training efficiency and reduces bandwidth by condensing large datasets into smaller synthetic ones. It enables models to achieve performance comparable to those trained on the raw full dataset and has become a widely adopted method for data sharing. However, security concerns in DD remain underexplored. Existing studies typically assume that malicious behavior originates from dataset owners during the initial distillation process, where backdoors are injected into raw datasets. In contrast, this work is the first to address a more realistic and concerning threat: attackers may intercept the dataset distribution process, inject backdoors into the distilled datasets, and redistribute them to users. While distilled datasets were previously considered resistant to backdoor attacks, we demonstrate that they remain vulnerable to such attacks. Furthermore, we show that attackers do not even require access to any raw data to inject the backdoors successfully. Specifically, our approach reconstructs conceptual archetypes for each class from the model trained on the distilled dataset. Backdoors are then injected into these archetypes to update the distilled dataset. Moreover, we ensure the updated dataset not only retains the backdoor but also preserves the original optimization trajectory, thus maintaining the knowledge of the raw dataset. To achieve this, a hybrid loss is designed to integrate backdoor information along the benign optimization trajectory, ensuring that previously learned information is not forgotten. Extensive experiments demonstrate that distilled datasets are highly vulnerable to backdoor attacks, with risks pervasive across various raw datasets, distillation methods, and downstream training strategies. Moreover, our attack method is efficient, capable of synthesizing a malicious distilled dataset in under one minute in certain cases.
2502.04230
XAttnMark: Learning Robust Audio Watermarking with Cross-Attention
cs.SD cs.AI cs.CR cs.LG eess.AS
The rapid proliferation of generative audio synthesis and editing technologies has raised significant concerns about copyright infringement, data provenance, and the spread of misinformation through deepfake audio. Watermarking offers a proactive solution by embedding imperceptible, identifiable, and traceable marks into audio content. While recent neural network-based watermarking methods like WavMark and AudioSeal have improved robustness and quality, they struggle to achieve both robust detection and accurate attribution simultaneously. This paper introduces Cross-Attention Robust Audio Watermark (XAttnMark), which bridges this gap by leveraging partial parameter sharing between the generator and the detector, a cross-attention mechanism for efficient message retrieval, and a temporal conditioning module for improved message distribution. Additionally, we propose a psychoacoustic-aligned temporal-frequency masking loss that captures fine-grained auditory masking effects, enhancing watermark imperceptibility. Our approach achieves state-of-the-art performance in both detection and attribution, demonstrating superior robustness against a wide range of audio transformations, including challenging generative editing with strong editing strength. The project webpage is available at https://liuyixin-louis.github.io/xattnmark/.
2502.04233
Graph machine learning for flight delay prediction due to holding manouver
cs.LG cs.SI
Flight delays due to holding maneuvers are a critical and costly phenomenon in aviation, driven by the need to manage air traffic congestion and ensure safety. Holding maneuvers occur when aircraft are instructed to circle in designated airspace, often due to factors such as airport congestion, adverse weather, or air traffic control restrictions. This study models the prediction of flight delays due to holding maneuvers as a graph problem, leveraging advanced Graph Machine Learning (Graph ML) techniques to capture complex interdependencies in air traffic networks. Holding maneuvers, while crucial for safety, cause increased fuel usage, emissions, and passenger dissatisfaction, making accurate prediction essential for operational efficiency. Traditional machine learning models, typically using tabular data, often overlook spatial-temporal relations within air traffic data. To address this, we model the problem of predicting holding as edge feature prediction in a directed (multi)graph where we apply both CatBoost, enriched with graph features capturing network centrality and connectivity, and Graph Attention Networks (GATs), which excel in relational data contexts. Our results indicate that CatBoost outperforms GAT in this imbalanced dataset, effectively predicting holding events and offering interpretability through graph-based feature importance. Additionally, we discuss the model's potential operational impact through a web-based tool that allows users to simulate real-time delay predictions. This research underscores the viability of graph-based approaches for predictive analysis in aviation, with implications for enhancing fuel efficiency, reducing delays, and improving passenger experience.
2502.04234
A Classification System Approach in Predicting Chinese Censorship
cs.CL cs.LG cs.SI
This paper is dedicated to using a classifier to predict whether a Weibo post would be censored under the Chinese internet. Through randomized sampling from \citeauthor{Fu2021} and Chinese tokenizing strategies, we constructed a cleaned Chinese phrase dataset with binary censorship markings. Utilizing various probability-based information retrieval methods on the data, we were able to derive 4 logistic regression models for classification. Furthermore, we experimented with pre-trained transformers to perform similar classification tasks. After evaluating both the macro-F1 and ROC-AUC metrics, we concluded that the Fined-Tuned BERT model exceeds other strategies in performance.
2502.04235
MAGA: MAssive Genre-Audience Reformulation to Pretraining Corpus Expansion
cs.CL
Despite the remarkable capabilities of large language models across various tasks, their continued scaling faces a critical challenge: the scarcity of high-quality pretraining data. While model architectures continue to evolve, the natural language data struggles to scale up. To tackle this bottleneck, we propose \textbf{MA}ssive \textbf{G}enre-\textbf{A}udience~(MAGA) reformulation method, which systematic synthesizes diverse, contextually-rich pretraining data from existing corpus. This work makes three main contributions: (1) We propose MAGA reformulation method, a lightweight and scalable approach for pretraining corpus expansion, and build a 770B tokens MAGACorpus. (2) We evaluate MAGACorpus with different data budget scaling strategies, demonstrating consistent improvements across various model sizes (134M-13B), establishing the necessity for next-generation large-scale synthetic pretraining language models. (3) Through comprehensive analysis, we investigate prompt engineering's impact on synthetic training collapse and reveal limitations in conventional collapse detection metrics using validation losses. Our work shows that MAGA can substantially expand training datasets while maintaining quality, offering a reliably pathway for scaling models beyond data limitations.
2502.04240
Memory-dependent abstractions of stochastic systems through the lens of transfer operators
eess.SY cs.SY
With the increasing ubiquity of safety-critical autonomous systems operating in uncertain environments, there is a need for mathematical methods for formal verification of stochastic models. Towards formally verifying properties of stochastic systems, methods based on discrete, finite Markov approximations -- abstractions -- thereof have surged in recent years. These are found in contexts where: either a) one only has partial, discrete observations of the underlying continuous stochastic process, or b) the original system is too complex to analyze, so one partitions the continuous state-space of the original system to construct a handleable, finite-state model thereof. In both cases, the abstraction is an approximation of the discrete stochastic process that arises precisely from the discretization of the underlying continuous process. The fact that the abstraction is Markov and the discrete process is not (even though the original one is) leads to approximation errors. Towards accounting for non-Markovianity, we introduce memory-dependent abstractions for stochastic systems, capturing dynamics with memory effects. Our contribution is twofold. First, we provide a formalism for memory-dependent abstractions based on transfer operators. Second, we quantify the approximation error by upper bounding the total variation distance between the true continuous state distribution and its discrete approximation.
2502.04242
A Theoretical Framework for Data Efficient Multi-Source Transfer Learning Based on Cram\'er-Rao Bound
cs.LG cs.AI
Multi-source transfer learning provides an effective solution to data scarcity in real-world supervised learning scenarios by leveraging multiple source tasks. In this field, existing works typically use all available samples from sources in training, which constrains their training efficiency and may lead to suboptimal results. To address this, we propose a theoretical framework that answers the question: what is the optimal quantity of source samples needed from each source task to jointly train the target model? Specifically, we introduce a generalization error measure that aligns with cross-entropy loss, and minimize it based on the Cram\'er-Rao Bound to determine the optimal transfer quantity for each source task. Additionally, we develop an architecture-agnostic and data-efficient algorithm OTQMS to implement our theoretical results for training deep multi-source transfer learning models. Experimental studies on diverse architectures and two real-world benchmark datasets show that our proposed algorithm significantly outperforms state-of-the-art approaches in both accuracy and data efficiency. The code and supplementary materials are available in https://anonymous.4open.science/r/Materials.
2502.04244
An object detection approach for lane change and overtake detection from motion profiles
cs.CV
In the application domain of fleet management and driver monitoring, it is very challenging to obtain relevant driving events and activities from dashcam footage while minimizing the amount of information stored and analyzed. In this paper, we address the identification of overtake and lane change maneuvers with a novel object detection approach applied to motion profiles, a compact representation of driving video footage into a single image. To train and test our model we created an internal dataset of motion profile images obtained from a heterogeneous set of dashcam videos, manually labeled with overtake and lane change maneuvers by the ego-vehicle. In addition to a standard object-detection approach, we show how the inclusion of CoordConvolution layers further improves the model performance, in terms of mAP and F1 score, yielding state-of-the art performance when compared to other baselines from the literature. The extremely low computational requirements of the proposed solution make it especially suitable to run in device.
2502.04245
TriNER: A Series of Named Entity Recognition Models For Hindi, Bengali & Marathi
cs.CL cs.AI cs.LG
India's rich cultural and linguistic diversity poses various challenges in the domain of Natural Language Processing (NLP), particularly in Named Entity Recognition (NER). NER is a NLP task that aims to identify and classify tokens into different entity groups like Person, Location, Organization, Number, etc. This makes NER very useful for downstream tasks like context-aware anonymization. This paper details our work to build a multilingual NER model for the three most spoken languages in India - Hindi, Bengali & Marathi. We train a custom transformer model and fine tune a few pretrained models, achieving an F1 Score of 92.11 for a total of 6 entity groups. Through this paper, we aim to introduce a single model to perform NER and significantly reduce the inconsistencies in entity groups and tag names, across the three languages.
2502.04247
Student-t processes as infinite-width limits of posterior Bayesian neural networks
stat.ML cs.LG math.PR
The asymptotic properties of Bayesian Neural Networks (BNNs) have been extensively studied, particularly regarding their approximations by Gaussian processes in the infinite-width limit. We extend these results by showing that posterior BNNs can be approximated by Student-t processes, which offer greater flexibility in modeling uncertainty. Specifically, we show that, if the parameters of a BNN follow a Gaussian prior distribution, and the variance of both the last hidden layer and the Gaussian likelihood function follows an Inverse-Gamma prior distribution, then the resulting posterior BNN converges to a Student-t process in the infinite-width limit. Our proof leverages the Wasserstein metric to establish control over the convergence rate of the Student-t process approximation.
2502.04248
Adapting to Evolving Adversaries with Regularized Continual Robust Training
cs.LG
Robust training methods typically defend against specific attack types, such as Lp attacks with fixed budgets, and rarely account for the fact that defenders may encounter new attacks over time. A natural solution is to adapt the defended model to new adversaries as they arise via fine-tuning, a method which we call continual robust training (CRT). However, when implemented naively, fine-tuning on new attacks degrades robustness on previous attacks. This raises the question: how can we improve the initial training and fine-tuning of the model to simultaneously achieve robustness against previous and new attacks? We present theoretical results which show that the gap in a model's robustness against different attacks is bounded by how far each attack perturbs a sample in the model's logit space, suggesting that regularizing with respect to this logit space distance can help maintain robustness against previous attacks. Extensive experiments on 3 datasets (CIFAR-10, CIFAR-100, and ImageNette) and over 100 attack combinations demonstrate that the proposed regularization improves robust accuracy with little overhead in training time. Our findings and open-source code lay the groundwork for the deployment of models robust to evolving attacks.
2502.04249
Free Energy Risk Metrics for Systemically Safe AI: Gatekeeping Multi-Agent Study
cs.AI cs.LG cs.MA physics.data-an stat.ML
We investigate the Free Energy Principle as a foundation for measuring risk in agentic and multi-agent systems. From these principles we introduce a Cumulative Risk Exposure metric that is flexible to differing contexts and needs. We contrast this to other popular theories for safe AI that hinge on massive amounts of data or describing arbitrarily complex world models. In our framework, stakeholders need only specify their preferences over system outcomes, providing straightforward and transparent decision rules for risk governance and mitigation. This framework naturally accounts for uncertainty in both world model and preference model, allowing for decision-making that is epistemically and axiologically humble, parsimonious, and future-proof. We demonstrate this novel approach in a simplified autonomous vehicle environment with multi-agent vehicles whose driving policies are mediated by gatekeepers that evaluate, in an online fashion, the risk to the collective safety in their neighborhood, and intervene through each vehicle's policy when appropriate. We show that the introduction of gatekeepers in an AV fleet, even at low penetration, can generate significant positive externalities in terms of increased system safety.
2502.04251
Combining Language and App UI Analysis for the Automated Assessment of Bug Reproduction Steps
cs.SE cs.LG
Bug reports are essential for developers to confirm software problems, investigate their causes, and validate fixes. Unfortunately, reports often miss important information or are written unclearly, which can cause delays, increased issue resolution effort, or even the inability to solve issues. One of the most common components of reports that are problematic is the steps to reproduce the bug(s) (S2Rs), which are essential to replicate the described program failures and reason about fixes. Given the proclivity for deficiencies in reported S2Rs, prior work has proposed techniques that assist reporters in writing or assessing the quality of S2Rs. However, automated understanding of S2Rs is challenging, and requires linking nuanced natural language phrases with specific, semantically related program information. Prior techniques often struggle to form such language to program connections - due to issues in language variability and limitations of information gleaned from program analyses. To more effectively tackle the problem of S2R quality annotation, we propose a new technique called AstroBR, which leverages the language understanding capabilities of LLMs to identify and extract the S2Rs from bug reports and map them to GUI interactions in a program state model derived via dynamic analysis. We compared AstroBR to a related state-of-the-art approach and we found that AstroBR annotates S2Rs 25.2% better (in terms of F1 score) than the baseline. Additionally, AstroBR suggests more accurate missing S2Rs than the baseline (by 71.4% in terms of F1 score).
2502.04256
Work in Progress: AI-Powered Engineering-Bridging Theory and Practice
eess.SY cs.SE cs.SY
This paper explores how generative AI can help automate and improve key steps in systems engineering. It examines AI's ability to analyze system requirements based on INCOSE's "good requirement" criteria, identifying well-formed and poorly written requirements. The AI does not just classify requirements but also explains why some do not meet the standards. By comparing AI assessments with those of experienced engineers, the study evaluates the accuracy and reliability of AI in identifying quality issues. Additionally, it explores AI's ability to classify functional and non-functional requirements and generate test specifications based on these classifications. Through both quantitative and qualitative analysis, the research aims to assess AI's potential to streamline engineering processes and improve learning outcomes. It also highlights the challenges and limitations of AI, ensuring its safe and ethical use in professional and academic settings.
2502.04260
Realistic Image-to-Image Machine Unlearning via Decoupling and Knowledge Retention
cs.LG
Machine Unlearning allows participants to remove their data from a trained machine learning model in order to preserve their privacy, and security. However, the machine unlearning literature for generative models is rather limited. The literature for image-to-image generative model (I2I model) considers minimizing the distance between Gaussian noise and the output of I2I model for forget samples as machine unlearning. However, we argue that the machine learning model performs fairly well on unseen data i.e., a retrained model will be able to catch generic patterns in the data and hence will not generate an output which is equivalent to Gaussian noise. In this paper, we consider that the model after unlearning should treat forget samples as out-of-distribution (OOD) data, i.e., the unlearned model should no longer recognize or encode the specific patterns found in the forget samples. To achieve this, we propose a framework which decouples the model parameters with gradient ascent, ensuring that forget samples are OOD for unlearned model with theoretical guarantee. We also provide $(\epsilon, \delta)$-unlearning guarantee for model updates with gradient ascent. The unlearned model is further fine-tuned on the remaining samples to maintain its performance. We also propose an attack model to ensure that the unlearned model has effectively removed the influence of forget samples. Extensive empirical evaluation on two large-scale datasets, ImageNet-1K and Places365 highlights the superiority of our approach. To show comparable performance with retrained model, we also show the comparison of a simple AutoEncoder on various baselines on CIFAR-10 dataset.
2502.04262
Efficient Randomized Experiments Using Foundation Models
cs.LG stat.ME stat.ML
Randomized experiments are the preferred approach for evaluating the effects of interventions, but they are costly and often yield estimates with substantial uncertainty. On the other hand, in silico experiments leveraging foundation models offer a cost-effective alternative that can potentially attain higher statistical precision. However, the benefits of in silico experiments come with a significant risk: statistical inferences are not valid if the models fail to accurately predict experimental responses to interventions. In this paper, we propose a novel approach that integrates the predictions from multiple foundation models with experimental data while preserving valid statistical inference. Our estimator is consistent and asymptotically normal, with asymptotic variance no larger than the standard estimator based on experimental data alone. Importantly, these statistical properties hold even when model predictions are arbitrarily biased. Empirical results across several randomized experiments show that our estimator offers substantial precision gains, equivalent to a reduction of up to 20% in the sample size needed to match the same precision as the standard estimator based on experimental data alone.
2502.04263
Cross the Gap: Exposing the Intra-modal Misalignment in CLIP via Modality Inversion
cs.CV cs.AI cs.LG
Pre-trained multi-modal Vision-Language Models like CLIP are widely used off-the-shelf for a variety of applications. In this paper, we show that the common practice of individually exploiting the text or image encoders of these powerful multi-modal models is highly suboptimal for intra-modal tasks like image-to-image retrieval. We argue that this is inherently due to the CLIP-style inter-modal contrastive loss that does not enforce any intra-modal constraints, leading to what we call intra-modal misalignment. To demonstrate this, we leverage two optimization-based modality inversion techniques that map representations from their input modality to the complementary one without any need for auxiliary data or additional trained adapters. We empirically show that, in the intra-modal tasks of image-to-image and text-to-text retrieval, approaching these tasks inter-modally significantly improves performance with respect to intra-modal baselines on more than fifteen datasets. Additionally, we demonstrate that approaching a native inter-modal task (e.g. zero-shot image classification) intra-modally decreases performance, further validating our findings. Finally, we show that incorporating an intra-modal term in the pre-training objective or narrowing the modality gap between the text and image feature embedding spaces helps reduce the intra-modal misalignment. The code is publicly available at: https://github.com/miccunifi/Cross-the-Gap.
2502.04266
Digital Gatekeeping: An Audit of Search Engine Results shows tailoring of queries on the Israel-Palestine Conflict
cs.CY cs.IR
Search engines, often viewed as reliable gateways to information, tailor search results using customization algorithms based on user preferences, location, and more. While this can be useful for routine queries, it raises concerns when the topics are sensitive or contentious, possibly limiting exposure to diverse viewpoints and increasing polarization. To examine the extent of this tailoring, we focused on the Israel-Palestine conflict and developed a privacy-protecting tool to audit the behavior of three search engines: DuckDuckGo, Google and Yahoo. Our study focused on two main questions: (1) How do search results for the same query about the conflict vary among different users? and (2) Are these results influenced by the user's location and browsing history? Our findings revealed significant customization based on location and browsing preferences, unlike previous studies that found only mild personalization for general topics. Moreover, queries related to the conflict were more customized than unrelated queries, and the results were not neutral concerning the conflict's portrayal.
2502.04268
Point2RBox-v2: Rethinking Point-supervised Oriented Object Detection with Spatial Layout Among Instances
cs.CV cs.AI
With the rapidly increasing demand for oriented object detection (OOD), recent research involving weakly-supervised detectors for learning OOD from point annotations has gained great attention. In this paper, we rethink this challenging task setting with the layout among instances and present Point2RBox-v2. At the core are three principles: 1) Gaussian overlap loss. It learns an upper bound for each instance by treating objects as 2D Gaussian distributions and minimizing their overlap. 2) Voronoi watershed loss. It learns a lower bound for each instance through watershed on Voronoi tessellation. 3) Consistency loss. It learns the size/rotation variation between two output sets with respect to an input image and its augmented view. Supplemented by a few devised techniques, e.g. edge loss and copy-paste, the detector is further enhanced. To our best knowledge, Point2RBox-v2 is the first approach to explore the spatial layout among instances for learning point-supervised OOD. Our solution is elegant and lightweight, yet it is expected to give a competitive performance especially in densely packed scenes: 62.61%/86.15%/34.71% on DOTA/HRSC/FAIR1M. Code is available at https://github.com/VisionXLab/point2rbox-v2.
2502.04269
How does a Multilingual LM Handle Multiple Languages?
cs.CL cs.AI
Multilingual language models have significantly advanced due to rapid progress in natural language processing. Models like BLOOM 1.7B, trained on diverse multilingual datasets, aim to bridge linguistic gaps. However, their effectiveness in capturing linguistic knowledge, particularly for low-resource languages, remains an open question. This study critically examines MLMs capabilities in multilingual understanding, semantic representation, and cross-lingual knowledge transfer. While these models perform well for high-resource languages, they struggle with less-represented ones. Additionally, traditional evaluation methods often overlook their internal syntactic and semantic encoding. This research addresses key limitations through three objectives. First, it assesses semantic similarity by analyzing multilingual word embeddings for consistency using cosine similarity. Second, it examines BLOOM-1.7B and Qwen2 through Named Entity Recognition and sentence similarity tasks to understand their linguistic structures. Third, it explores cross-lingual knowledge transfer by evaluating generalization from high-resource to low-resource languages in sentiment analysis and text classification. By leveraging linguistic probing, performance metrics, and visualizations, this study provides insights into the strengths and limitations of MLMs. The findings aim to enhance multilingual NLP models, ensuring better support for both high- and low-resource languages, thereby promoting inclusivity in language technologies.
2502.04270
PILAF: Optimal Human Preference Sampling for Reward Modeling
cs.LG stat.ML
As large language models increasingly drive real-world applications, aligning them with human values becomes paramount. Reinforcement Learning from Human Feedback (RLHF) has emerged as a key technique, translating preference data into reward models when oracle human values remain inaccessible. In practice, RLHF mostly relies on approximate reward models, which may not consistently guide the policy toward maximizing the underlying human values. We propose Policy-Interpolated Learning for Aligned Feedback (PILAF), a novel response sampling strategy for preference labeling that explicitly aligns preference learning with maximizing the underlying oracle reward. PILAF is theoretically grounded, demonstrating optimality from both an optimization and a statistical perspective. The method is straightforward to implement and demonstrates strong performance in iterative and online RLHF settings where feedback curation is critical.
2502.04271
Variational decision diagrams for quantum-inspired machine learning applications
quant-ph cs.LG
Decision diagrams (DDs) have emerged as an efficient tool for simulating quantum circuits due to their capacity to exploit data redundancies in quantum states and quantum operations, enabling the efficient computation of probability amplitudes. However, their application in quantum machine learning (QML) has remained unexplored. This paper introduces variational decision diagrams (VDDs), a novel graph structure that combines the structural benefits of DDs with the adaptability of variational methods for efficiently representing quantum states. We investigate the trainability of VDDs by applying them to the ground state estimation problem for transverse-field Ising and Heisenberg Hamiltonians. Analysis of gradient variance suggests that training VDDs is possible, as no signs of vanishing gradients--also known as barren plateaus--are observed. This work provides new insights into the use of decision diagrams in QML as an alternative to design and train variational ans\"atze.
2502.04273
Electrical Impedance Tomography for Anisotropic Media: a Machine Learning Approach to Classify Inclusions
math.NA cs.LG cs.NA
We consider the problem in Electrical Impedance Tomography (EIT) of identifying one or multiple inclusions in a background-conducting body $\Omega\subset\mathbb{R}^2$, from the knowledge of a finite number of electrostatic measurements taken on its boundary $\partial\Omega$ and modelled by the Dirichlet-to-Neumann (D-N) matrix. Once the presence of one inclusion in $\Omega$ is established, our model, combined with the machine learning techniques of Artificial Neural Networks (ANN) and Support Vector Machines (SVM), may be used to determine the size of the inclusion, the presence of multiple inclusions, and also that of anisotropy within the inclusion(s). Utilising both real and simulated datasets within a 16-electrode setup, we achieve a high rate of inclusion detection and show that two measurements are sufficient to achieve a good level of accuracy when predicting the size of an inclusion. This underscores the substantial potential of integrating machine learning approaches with the more classical analysis of EIT and the inverse inclusion problem to extract critical insights, such as the presence of anisotropy.
2502.04274
Orthogonal Representation Learning for Estimating Causal Quantities
cs.LG
Representation learning is widely used for estimating causal quantities (e.g., the conditional average treatment effect) from observational data. While existing representation learning methods have the benefit of allowing for end-to-end learning, they do not have favorable theoretical properties of Neyman-orthogonal learners, such as double robustness and quasi-oracle efficiency. Also, such representation learning methods often employ additional constraints, like balancing, which may even lead to inconsistent estimation. In this paper, we propose a novel class of Neyman-orthogonal learners for causal quantities defined at the representation level, which we call OR-learners. Our OR-learners have several practical advantages: they allow for consistent estimation of causal quantities based on any learned representation, while offering favorable theoretical properties including double robustness and quasi-oracle efficiency. In multiple experiments, we show that, under certain regularity conditions, our OR-learners improve existing representation learning methods and achieve state-of-the-art performance. To the best of our knowledge, our OR-learners are the first work to offer a unified framework of representation learning methods and Neyman-orthogonal learners for causal quantities estimation.
2502.04276
Gaussian Process Regression for Inverse Problems in Linear PDEs
stat.ML cs.LG math.AC
This paper introduces a computationally efficient algorithm in system theory for solving inverse problems governed by linear partial differential equations (PDEs). We model solutions of linear PDEs using Gaussian processes with priors defined based on advanced commutative algebra and algebraic analysis. The implementation of these priors is algorithmic and achieved using the Macaulay2 computer algebra software. An example application includes identifying the wave speed from noisy data for classical wave equations, which are widely used in physics. The method achieves high accuracy while enhancing computational efficiency.
2502.04281
DECAF: Learning to be Fair in Multi-agent Resource Allocation
cs.LG cs.CY cs.MA
A wide variety of resource allocation problems operate under resource constraints that are managed by a central arbitrator, with agents who evaluate and communicate preferences over these resources. We formulate this broad class of problems as Distributed Evaluation, Centralized Allocation (DECA) problems and propose methods to learn fair and efficient policies in centralized resource allocation. Our methods are applied to learning long-term fairness in a novel and general framework for fairness in multi-agent systems. We show three different methods based on Double Deep Q-Learning: (1) A joint weighted optimization of fairness and utility, (2) a split optimization, learning two separate Q-estimators for utility and fairness, and (3) an online policy perturbation to guide existing black-box utility functions toward fair solutions. Our methods outperform existing fair MARL approaches on multiple resource allocation domains, even when evaluated using diverse fairness functions, and allow for flexible online trade-offs between utility and fairness.
2502.04286
A Methodology for Studying Linguistic and Cultural Change in China, 1900-1950
cs.CL
This paper presents a quantitative approach to studying linguistic and cultural change in China during the first half of the twentieth century, a period that remains understudied in computational humanities research. The dramatic changes in Chinese language and culture during this time call for greater reflection on the tools and methods used for text analysis. This preliminary study offers a framework for analyzing Chinese texts from the late nineteenth and twentieth centuries, demonstrating how established methods such as word counts and word embeddings can provide new historical insights into the complex negotiations between Western modernity and Chinese cultural discourse.
2502.04288
Leveraging Geolocation in Clinical Records to Improve Alzheimer's Disease Diagnosis Using DMV Framework
cs.LG
Alzheimer's Disease (AD) early detection is critical for enabling timely intervention and improving patient outcomes. This paper presents a DMV framework using Llama3-70B and GPT-4o as embedding models to analyze clinical notes and predict a continuous risk score associated with early AD onset. Framing the task as a regression problem, we model the relationship between linguistic features in clinical notes (inputs) and a target variable (data value) that answers specific questions related to AD risk within certain topic categories. By leveraging a multi-faceted feature set that includes geolocation data, we capture additional environmental context potentially linked to AD. Our results demonstrate that the integration of the geolocation information significantly decreases the error of predicting early AD risk scores over prior models by 28.57% (Llama3-70B) and 33.47% (GPT4-o). Our findings suggest that this combined approach can enhance the predictive accuracy of AD risk assessment, supporting early diagnosis and intervention in clinical settings. Additionally, the framework's ability to incorporate geolocation data provides a more comprehensive risk assessment model that could help healthcare providers better understand and address environmental factors contributing to AD development.
2502.04289
Retro-Rank-In: A Ranking-Based Approach for Inorganic Materials Synthesis Planning
physics.chem-ph cs.LG
Retrosynthesis strategically plans the synthesis of a chemical target compound from simpler, readily available precursor compounds. This process is critical for synthesizing novel inorganic materials, yet traditional methods in inorganic chemistry continue to rely on trial-and-error experimentation. Emerging machine-learning approaches struggle to generalize to entirely new reactions due to their reliance on known precursors, as they frame retrosynthesis as a multi-label classification task. To address these limitations, we propose Retro-Rank-In, a novel framework that reformulates the retrosynthesis problem by embedding target and precursor materials into a shared latent space and learning a pairwise ranker on a bipartite graph of inorganic compounds. We evaluate Retro-Rank-In's generalizability on challenging retrosynthesis dataset splits designed to mitigate data duplicates and overlaps. For instance, for Cr2AlB2, it correctly predicts the verified precursor pair CrB + Al despite never seeing them in training, a capability absent in prior work. Extensive experiments show that Retro-Rank-In sets a new state-of-the-art, particularly in out-of-distribution generalization and candidate set ranking, offering a powerful tool for accelerating inorganic material synthesis.
2502.04290
Every Call is Precious: Global Optimization of Black-Box Functions with Unknown Lipschitz Constants
cs.LG cs.AI cs.SY eess.SY math.OC stat.ML
Optimizing expensive, non-convex, black-box Lipschitz continuous functions presents significant challenges, particularly when the Lipschitz constant of the underlying function is unknown. Such problems often demand numerous function evaluations to approximate the global optimum, which can be prohibitive in terms of time, energy, or resources. In this work, we introduce Every Call is Precious (ECP), a novel global optimization algorithm that minimizes unpromising evaluations by strategically focusing on potentially optimal regions. Unlike previous approaches, ECP eliminates the need to estimate the Lipschitz constant, thereby avoiding additional function evaluations. ECP guarantees no-regret performance for infinite evaluation budgets and achieves minimax-optimal regret bounds within finite budgets. Extensive ablation studies validate the algorithm's robustness, while empirical evaluations show that ECP outperforms 10 benchmark algorithms including Lipschitz, Bayesian, bandits, and evolutionary methods across 30 multi-dimensional non-convex synthetic and real-world optimization problems, which positions ECP as a competitive approach for global optimization.
2502.04293
GCE-Pose: Global Context Enhancement for Category-level Object Pose Estimation
cs.CV
A key challenge in model-free category-level pose estimation is the extraction of contextual object features that generalize across varying instances within a specific category. Recent approaches leverage foundational features to capture semantic and geometry cues from data. However, these approaches fail under partial visibility. We overcome this with a first-complete-then-aggregate strategy for feature extraction utilizing class priors. In this paper, we present GCE-Pose, a method that enhances pose estimation for novel instances by integrating category-level global context prior. GCE-Pose performs semantic shape reconstruction with a proposed Semantic Shape Reconstruction (SSR) module. Given an unseen partial RGB-D object instance, our SSR module reconstructs the instance's global geometry and semantics by deforming category-specific 3D semantic prototypes through a learned deep Linear Shape Model. We further introduce a Global Context Enhanced (GCE) feature fusion module that effectively fuses features from partial RGB-D observations and the reconstructed global context. Extensive experiments validate the impact of our global context prior and the effectiveness of the GCE fusion module, demonstrating that GCE-Pose significantly outperforms existing methods on challenging real-world datasets HouseCat6D and NOCS-REAL275. Our project page is available at https://colin-de.github.io/GCE-Pose/.
2502.04294
Prediction-Powered E-Values
stat.ML cs.LG stat.ME
Quality statistical inference requires a sufficient amount of data, which can be missing or hard to obtain. To this end, prediction-powered inference has risen as a promising methodology, but existing approaches are largely limited to Z-estimation problems such as inference of means and quantiles. In this paper, we apply ideas of prediction-powered inference to e-values. By doing so, we inherit all the usual benefits of e-values -- such as anytime-validity, post-hoc validity and versatile sequential inference -- as well as greatly expand the set of inferences achievable in a prediction-powered manner. In particular, we show that every inference procedure that can be framed in terms of e-values has a prediction-powered counterpart, given by our method. We showcase the effectiveness of our framework across a wide range of inference tasks, from simple hypothesis testing and confidence intervals to more involved procedures for change-point detection and causal discovery, which were out of reach of previous techniques. Our approach is modular and easily integrable into existing algorithms, making it a compelling choice for practical applications.
2502.04295
Beyond Prompt Content: Enhancing LLM Performance via Content-Format Integrated Prompt Optimization
cs.CL
Large Language Models (LLMs) have shown significant capability across various tasks, with their real-world effectiveness often driven by prompt design. While recent research has focused on optimizing prompt content, the role of prompt formatting, a critical but often overlooked dimension, has received limited systematic investigation. In this paper, we introduce Content-Format Integrated Prompt Optimization (CFPO), an innovative methodology that jointly optimizes both prompt content and formatting through an iterative refinement process. CFPO leverages natural language mutations to explore content variations and employs a dynamic format exploration strategy that systematically evaluates diverse format options. Our extensive evaluations across multiple tasks and open-source LLMs demonstrate that CFPO demonstrates measurable performance improvements compared to content-only optimization methods. This highlights the importance of integrated content-format optimization and offers a practical, model-agnostic approach to enhancing LLM performance. Code is available at https://github.com/HenryLau7/CFPO.
2502.04296
Learning Real-World Action-Video Dynamics with Heterogeneous Masked Autoregression
cs.RO cs.CV cs.LG
We propose Heterogeneous Masked Autoregression (HMA) for modeling action-video dynamics to generate high-quality data and evaluation in scaling robot learning. Building interactive video world models and policies for robotics is difficult due to the challenge of handling diverse settings while maintaining computational efficiency to run in real time. HMA uses heterogeneous pre-training from observations and action sequences across different robotic embodiments, domains, and tasks. HMA uses masked autoregression to generate quantized or soft tokens for video predictions. \ourshort achieves better visual fidelity and controllability than the previous robotic video generation models with 15 times faster speed in the real world. After post-training, this model can be used as a video simulator from low-level action inputs for evaluating policies and generating synthetic data. See this link https://liruiw.github.io/hma for more information.
2502.04297
Statistical guarantees for continuous-time policy evaluation: blessing of ellipticity and new tradeoffs
cs.LG math.OC math.PR math.ST stat.TH
We study the estimation of the value function for continuous-time Markov diffusion processes using a single, discretely observed ergodic trajectory. Our work provides non-asymptotic statistical guarantees for the least-squares temporal-difference (LSTD) method, with performance measured in the first-order Sobolev norm. Specifically, the estimator attains an $O(1 / \sqrt{T})$ convergence rate when using a trajectory of length $T$; notably, this rate is achieved as long as $T$ scales nearly linearly with both the mixing time of the diffusion and the number of basis functions employed. A key insight of our approach is that the ellipticity inherent in the diffusion process ensures robust performance even as the effective horizon diverges to infinity. Moreover, we demonstrate that the Markovian component of the statistical error can be controlled by the approximation error, while the martingale component grows at a slower rate relative to the number of basis functions. By carefully balancing these two sources of error, our analysis reveals novel trade-offs between approximation and statistical errors.
2502.04299
MotionCanvas: Cinematic Shot Design with Controllable Image-to-Video Generation
cs.CV
This paper presents a method that allows users to design cinematic video shots in the context of image-to-video generation. Shot design, a critical aspect of filmmaking, involves meticulously planning both camera movements and object motions in a scene. However, enabling intuitive shot design in modern image-to-video generation systems presents two main challenges: first, effectively capturing user intentions on the motion design, where both camera movements and scene-space object motions must be specified jointly; and second, representing motion information that can be effectively utilized by a video diffusion model to synthesize the image animations. To address these challenges, we introduce MotionCanvas, a method that integrates user-driven controls into image-to-video (I2V) generation models, allowing users to control both object and camera motions in a scene-aware manner. By connecting insights from classical computer graphics and contemporary video generation techniques, we demonstrate the ability to achieve 3D-aware motion control in I2V synthesis without requiring costly 3D-related training data. MotionCanvas enables users to intuitively depict scene-space motion intentions, and translates them into spatiotemporal motion-conditioning signals for video diffusion models. We demonstrate the effectiveness of our method on a wide range of real-world image content and shot-design scenarios, highlighting its potential to enhance the creative workflows in digital content creation and adapt to various image and video editing applications.
2502.04302
Strong Equivalence in Answer Set Programming with Constraints
cs.AI cs.LO
We investigate the concept of strong equivalence within the extended framework of Answer Set Programming with constraints. Two groups of rules are considered strongly equivalent if, informally speaking, they have the same meaning in any context. We demonstrate that, under certain assumptions, strong equivalence between rule sets in this extended setting can be precisely characterized by their equivalence in the logic of Here-and-There with constraints. Furthermore, we present a translation from the language of several clingo-based answer set solvers that handle constraints into the language of Here-and-There with constraints. This translation enables us to leverage the logic of Here-and-There to reason about strong equivalence within the context of these solvers. We also explore the computational complexity of determining strong equivalence in this context.
2502.04306
ScoreFlow: Mastering LLM Agent Workflows via Score-based Preference Optimization
cs.CL
Recent research has leveraged large language model multi-agent systems for complex problem-solving while trying to reduce the manual effort required to build them, driving the development of automated agent workflow optimization methods. However, existing methods remain inflexible due to representational limitations, a lack of adaptability, and poor scalability when relying on discrete optimization techniques. We address these challenges with ScoreFlow, a simple yet high-performance framework that leverages efficient gradient-based optimization in a continuous space. ScoreFlow incorporates Score-DPO, a novel variant of the direct preference optimization method that accounts for quantitative feedback. Across six benchmarks spanning question answering, coding, and mathematical reasoning, ScoreFlow achieves an 8.2% improvement over existing baselines. Moreover, it empowers smaller models to outperform larger ones with lower inference costs. Project: https://github.com/Gen-Verse/ScoreFlow
2502.04307
DexterityGen: Foundation Controller for Unprecedented Dexterity
cs.RO cs.AI cs.LG cs.SY eess.SY
Teaching robots dexterous manipulation skills, such as tool use, presents a significant challenge. Current approaches can be broadly categorized into two strategies: human teleoperation (for imitation learning) and sim-to-real reinforcement learning. The first approach is difficult as it is hard for humans to produce safe and dexterous motions on a different embodiment without touch feedback. The second RL-based approach struggles with the domain gap and involves highly task-specific reward engineering on complex tasks. Our key insight is that RL is effective at learning low-level motion primitives, while humans excel at providing coarse motion commands for complex, long-horizon tasks. Therefore, the optimal solution might be a combination of both approaches. In this paper, we introduce DexterityGen (DexGen), which uses RL to pretrain large-scale dexterous motion primitives, such as in-hand rotation or translation. We then leverage this learned dataset to train a dexterous foundational controller. In the real world, we use human teleoperation as a prompt to the controller to produce highly dexterous behavior. We evaluate the effectiveness of DexGen in both simulation and real world, demonstrating that it is a general-purpose controller that can realize input dexterous manipulation commands and significantly improves stability by 10-100x measured as duration of holding objects across diverse tasks. Notably, with DexGen we demonstrate unprecedented dexterous skills including diverse object reorientation and dexterous tool use such as pen, syringe, and screwdriver for the first time.
2502.04308
HOG-Diff: Higher-Order Guided Diffusion for Graph Generation
cs.LG cs.AI cs.SI physics.soc-ph
Graph generation is a critical yet challenging task as empirical analyses require a deep understanding of complex, non-Euclidean structures. Although diffusion models have recently made significant achievements in graph generation, these models typically adapt from the frameworks designed for image generation, making them ill-suited for capturing the topological properties of graphs. In this work, we propose a novel Higher-order Guided Diffusion (HOG-Diff) model that follows a coarse-to-fine generation curriculum and is guided by higher-order information, enabling the progressive generation of plausible graphs with inherent topological structures. We further prove that our model exhibits a stronger theoretical guarantee than classical diffusion frameworks. Extensive experiments on both molecular and generic graph generation tasks demonstrate that our method consistently outperforms or remains competitive with state-of-the-art baselines. Our code is available at https://github.com/Yiminghh/HOG-Diff.
2502.04309
Targeted Learning for Data Fairness
cs.LG stat.ML
Data and algorithms have the potential to produce and perpetuate discrimination and disparate treatment. As such, significant effort has been invested in developing approaches to defining, detecting, and eliminating unfair outcomes in algorithms. In this paper, we focus on performing statistical inference for fairness. Prior work in fairness inference has largely focused on inferring the fairness properties of a given predictive algorithm. Here, we expand fairness inference by evaluating fairness in the data generating process itself, referred to here as data fairness. We perform inference on data fairness using targeted learning, a flexible framework for nonparametric inference. We derive estimators demographic parity, equal opportunity, and conditional mutual information. Additionally, we find that our estimators for probabilistic metrics exploit double robustness. To validate our approach, we perform several simulations and apply our estimators to real data.
2502.04310
Finding Pegasus: Enhancing Unsupervised Anomaly Detection in High-Dimensional Data using a Manifold-Based Approach
cs.LG astro-ph.CO
Unsupervised machine learning methods are well suited to searching for anomalies at scale but can struggle with the high-dimensional representation of many modern datasets, hence dimensionality reduction (DR) is often performed first. In this paper we analyse unsupervised anomaly detection (AD) from the perspective of the manifold created in DR. We present an idealised illustration, "Finding Pegasus", and a novel formal framework with which we categorise AD methods and their results into "on manifold" and "off manifold". We define these terms and show how they differ. We then use this insight to develop an approach of combining AD methods which significantly boosts AD recall without sacrificing precision in situations employing high DR. When tested on MNIST data, our approach of combining AD methods improves recall by as much as 16 percent compared with simply combining with the best standalone AD method (Isolation Forest), a result which shows great promise for its application to real-world data.
2502.04312
Consistency of augmentation graph and network approximability in contrastive learning
cs.LG math.AP math.SP
Contrastive learning leverages data augmentation to develop feature representation without relying on large labeled datasets. However, despite its empirical success, the theoretical foundations of contrastive learning remain incomplete, with many essential guarantees left unaddressed, particularly the realizability assumption concerning neural approximability of an optimal spectral contrastive loss solution. In this work, we overcome these limitations by analyzing the pointwise and spectral consistency of the augmentation graph Laplacian. We establish that, under specific conditions for data generation and graph connectivity, as the augmented dataset size increases, the augmentation graph Laplacian converges to a weighted Laplace-Beltrami operator on the natural data manifold. These consistency results ensure that the graph Laplacian spectrum effectively captures the manifold geometry. Consequently, they give way to a robust framework for establishing neural approximability, directly resolving the realizability assumption in a current paradigm.
2502.04313
Great Models Think Alike and this Undermines AI Oversight
cs.LG cs.AI cs.CL
As Language Model (LM) capabilities advance, evaluating and supervising them at scale is getting harder for humans. There is hope that other language models can automate both these tasks, which we refer to as "AI Oversight". We study how model similarity affects both aspects of AI oversight by proposing a probabilistic metric for LM similarity based on overlap in model mistakes. Using this metric, we first show that LLM-as-a-judge scores favor models similar to the judge, generalizing recent self-preference results. Then, we study training on LM annotations, and find complementary knowledge between the weak supervisor and strong student model plays a crucial role in gains from "weak-to-strong generalization". As model capabilities increase, it becomes harder to find their mistakes, and we might defer more to AI oversight. However, we observe a concerning trend -- model mistakes are becoming more similar with increasing capabilities, pointing to risks from correlated failures. Our work underscores the importance of reporting and correcting for model similarity, especially in the emerging paradigm of AI oversight.
2502.04314
BOUQuET: dataset, Benchmark and Open initiative for Universal Quality Evaluation in Translation
cs.CL
This paper presents BOUQuET, a multicentric and multi-register/domain dataset and benchmark, and its broader collaborative extension initiative. This dataset is handcrafted in non-English languages first, each of these source languages being represented among the 23 languages commonly used by half of the world's population and therefore having the potential to serve as pivot languages that will enable more accurate translations. The dataset is specially designed to avoid contamination and be multicentric, so as to enforce representation of multilingual language features. In addition, the dataset goes beyond the sentence level, as it is organized in paragraphs of various lengths. Compared with related machine translation (MT) datasets, we show that BOUQuET has a broader representation of domains while simplifying the translation task for non-experts. Therefore, BOUQuET is specially suitable for the open initiative and call for translation participation that we are launching to extend it to a multi-way parallel corpus to any written language.
2502.04315
ChameleonLLM: Batch-Aware Dynamic Low-Rank Adaptation via Inference-Time Clusters
cs.CL cs.AI cs.LG
Recent advances in large language models (LLMs) have shown remarkable performance across diverse tasks. However, these models are typically deployed with fixed weights, which limits their ability to adapt dynamically to the variability inherent in real-world data during inference. This paper introduces ChameleonLLM, a novel framework that enables inference-time adaptation of LLMs by leveraging batch-aware clustering and on-the-fly generation of low-rank updates. Unlike traditional fine-tuning approaches such as Low-Rank Adaptation (LoRA) or methods that rely on a fixed set of pre-learned uniforms (changeable masks), our method dynamically generates adaptive modifications to the decoder weights based on the aggregated statistics of clustered batches. By intelligently grouping similar inputs and computing context-aware low-rank updates via a hyper-network, ChameleonLLM achieves significant performance gains, outperforming conventional LoRA methods while eliminating the overhead of maintaining multiple expert models. Our experiments highlight the potential of our approach to serve as a versatile and highly adaptive solution for language model inference. ChameleonLLM is open-sourced to ensure the reproducibility of our experiments: https://anonymous.4open.science/r/ChamaleonLLM/
2502.04317
Factorized Implicit Global Convolution for Automotive Computational Fluid Dynamics Prediction
cs.CV
Computational Fluid Dynamics (CFD) is crucial for automotive design, requiring the analysis of large 3D point clouds to study how vehicle geometry affects pressure fields and drag forces. However, existing deep learning approaches for CFD struggle with the computational complexity of processing high-resolution 3D data. We propose Factorized Implicit Global Convolution (FIGConv), a novel architecture that efficiently solves CFD problems for very large 3D meshes with arbitrary input and output geometries. FIGConv achieves quadratic complexity $O(N^2)$, a significant improvement over existing 3D neural CFD models that require cubic complexity $O(N^3)$. Our approach combines Factorized Implicit Grids to approximate high-resolution domains, efficient global convolutions through 2D reparameterization, and a U-shaped architecture for effective information gathering and integration. We validate our approach on the industry-standard Ahmed body dataset and the large-scale DrivAerNet dataset. In DrivAerNet, our model achieves an $R^2$ value of 0.95 for drag prediction, outperforming the previous state-of-the-art by a significant margin. This represents a 40% improvement in relative mean squared error and a 70% improvement in absolute mean squared error over previous methods.
2502.04318
sshELF: Single-Shot Hierarchical Extrapolation of Latent Features for 3D Reconstruction from Sparse-Views
cs.CV
Reconstructing unbounded outdoor scenes from sparse outward-facing views poses significant challenges due to minimal view overlap. Previous methods often lack cross-scene understanding and their primitive-centric formulations overload local features to compensate for missing global context, resulting in blurriness in unseen parts of the scene. We propose sshELF, a fast, single-shot pipeline for sparse-view 3D scene reconstruction via hierarchal extrapolation of latent features. Our key insights is that disentangling information extrapolation from primitive decoding allows efficient transfer of structural patterns across training scenes. Our method: (1) learns cross-scene priors to generate intermediate virtual views to extrapolate to unobserved regions, (2) offers a two-stage network design separating virtual view generation from 3D primitive decoding for efficient training and modular model design, and (3) integrates a pre-trained foundation model for joint inference of latent features and texture, improving scene understanding and generalization. sshELF can reconstruct 360 degree scenes from six sparse input views and achieves competitive results on synthetic and real-world datasets. We find that sshELF faithfully reconstructs occluded regions, supports real-time rendering, and provides rich latent features for downstream applications. The code will be released.
2502.04320
ConceptAttention: Diffusion Transformers Learn Highly Interpretable Features
cs.CV cs.LG
Do the rich representations of multi-modal diffusion transformers (DiTs) exhibit unique properties that enhance their interpretability? We introduce ConceptAttention, a novel method that leverages the expressive power of DiT attention layers to generate high-quality saliency maps that precisely locate textual concepts within images. Without requiring additional training, ConceptAttention repurposes the parameters of DiT attention layers to produce highly contextualized concept embeddings, contributing the major discovery that performing linear projections in the output space of DiT attention layers yields significantly sharper saliency maps compared to commonly used cross-attention mechanisms. Remarkably, ConceptAttention even achieves state-of-the-art performance on zero-shot image segmentation benchmarks, outperforming 11 other zero-shot interpretability methods on the ImageNet-Segmentation dataset and on a single-class subset of PascalVOC. Our work contributes the first evidence that the representations of multi-modal DiT models like Flux are highly transferable to vision tasks like segmentation, even outperforming multi-modal foundation models like CLIP.
2502.04321
Variation of sentence length across time and genre
cs.CL
The goal of this paper is threefold: i) to present some practical aspects of using full-text version of Corpus of Historical American English (COHA), the largest diachronic multi-genre corpus of the English language, in the investigation of a linguistic trend of change; ii) to test a widely held assumption that sentence length in written English has been steadily decreasing over the past few centuries; iii) to point to a possible link between the changes in sentence length and changes in the English syntactic usage. The empirical proof of concept for iii) is provided by the decline in the frequency of the non-finite purpose subordinator in order to. Sentence length, genre and the likelihood of occurrence of in order to are shown to be interrelated.
2502.04322
Speak Easy: Eliciting Harmful Jailbreaks from LLMs with Simple Interactions
cs.LG cs.AI cs.CL cs.CY
Despite extensive safety alignment efforts, large language models (LLMs) remain vulnerable to jailbreak attacks that elicit harmful behavior. While existing studies predominantly focus on attack methods that require technical expertise, two critical questions remain underexplored: (1) Are jailbroken responses truly useful in enabling average users to carry out harmful actions? (2) Do safety vulnerabilities exist in more common, simple human-LLM interactions? In this paper, we demonstrate that LLM responses most effectively facilitate harmful actions when they are both actionable and informative--two attributes easily elicited in multi-step, multilingual interactions. Using this insight, we propose HarmScore, a jailbreak metric that measures how effectively an LLM response enables harmful actions, and Speak Easy, a simple multi-step, multilingual attack framework. Notably, by incorporating Speak Easy into direct request and jailbreak baselines, we see an average absolute increase of 0.319 in Attack Success Rate and 0.426 in HarmScore in both open-source and proprietary LLMs across four safety benchmarks. Our work reveals a critical yet often overlooked vulnerability: Malicious users can easily exploit common interaction patterns for harmful intentions.
2502.04323
The Uniformly Rotated Mondrian Kernel
cs.LG math.PR
First proposed by Rahimi and Recht, random features are used to decrease the computational cost of kernel machines in large-scale problems. The Mondrian kernel is one such example of a fast random feature approximation of the Laplace kernel, generated by a computationally efficient hierarchical random partition of the input space known as the Mondrian process. In this work, we study a variation of this random feature map by using uniformly randomly rotated Mondrian processes to approximate a kernel that is invariant under rotations. We obtain a closed-form expression for this isotropic kernel, as well as a uniform convergence rate of the uniformly rotated Mondrian kernel to this limit. To this end, we utilize techniques from the theory of stationary random tessellations in stochastic geometry and prove a new result on the geometry of the typical cell of the superposition of uniformly random rotations of Mondrian tessellations. Finally, we test the empirical performance of this random feature map on both synthetic and real-world datasets, demonstrating its improved performance over the Mondrian kernel on a debiased dataset.
2502.04324
Can Grammarly and ChatGPT accelerate language change? AI-powered technologies and their impact on the English language: wordiness vs. conciseness
cs.CL cs.CY
The proliferation of NLP-powered language technologies, AI-based natural language generation models, and English as a mainstream means of communication among both native and non-native speakers make the output of AI-powered tools especially intriguing to linguists. This paper investigates how Grammarly and ChatGPT affect the English language regarding wordiness vs. conciseness. A case study focusing on the purpose subordinator in order to is presented to illustrate the way in which Grammarly and ChatGPT recommend shorter grammatical structures instead of longer and more elaborate ones. Although the analysed sentences were produced by native speakers, are perfectly correct, and were extracted from a language corpus of contemporary English, both Grammarly and ChatGPT suggest more conciseness and less verbosity, even for relatively short sentences. The present article argues that technologies such as Grammarly not only mirror language change but also have the potential to facilitate or accelerate it.
2502.04326
WorldSense: Evaluating Real-world Omnimodal Understanding for Multimodal LLMs
cs.CV cs.AI
In this paper, we introduce WorldSense, the first benchmark to assess the multi-modal video understanding, that simultaneously encompasses visual, audio, and text inputs. In contrast to existing benchmarks, our WorldSense has several features: (i) collaboration of omni-modality, we design the evaluation tasks to feature a strong coupling of audio and video, requiring models to effectively utilize the synergistic perception of omni-modality; (ii) diversity of videos and tasks, WorldSense encompasses a diverse collection of 1,662 audio-visual synchronised videos, systematically categorized into 8 primary domains and 67 fine-grained subcategories to cover the broad scenarios, and 3,172 multi-choice QA pairs across 26 distinct tasks to enable the comprehensive evaluation; (iii) high-quality annotations, all the QA pairs are manually labeled by 80 expert annotators with multiple rounds of correction to ensure quality. Based on our WorldSense, we extensively evaluate various state-of-the-art models. The experimental results indicate that existing models face significant challenges in understanding real-world scenarios (48.0% best accuracy). We hope our WorldSense can provide a platform for evaluating the ability in constructing and understanding coherent contexts from omni-modality.
2502.04327
Value-Based Deep RL Scales Predictably
cs.LG
Scaling data and compute is critical to the success of machine learning. However, scaling demands predictability: we want methods to not only perform well with more compute or data, but also have their performance be predictable from small-scale runs, without running the large-scale experiment. In this paper, we show that value-based off-policy RL methods are predictable despite community lore regarding their pathological behavior. First, we show that data and compute requirements to attain a given performance level lie on a Pareto frontier, controlled by the updates-to-data (UTD) ratio. By estimating this frontier, we can predict this data requirement when given more compute, and this compute requirement when given more data. Second, we determine the optimal allocation of a total resource budget across data and compute for a given performance and use it to determine hyperparameters that maximize performance for a given budget. Third, this scaling behavior is enabled by first estimating predictable relationships between hyperparameters, which is used to manage effects of overfitting and plasticity loss unique to RL. We validate our approach using three algorithms: SAC, BRO, and PQL on DeepMind Control, OpenAI gym, and IsaacGym, when extrapolating to higher levels of data, compute, budget, or performance.
2502.04328
Ola: Pushing the Frontiers of Omni-Modal Language Model with Progressive Modality Alignment
cs.CV cs.CL cs.MM cs.SD eess.AS eess.IV
Recent advances in large language models, particularly following GPT-4o, have sparked increasing interest in developing omni-modal models capable of understanding more modalities. While some open-source alternatives have emerged, there is still a notable lag behind specialized single-modality models in performance. In this paper, we present Ola, an Omni-modal language model that achieves competitive performance across image, video, and audio understanding compared to specialized counterparts. The core design of Ola lies in its progressive modality alignment strategy that extends the supporting modality of the language model progressively. Our training pipeline begins with the most distinct modalities: image and text, then gradually expands the skill sets of the model using speech data that connects language and audio knowledge, and video data that connects all modalities. The progressive learning pipeline also enables us to maintain a relatively small size of the cross-modal alignment data, making developing omni-modal from existing vision-language models easy and less costly. Moreover, to unlock an advanced interactive experience like GPT-4o, we further design a sentence-wise decoding solution for streaming speech generation. Extensive experiments demonstrate that Ola surpasses existing open omni-modal LLMs across all modalities while achieving highly competitive performance compared to state-of-the-art specialized models of similar sizes. We aim to make Ola a fully open omni-modal understanding solution to advance future research in this emerging field. Model weights, code, and data are open-sourced at https://github.com/Ola-Omni/Ola.
2502.04329
SMART: Advancing Scalable Map Priors for Driving Topology Reasoning
cs.CV cs.RO
Topology reasoning is crucial for autonomous driving as it enables comprehensive understanding of connectivity and relationships between lanes and traffic elements. While recent approaches have shown success in perceiving driving topology using vehicle-mounted sensors, their scalability is hindered by the reliance on training data captured by consistent sensor configurations. We identify that the key factor in scalable lane perception and topology reasoning is the elimination of this sensor-dependent feature. To address this, we propose SMART, a scalable solution that leverages easily available standard-definition (SD) and satellite maps to learn a map prior model, supervised by large-scale geo-referenced high-definition (HD) maps independent of sensor settings. Attributed to scaled training, SMART alone achieves superior offline lane topology understanding using only SD and satellite inputs. Extensive experiments further demonstrate that SMART can be seamlessly integrated into any online topology reasoning methods, yielding significant improvements of up to 28% on the OpenLane-V2 benchmark.
2502.04339
Analysis of Diffusion Models for Manifold Data
math.ST cond-mat.dis-nn cs.IT cs.LG math.IT math.PR stat.TH
We analyze the time reversed dynamics of generative diffusion models. If the exact empirical score function is used in a regime of large dimension and exponentially large number of samples, these models are known to undergo transitions between distinct dynamical regimes. We extend this analysis and compute the transitions for an analytically tractable manifold model where the statistical model for the data is a mixture of lower dimensional Gaussians embedded in higher dimensional space. We compute the so-called speciation and collapse transition times, as a function of the ratio of manifold-to-ambient space dimensions, and other characteristics of the data model. An important tool used in our analysis is the exact formula for the mutual information (or free energy) of Generalized Linear Models.
2502.04341
Comparative Analysis of Community Detection Algorithms on the SNAP Social Circles Dataset
cs.SI cs.AI
In network research, Community Detection has always been a topic of significant interest in network science, with numerous papers and algorithms proposing to uncover the underlying structures within networks. In this paper, we conduct a comparative analysis of several prominent community detection algorithms applied to the SNAP Social Circles Dataset, derived from the Facebook Social Media network. The algorithms implemented include Louvain, Girvan-Newman, Spectral Clustering, K-Means Clustering, etc. We evaluate the performance of these algorithms based on various metrics such as modularity, normalized cut-ratio, silhouette score, compactness, and separability. Our findings reveal insights into the effectiveness of each algorithm in detecting various meaningful communities within the social network, shedding light on their strength and limitations. This research contributes to the understanding of community detection methods and provides valuable guidance for their application in analyzing real-world social networks.
2502.04342
Tutorial on Using Machine Learning and Deep Learning Models for Mental Illness Detection
cs.CL cs.AI cs.LG
Social media has become an important source for understanding mental health, providing researchers with a way to detect conditions like depression from user-generated posts. This tutorial provides practical guidance to address common challenges in applying machine learning and deep learning methods for mental health detection on these platforms. It focuses on strategies for working with diverse datasets, improving text preprocessing, and addressing issues such as imbalanced data and model evaluation. Real-world examples and step-by-step instructions demonstrate how to apply these techniques effectively, with an emphasis on transparency, reproducibility, and ethical considerations. By sharing these approaches, this tutorial aims to help researchers build more reliable and widely applicable models for mental health research, contributing to better tools for early detection and intervention.
2502.04343
Synergistic Traffic Assignment
cs.GT cs.MA math.OC
Traffic assignment analyzes traffic flows in road networks that emerge due to traveler interaction. Traditionally, travelers are assumed to use private cars, so road costs grow with the number of users due to congestion. However, in sustainable transit systems, travelers share vehicles s.t. more users on a road lead to higher sharing potential and reduced cost per user. Thus, we invert the usual avoidant traffic assignment (ATA) and instead consider synergistic traffic assignment (STA) where road costs decrease with use. We find that STA is significantly different from ATA from a game-theoretical point of view. We show that a simple iterative best-response method with simultaneous updates converges to an equilibrium state. This enables efficient computation of equilibria using optimized speedup techniques for shortest-path queries. In contrast, ATA requires slower sequential updates or more complicated iteration schemes that only approximate an equilibrium. Experiments with a realistic scenario for the city of Stuttgart indicate that STA indeed quickly converges to an equilibrium. We envision STA as a part of software-defined transportation systems that dynamically adapt to current travel demand. As a first demonstration, we show that an STA equilibrium can be used to incorporate traveler synergism in a simple bus line planning algorithm to potentially greatly reduce the required vehicle resources.
2502.04345
JingFang: A Traditional Chinese Medicine Large Language Model of Expert-Level Medical Diagnosis and Syndrome Differentiation-Based Treatment
cs.CL cs.AI cs.LG
Traditional Chinese medicine (TCM) plays a vital role in health protection and disease treatment, but its practical application requires extensive medical knowledge and clinical experience. Existing TCM Large Language Models (LLMs) exhibit critical limitations of uncomprehensive medical consultation and diagnoses, and inaccurate syndrome differentiation-based treatment. To address these issues, this study establishes JingFang (JF): a novel TCM Large Language Model that demonstrates the expert-level capability of medical diagnosis and syndrome differentiation-based treatment. We innovate a Multi-agent Dynamic Collaborative Chain-of-Thought Mechanism (MDCCTM) for medical consultation, enabling JF with effective and accurate diagnostic ability. In addition, a Syndrome Agent and a Dual-Stage Retrieval Scheme (DSRS) are developed to significantly enhance the capacity of JF for disease treatment based on syndrome differentiation. JingFang not only facilitates the application of LLMs but also promotes the effective practice of TCM in human health protection and disease treatment.
2502.04346
Multi-Lingual Cyber Threat Detection in Tweets/X Using ML, DL, and LLM: A Comparative Analysis
cs.CL cs.AI
Cyber threat detection has become an important area of focus in today's digital age due to the growing spread of fake information and harmful content on social media platforms such as Twitter (now 'X'). These cyber threats, often disguised within tweets, pose significant risks to individuals, communities, and even nations, emphasizing the need for effective detection systems. While previous research has explored tweet-based threats, much of the work is limited to specific languages, domains, or locations, or relies on single-model approaches, reducing their applicability to diverse real-world scenarios. To address these gaps, our study focuses on multi-lingual tweet cyber threat detection using a variety of advanced models. The research was conducted in three stages: (1) We collected and labeled tweet datasets in four languages English, Chinese, Russian, and Arabic employing both manual and polarity-based labeling methods to ensure high-quality annotations. (2) Each dataset was analyzed individually using machine learning (ML) and deep learning (DL) models to assess their performance on distinct languages. (3) Finally, we combined all four datasets into a single multi-lingual dataset and applied DL and large language model (LLM) architectures to evaluate their efficacy in identifying cyber threats across various languages. Our results show that among machine learning models, Random Forest (RF) attained the highest performance; however, the Bi-LSTM architecture consistently surpassed other DL and LLM architectures across all datasets. These findings underline the effectiveness of Bi-LSTM in multilingual cyber threat detection. The code for this paper can be found at this link: https://github.com/Mmurrad/Tweet-Data-Classification.git.
2502.04347
SCALM: Detecting Bad Practices in Smart Contracts Through LLMs
cs.CL cs.AI
As the Ethereum platform continues to mature and gain widespread usage, it is crucial to maintain high standards of smart contract writing practices. While bad practices in smart contracts may not directly lead to security issues, they do elevate the risk of encountering problems. Therefore, to understand and avoid these bad practices, this paper introduces the first systematic study of bad practices in smart contracts, delving into over 35 specific issues. Specifically, we propose a large language models (LLMs)-based framework, SCALM. It combines Step-Back Prompting and Retrieval-Augmented Generation (RAG) to identify and address various bad practices effectively. Our extensive experiments using multiple LLMs and datasets have shown that SCALM outperforms existing tools in detecting bad practices in smart contracts.
2502.04348
Prompt-based Depth Pruning of Large Language Models
cs.CL cs.AI
Depth pruning aims to reduce the inference cost of a large language model without any hardware-specific complications, by simply removing several less important transformer blocks. However, our empirical findings suggest that the importance of a transformer block may be highly task-dependent -- a block that is crucial for a task can be removed without degrading the accuracy on another task. Based on this observation, we develop a dynamic depth pruning algorithm, coined PuDDing (Prompt-routed Dynamic Depth Pruning), which determines which blocks to omit from the model based on the input prompt. PuDDing operates by training a lightweight router to predict the best omission set among a set of options, where this option set has also been constructed in a data-driven manner. Empirical results on commonsense reasoning benchmarks demonstrate that PuDDing effectively accelerates the inference language models, and achieves better on-task performance than static depth pruning baselines.
2502.04349
Dynamic benchmarking framework for LLM-based conversational data capture
cs.CL cs.AI
The rapid evolution of large language models (LLMs) has transformed conversational agents, enabling complex human-machine interactions. However, evaluation frameworks often focus on single tasks, failing to capture the dynamic nature of multi-turn dialogues. This paper introduces a dynamic benchmarking framework to assess LLM-based conversational agents through interactions with synthetic users. The framework integrates generative agent simulation to evaluate performance on key dimensions: information extraction, context awareness, and adaptive engagement. By simulating various aspects of user behavior, our work provides a scalable, automated, and flexible benchmarking approach. Experimental evaluation - within a loan application use case - demonstrates the framework's effectiveness under one-shot and few-shot extraction conditions. Results show that adaptive strategies improve data extraction accuracy, especially when handling ambiguous responses. Future work will extend its applicability to broader domains and incorporate additional metrics (e.g., conversational coherence, user engagement). This study contributes a structured, scalable approach to evaluating LLM-based conversational agents, facilitating real-world deployment.
2502.04350
CodeSteer: Symbolic-Augmented Language Models via Code/Text Guidance
cs.CL cs.AI cs.LG cs.SC cs.SE
Existing methods fail to effectively steer Large Language Models (LLMs) between textual reasoning and code generation, leaving symbolic computing capabilities underutilized. We introduce CodeSteer, an effective method for guiding LLM code/text generation. We construct a comprehensive benchmark SymBench comprising 37 symbolic tasks with adjustable complexity and also synthesize datasets of 12k multi-round guidance/generation trajectories and 5.5k guidance comparison pairs. We fine-tune the Llama-3-8B model with a newly designed multi-round supervised fine-tuning (SFT) and direct preference optimization (DPO). The resulting model, CodeSteerLLM, augmented with the proposed symbolic and self-answer checkers, effectively guides the code/text generation of larger models. Augmenting GPT-4o with CodeSteer raises its average performance score from 53.3 to 86.4, even outperforming the existing best LLM OpenAI o1 (82.7), o1-preview (74.8), and DeepSeek R1 (76.8) across all 37 tasks (28 seen, 9 unseen). Trained for GPT-4o, CodeSteer demonstrates superior generalizability, providing an average 41.8 performance boost on Claude, Mistral, and GPT-3.5. CodeSteer-guided LLMs fully harness symbolic computing to maintain strong performance on highly complex tasks. Models, Datasets, and Codes are available at https://github.com/yongchao98/CodeSteer-v1.0.
2502.04351
NER4all or Context is All You Need: Using LLMs for low-effort, high-performance NER on historical texts. A humanities informed approach
cs.CL cs.AI
Named entity recognition (NER) is a core task for historical research in automatically establishing all references to people, places, events and the like. Yet, do to the high linguistic and genre diversity of sources, only limited canonisation of spellings, the level of required historical domain knowledge, and the scarcity of annotated training data, established approaches to natural language processing (NLP) have been both extremely expensive and yielded only unsatisfactory results in terms of recall and precision. Our paper introduces a new approach. We demonstrate how readily-available, state-of-the-art LLMs significantly outperform two leading NLP frameworks, spaCy and flair, for NER in historical documents by seven to twentytwo percent higher F1-Scores. Our ablation study shows how providing historical context to the task and a bit of persona modelling that turns focus away from a purely linguistic approach are core to a successful prompting strategy. We also demonstrate that, contrary to our expectations, providing increasing numbers of examples in few-shot approaches does not improve recall or precision below a threshold of 16-shot. In consequence, our approach democratises access to NER for all historians by removing the barrier of scripting languages and computational skills required for established NLP tools and instead leveraging natural language prompts and consumer-grade tools and frontends.
2502.04352
Investigating the Robustness of Deductive Reasoning with Large Language Models
cs.CL cs.AI
Large Language Models (LLMs) have been shown to achieve impressive results for many reasoning-based Natural Language Processing (NLP) tasks, suggesting a degree of deductive reasoning capability. However, it remains unclear to which extent LLMs, in both informal and autoformalisation methods, are robust on logical deduction tasks. Moreover, while many LLM-based deduction methods have been proposed, there is a lack of a systematic study that analyses the impact of their design components. Addressing these two challenges, we propose the first study of the robustness of LLM-based deductive reasoning methods. We devise a framework with two families of perturbations: adversarial noise and counterfactual statements, which jointly generate seven perturbed datasets. We organize the landscape of LLM reasoners according to their reasoning format, formalisation syntax, and feedback for error recovery. The results show that adversarial noise affects autoformalisation, while counterfactual statements influence all approaches. Detailed feedback does not improve overall accuracy despite reducing syntax errors, pointing to the challenge of LLM-based methods to self-correct effectively.
2502.04353
CognArtive: Large Language Models for Automating Art Analysis and Decoding Aesthetic Elements
cs.CL cs.AI cs.CV
Art, as a universal language, can be interpreted in diverse ways, with artworks embodying profound meanings and nuances. The advent of Large Language Models (LLMs) and the availability of Multimodal Large Language Models (MLLMs) raise the question of how these transformative models can be used to assess and interpret the artistic elements of artworks. While research has been conducted in this domain, to the best of our knowledge, a deep and detailed understanding of the technical and expressive features of artworks using LLMs has not been explored. In this study, we investigate the automation of a formal art analysis framework to analyze a high-throughput number of artworks rapidly and examine how their patterns evolve over time. We explore how LLMs can decode artistic expressions, visual elements, composition, and techniques, revealing emerging patterns that develop across periods. Finally, we discuss the strengths and limitations of LLMs in this context, emphasizing their ability to process vast quantities of art-related data and generate insightful interpretations. Due to the exhaustive and granular nature of the results, we have developed interactive data visualizations, available online https://cognartive.github.io/, to enhance understanding and accessibility.
2502.04354
Reviving The Classics: Active Reward Modeling in Large Language Model Alignment
cs.CL cs.AI cs.LG
Building neural reward models from human preferences is a pivotal component in reinforcement learning from human feedback (RLHF) and large language model alignment research. Given the scarcity and high cost of human annotation, how to select the most informative pairs to annotate is an essential yet challenging open problem. In this work, we highlight the insight that an ideal comparison dataset for reward modeling should balance exploration of the representation space and make informative comparisons between pairs with moderate reward differences. Technically, challenges arise in quantifying the two objectives and efficiently prioritizing the comparisons to be annotated. To address this, we propose the Fisher information-based selection strategies, adapt theories from the classical experimental design literature, and apply them to the final linear layer of the deep neural network-based reward modeling tasks. Empirically, our method demonstrates remarkable performance, high computational efficiency, and stability compared to other selection methods from deep learning and classical statistical literature across multiple open-source LLMs and datasets. Further ablation studies reveal that incorporating cross-prompt comparisons in active reward modeling significantly enhances labeling efficiency, shedding light on the potential for improved annotation strategies in RLHF.
2502.04355
LLM-ProS: Analyzing Large Language Models' Performance in Competitive Problem Solving
cs.CL cs.AI
The rapid advancement of large language models has opened new avenues for automating complex problem-solving tasks such as algorithmic coding and competitive programming. This paper introduces a novel evaluation technique, LLM-ProS, to assess the performance of state-of-the-art LLMs on International Collegiate Programming Contest (ICPC) problems. Using a curated dataset of 166 World Finals problems from 2011 to 2024, we benchmark the models' reasoning, accuracy, and efficiency. We evaluate the five models-GPT-4o, Mistral Large, Llama-3.1-405B, and the o1 family, consisting of o1-mini and o1-preview, across critical metrics like correctness, resource utilization, and response calibration. Our results reveal significant differences in the models' abilities to generalize, adapt, and solve novel problems. We also investigated the impact of training methodologies, dataset contamination, and chain-of-thought reasoning on model performance. The findings provide new insights into optimizing LLMs for algorithmic tasks, highlighting both strengths and limitations of current models.