id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2502.12499
GPU Memory Usage Optimization for Backward Propagation in Deep Network Training
cs.LG cs.DS
In modern Deep Learning, it has been a trend to design larger Deep Neural Networks (DNNs) for the execution of more complex tasks and better accuracy. On the other hand, Convolutional Neural Networks (CNNs) have become the standard method for most of computer vision tasks. However, the memory allocation for the intermediate data in convolution layers can cause severe memory pressure during model training. Many solutions have been proposed to resolve the problem. Besides hardware-dependent solutions, a general methodology rematerialization can reduce GPU memory usage by trading computation for memory efficiently. The idea is to select a set of intermediate results during the forward phase as checkpoints, and only save them in memory to reduce memory usage. The backward phase recomputes the intermediate data from the closest checkpoints in memory as needed. This recomputation increases execution time but saves memory by not storing all intermediate results in memory during the forward phase. In this paper, we will focus on efficiently finding the optimal checkpoint subset to achieve the least peak memory usage during the model training. We first describe the theoretical background of the training of a neural network using mathematical equations. We use these equations to identify all essential data required during both forward and backward phases to compute the gradient of weights of the model. We first identify the checkpoint selection problem and propose a dynamic programming algorithm with time complexity O(n3) to solve the problem of finding the optimal checkpoint subset. With extensive experiments, we formulate a more accurate description of the problem using our theoretical analysis and revise the objective function based on the tracing, and propose an O(n)-time algorithm for finding the optimal checkpoint subset.
2502.12501
Crowd Comparative Reasoning: Unlocking Comprehensive Evaluations for LLM-as-a-Judge
cs.CL
LLM-as-a-Judge, which generates chain-of-thought (CoT) judgments, has become a widely adopted auto-evaluation method. However, its reliability is compromised by the CoT reasoning's inability to capture comprehensive and deeper details, often leading to incomplete outcomes. Existing methods mainly rely on majority voting or criteria expansion, which is insufficient to address the limitation in CoT. We propose Crowd-based Comparative Evaluation, which introduces additional crowd responses to compare with the candidate responses, thereby exposing deeper and more comprehensive details within the candidate responses. This process effectively guides LLM-as-a-Judge to provide a more detailed CoT judgment. Extensive experiments demonstrate that our approach enhances evaluation reliability, achieving an average accuracy gain of 6.7% across five benchmarks. Moreover, our method produces higher-quality CoTs that facilitate judge distillation and exhibit superior performance in rejection sampling for supervised fine-tuning (SFT), referred to as crowd rejection sampling, thereby enabling more efficient SFT. Our analysis confirms that CoTs generated by ours are more comprehensive and of higher quality, and evaluation accuracy improves as inference scales.
2502.12502
Efficient OpAmp Adaptation for Zoom Attention to Golden Contexts
cs.CL
Large language models (LLMs) have shown significant promise in question-answering (QA) tasks, particularly in retrieval-augmented generation (RAG) scenarios and long-context applications. However, their performance is hindered by noisy reference documents, which often distract from essential information. Despite fine-tuning efforts, Transformer-based architectures struggle to prioritize relevant content. This is evidenced by their tendency to allocate disproportionate attention to irrelevant or later-positioned documents. Recent work proposes the differential attention mechanism to address this issue, but this mechanism is limited by an unsuitable common-mode rejection ratio (CMRR) and high computational costs. Inspired by the operational amplifier (OpAmp), we propose the OpAmp adaptation to address these challenges, which is implemented with adapters efficiently. By integrating the adapter into pre-trained Transformer blocks, our approach enhances focus on the golden context without costly training from scratch. Empirical evaluations on noisy-context benchmarks reveal that our Qwen2.5-OpAmp-72B model, trained with our OpAmp adaptation, surpasses the performance of state-of-the-art LLMs, including DeepSeek-V3 and GPT-4o.
2502.12507
Mixture of Attention Yields Accurate Results for Tabular Data
cs.LG cs.AI
Tabular data inherently exhibits significant feature heterogeneity, but existing transformer-based methods lack specialized mechanisms to handle this property. To bridge the gap, we propose MAYA, an encoder-decoder transformer-based framework. In the encoder, we design a Mixture of Attention (MOA) that constructs multiple parallel attention branches and averages the features at each branch, effectively fusing heterogeneous features while limiting parameter growth. Additionally, we employ collaborative learning with a dynamic consistency weight constraint to produce more robust representations. In the decoder stage, cross-attention is utilized to seamlessly integrate tabular data with corresponding label features. This dual-attention mechanism effectively captures both intra-instance and inter-instance interactions. We evaluate the proposed method on a wide range of datasets and compare it with other state-of-the-art transformer-based methods. Extensive experiments demonstrate that our model achieves superior performance among transformer-based methods in both tabular classification and regression tasks.
2502.12508
Understanding Generalization in Transformers: Error Bounds and Training Dynamics Under Benign and Harmful Overfitting
cs.LG
Transformers serve as the foundational architecture for many successful large-scale models, demonstrating the ability to overfit the training data while maintaining strong generalization on unseen data, a phenomenon known as benign overfitting. However, research on how the training dynamics influence error bounds within the context of benign overfitting has been limited. This paper addresses this gap by developing a generalization theory for a two-layer transformer with labeled flip noise. Specifically, we present generalization error bounds for both benign and harmful overfitting under varying signal-to-noise ratios (SNR), where the training dynamics are categorized into three distinct stages, each with its corresponding error bounds. Additionally, we conduct extensive experiments to identify key factors that influence test errors in transformers. Our experimental results align closely with the theoretical predictions, validating our findings.
2502.12509
LegalCore: A Dataset for Legal Documents Event Coreference Resolution
cs.CL cs.AI
Recognizing events and their coreferential mentions in a document is essential for understanding semantic meanings of text. The existing research on event coreference resolution is mostly limited to news articles. In this paper, we present the first dataset for the legal domain, LegalCore, which has been annotated with comprehensive event and event coreference information. The legal contract documents we annotated in this dataset are several times longer than news articles, with an average length of around 25k tokens per document. The annotations show that legal documents have dense event mentions and feature both short-distance and super long-distance coreference links between event mentions. We further benchmark mainstream Large Language Models (LLMs) on this dataset for both event detection and event coreference resolution tasks, and find that this dataset poses significant challenges for state-of-the-art open-source and proprietary LLMs, which perform significantly worse than a supervised baseline. We will publish the dataset as well as the code.
2502.12510
Aspect-Guided Multi-Level Perturbation Analysis of Large Language Models in Automated Peer Review
cs.CL
We propose an aspect-guided, multi-level perturbation framework to evaluate the robustness of Large Language Models (LLMs) in automated peer review. Our framework explores perturbations in three key components of the peer review process-papers, reviews, and rebuttals-across several quality aspects, including contribution, soundness, presentation, tone, and completeness. By applying targeted perturbations and examining their effects on both LLM-as-Reviewer and LLM-as-Meta-Reviewer, we investigate how aspect-based manipulations, such as omitting methodological details from papers or altering reviewer conclusions, can introduce significant biases in the review process. We identify several potential vulnerabilities: review conclusions that recommend a strong reject may significantly influence meta-reviews, negative or misleading reviews may be wrongly interpreted as thorough, and incomplete or hostile rebuttals can unexpectedly lead to higher acceptance rates. Statistical tests show that these biases persist under various Chain-of-Thought prompting strategies, highlighting the lack of robust critical evaluation in current LLMs. Our framework offers a practical methodology for diagnosing these vulnerabilities, thereby contributing to the development of more reliable and robust automated reviewing systems.
2502.12511
Myna: Masking-Based Contrastive Learning of Musical Representations
cs.SD cs.AI cs.LG
We present Myna, a simple yet effective approach for self-supervised musical representation learning. Built on a contrastive learning framework, Myna introduces two key innovations: (1) the use of a Vision Transformer (ViT) on mel-spectrograms as the backbone and (2) a novel data augmentation strategy, token masking, that masks 90 percent of spectrogram tokens. These innovations deliver both effectiveness and efficiency: (i) Token masking enables a significant increase in per-GPU batch size, from 48 or 120 in prior methods (CLMR, MULE) to 4096. (ii) By avoiding traditional augmentations, Myna retains pitch sensitivity, enhancing performance in tasks like key detection. (iii) The use of vertical patches allows the model to better capture critical features for key detection. Our hybrid model, Myna-22M-Hybrid, processes both 16x16 and 128x2 patches, achieving state-of-the-art results. Trained on a single GPU, it outperforms MULE (62M) on average and rivals MERT-95M, which was trained on 16 and 64 GPUs, respectively. Additionally, it surpasses MERT-95M-public, establishing itself as the best-performing model trained on publicly available data. We release our code and models to promote reproducibility and facilitate future research.
2502.12513
RealSyn: An Effective and Scalable Multimodal Interleaved Document Transformation Paradigm
cs.CV
After pre-training on extensive image-text pairs, Contrastive Language-Image Pre-training (CLIP) demonstrates promising performance on a wide variety of benchmarks. However, a substantial volume of non-paired data, such as multimodal interleaved documents, remains underutilized for vision-language representation learning. To fully leverage these unpaired documents, we initially establish a Real-World Data Extraction pipeline to extract high-quality images and texts. Then we design a hierarchical retrieval method to efficiently associate each image with multiple semantically relevant realistic texts. To further enhance fine-grained visual information, we propose an image semantic augmented generation module for synthetic text production. Furthermore, we employ a semantic balance sampling strategy to improve dataset diversity, enabling better learning of long-tail concepts. Based on these innovations, we construct RealSyn, a dataset combining realistic and synthetic texts, available in three scales: 15M, 30M, and 100M. Extensive experiments demonstrate that RealSyn effectively advances vision-language representation learning and exhibits strong scalability. Models pre-trained on RealSyn achieve state-of-the-art performance on multiple downstream tasks. To facilitate future research, the RealSyn dataset and pre-trained model weights are released at https://github.com/deepglint/RealSyn.
2502.12514
Memory-updated-based Framework for 100% Reliable Flexible Flat Cables Insertion
cs.RO
Automatic assembly lines have increasingly replaced human labor in various tasks; however, the automation of Flexible Flat Cable (FFC) insertion remains unrealized due to its high requirement for effective feedback and dynamic operation, limiting approximately 11% of global industrial capacity. Despite lots of approaches, like vision-based tactile sensors and reinforcement learning, having been proposed, the implementation of human-like high-reliable insertion (i.e., with a 100% success rate in completed insertion) remains a big challenge. Drawing inspiration from human behavior in FFC insertion, which involves sensing three-dimensional forces, translating them into physical concepts, and continuously improving estimates, we propose a novel framework. This framework includes a sensing module for collecting three-dimensional tactile data, a perception module for interpreting this data into meaningful physical signals, and a memory module based on Bayesian theory for reliability estimation and control. This strategy enables the robot to accurately assess its physical state and generate reliable status estimations and corrective actions. Experimental results demonstrate that the robot using this framework can detect alignment errors of 0.5 mm with an accuracy of 97.92% and then achieve a 100% success rate in all completed tests after a few iterations. This work addresses the challenges of unreliable perception and control in complex insertion tasks, highlighting the path toward the development of fully automated production lines.
2502.12516
Can LLMs Extract Frame-Semantic Arguments?
cs.CL
Frame-semantic parsing is a critical task in natural language understanding, yet the ability of large language models (LLMs) to extract frame-semantic arguments remains underexplored. This paper presents a comprehensive evaluation of LLMs on frame-semantic argument identification, analyzing the impact of input representation formats, model architectures, and generalization to unseen and out-of-domain samples. Our experiments, spanning models from 0.5B to 78B parameters, reveal that JSON-based representations significantly enhance performance, and while larger models generally perform better, smaller models can achieve competitive results through fine-tuning. We also introduce a novel approach to frame identification leveraging predicted frame elements, achieving state-of-the-art performance on ambiguous targets. Despite strong generalization capabilities, our analysis finds that LLMs still struggle with out-of-domain data.
2502.12518
New Constant Dimension Codes From the Inserting Mixed Dimension Construction and Multilevel Construction
cs.IT math.IT
Constant dimension codes (CDCs) are essential for error correction in random network coding. A fundamental problem of CDCs is to determine their maximal possible size for given parameters. Inserting construction and multilevel construction are two effective techniques for constructing CDCs. We first provide a sufficient condition for a subspace to be added to the code from the mixed dimension construction in Lao et al. (IEEE Trans. Inf. Theory 69(7): 4333-4344, 2023). By appropriately combining matrix blocks from small CDCs and rank-metric codes, we introduce three inserting constructions based on the mixed dimension construction. Furthermore, the mixed dimension construction and these inserting constructions are improved by the multilevel construction that is based on lifting rank-restricted Ferrers diagram rank-metric codes. Our constructions yield some new lower bounds for CDCs, which are superior to the previously best-known ones.
2502.12520
SAFEERASER: Enhancing Safety in Multimodal Large Language Models through Multimodal Machine Unlearning
cs.CV
As Multimodal Large Language Models (MLLMs) develop, their potential security issues have become increasingly prominent. Machine Unlearning (MU), as an effective strategy for forgetting specific knowledge in training data, has been widely used in privacy protection. However, MU for safety in MLLM has yet to be fully explored. To address this issue, we propose SAFEERASER, a safety unlearning benchmark for MLLMs, consisting of 3,000 images and 28.8K VQA pairs. We comprehensively evaluate unlearning methods from two perspectives: forget quality and model utility. Our findings show that existing MU methods struggle to maintain model performance while implementing the forget operation and often suffer from over-forgetting. Hence, we introduce Prompt Decouple (PD) Loss to alleviate over-forgetting through decouple prompt during unlearning process. To quantitatively measure over-forgetting mitigated by PD Loss, we propose a new metric called Safe Answer Refusal Rate (SARR). Experimental results demonstrate that combining PD Loss with existing unlearning methods can effectively prevent over-forgetting and achieve a decrease of 79.5% in the SARR metric of LLaVA-7B and LLaVA-13B, while maintaining forget quality and model utility. Our code and dataset will be released upon acceptance. Warning: This paper contains examples of harmful language and images, and reader discretion is recommended.
2502.12521
Inference-Time Computations for LLM Reasoning and Planning: A Benchmark and Insights
cs.AI cs.LG
We examine the reasoning and planning capabilities of large language models (LLMs) in solving complex tasks. Recent advances in inference-time techniques demonstrate the potential to enhance LLM reasoning without additional training by exploring intermediate steps during inference. Notably, OpenAI's o1 model shows promising performance through its novel use of multi-step reasoning and verification. Here, we explore how scaling inference-time techniques can improve reasoning and planning, focusing on understanding the tradeoff between computational cost and performance. To this end, we construct a comprehensive benchmark, known as Sys2Bench, and perform extensive experiments evaluating existing inference-time techniques on eleven diverse tasks across five categories, including arithmetic reasoning, logical reasoning, common sense reasoning, algorithmic reasoning, and planning. Our findings indicate that simply scaling inference-time computation has limitations, as no single inference-time technique consistently performs well across all reasoning and planning tasks.
2502.12523
Cohesive Subgraph Discovery in Hypergraphs: A Locality-Driven Indexing Framework
cs.SI
Hypergraphs are increasingly employed to model complex, diverse relationships in modern networks, effectively capturing higher-order interactions. A critical challenge in this domain is the discovery of cohesive subgraphs, which provides valuable insights into hypergraph structures. However, selecting suitable parameters for this task remains unresolved. To address this, we propose an efficient indexing framework designed for online retrieval of cohesive subgraphs. Our approach enables rapid identification of desired structures without requiring exhaustive graph traversals, thus ensuring scalability and practicality. This framework has broad applicability, supporting informed decision-making across various domains by offering a comprehensive view of network landscapes. Extensive experiments on real-world datasets demonstrate the effectiveness and efficiency of our proposed indexing technique.
2502.12524
YOLOv12: Attention-Centric Real-Time Object Detectors
cs.CV cs.AI
Enhancing the network architecture of the YOLO framework has been crucial for a long time, but has focused on CNN-based improvements despite the proven superiority of attention mechanisms in modeling capabilities. This is because attention-based models cannot match the speed of CNN-based models. This paper proposes an attention-centric YOLO framework, namely YOLOv12, that matches the speed of previous CNN-based ones while harnessing the performance benefits of attention mechanisms. YOLOv12 surpasses all popular real-time object detectors in accuracy with competitive speed. For example, YOLOv12-N achieves 40.6% mAP with an inference latency of 1.64 ms on a T4 GPU, outperforming advanced YOLOv10-N / YOLOv11-N by 2.1%/1.2% mAP with a comparable speed. This advantage extends to other model scales. YOLOv12 also surpasses end-to-end real-time detectors that improve DETR, such as RT-DETR / RT-DETRv2: YOLOv12-S beats RT-DETR-R18 / RT-DETRv2-R18 while running 42% faster, using only 36% of the computation and 45% of the parameters. More comparisons are shown in Figure 1.
2502.12525
From Abstract to Actionable: Pairwise Shapley Values for Explainable AI
cs.LG cs.AI
Explainable AI (XAI) is critical for ensuring transparency, accountability, and trust in machine learning systems as black-box models are increasingly deployed within high-stakes domains. Among XAI methods, Shapley values are widely used for their fairness and consistency axioms. However, prevalent Shapley value approximation methods commonly rely on abstract baselines or computationally intensive calculations, which can limit their interpretability and scalability. To address such challenges, we propose Pairwise Shapley Values, a novel framework that grounds feature attributions in explicit, human-relatable comparisons between pairs of data instances proximal in feature space. Our method introduces pairwise reference selection combined with single-value imputation to deliver intuitive, model-agnostic explanations while significantly reducing computational overhead. Here, we demonstrate that Pairwise Shapley Values enhance interpretability across diverse regression and classification scenarios--including real estate pricing, polymer property prediction, and drug discovery datasets. We conclude that the proposed methods enable more transparent AI systems and advance the real-world applicability of XAI.
2502.12527
Comprehensive Assessment and Analysis for NSFW Content Erasure in Text-to-Image Diffusion Models
cs.CV
Text-to-image (T2I) diffusion models have gained widespread application across various domains, demonstrating remarkable creative potential. However, the strong generalization capabilities of these models can inadvertently led they to generate NSFW content even with efforts on filtering NSFW content from the training dataset, posing risks to their safe deployment. While several concept erasure methods have been proposed to mitigate this issue, a comprehensive evaluation of their effectiveness remains absent. To bridge this gap, we present the first systematic investigation of concept erasure methods for NSFW content and its sub-themes in text-to-image diffusion models. At the task level, we provide a holistic evaluation of 11 state-of-the-art baseline methods with 14 variants. Specifically, we analyze these methods from six distinct assessment perspectives, including three conventional perspectives, i.e., erasure proportion, image quality, and semantic alignment, and three new perspectives, i.e., excessive erasure, the impact of explicit and implicit unsafe prompts, and robustness. At the tool level, we perform a detailed toxicity analysis of NSFW datasets and compare the performance of different NSFW classifiers, offering deeper insights into their performance alongside a compilation of comprehensive evaluation metrics. Our benchmark not only systematically evaluates concept erasure methods, but also delves into the underlying factors influencing their performance at the insight level. By synthesizing insights from various evaluation perspectives, we provide a deeper understanding of the challenges and opportunities in the field, offering actionable guidance and inspiration for advancing research and practical applications in concept erasure.
2502.12528
Contextual Linear Bandits with Delay as Payoff
cs.LG
A recent work by Schlisselberg et al. (2024) studies a delay-as-payoff model for stochastic multi-armed bandits, where the payoff (either loss or reward) is delayed for a period that is proportional to the payoff itself. While this captures many real-world applications, the simple multi-armed bandit setting limits the practicality of their results. In this paper, we address this limitation by studying the delay-as-payoff model for contextual linear bandits. Specifically, we start from the case with a fixed action set and propose an efficient algorithm whose regret overhead compared to the standard no-delay case is at most $D\Delta_{\max}\log T$, where $T$ is the total horizon, $D$ is the maximum delay, and $\Delta_{\max}$ is the maximum suboptimality gap. When payoff is loss, we also show further improvement of the bound, demonstrating a separation between reward and loss similar to Schlisselberg et al. (2024). Contrary to standard linear bandit algorithms that construct least squares estimator and confidence ellipsoid, the main novelty of our algorithm is to apply a phased arm elimination procedure by only picking actions in a volumetric spanner of the action set, which addresses challenges arising from both payoff-dependent delays and large action sets. We further extend our results to the case with varying action sets by adopting the reduction from Hanna et al. (2023). Finally, we implement our algorithm and showcase its effectiveness and superior performance in experiments.
2502.12529
Alternating Regret for Online Convex Optimization
cs.LG
Motivated by alternating learning dynamics in two-player games, a recent work by Cevher et al.(2024) shows that $o(\sqrt{T})$ alternating regret is possible for any $T$-round adversarial Online Linear Optimization (OLO) problem, and left as an open question whether the same is true for general Online Convex Optimization (OCO). We answer this question in the affirmative by showing that the continuous Hedge algorithm achieves $\tilde{\mathcal{O}}(d^{\frac{2}{3}}T^{\frac{1}{3}})$ alternating regret for any adversarial $d$-dimensional OCO problems. We show that this implies an alternating learning dynamic that finds a Nash equilibrium for any convex-concave zero-sum games or a coarse correlated equilibrium for any convex two-player general-sum games at a rate of $\tilde{\mathcal{O}}(d^{\frac{2}{3}}/T^{\frac{2}{3}})$. To further improve the time complexity and/or the dimension dependence, we propose another simple algorithm, Follow-the-Regularized-Leader with a regularizer whose convex conjugate is 3rd-order smooth, for OCO with smooth and self-concordant loss functions (such as linear or quadratic losses). We instantiate our algorithm with different regularizers and show that, for example, when the decision set is the $\ell_2$ ball, our algorithm achieves $\tilde{\mathcal{O}}(T^{\frac{2}{5}})$ alternating regret with no dimension dependence (and a better $\tilde{\mathcal{O}}(T^{\frac{1}{3}})$ bound for quadratic losses). We complement our results by showing some algorithm-specific alternating regret lower bounds, including a somewhat surprising $\Omega(\sqrt{T})$ lower bound for a Regret Matching variant that is widely used in alternating learning dynamics.
2502.12530
Policy-to-Language: Train LLMs to Explain Decisions with Flow-Matching Generated Rewards
cs.CL cs.LG
As humans increasingly share environments with diverse agents powered by RL, LLMs, and beyond, the ability to explain their policies in natural language will be vital for reliable coexistence. In this paper, we build a model-agnostic explanation generator based on an LLM. The technical novelty is that the rewards for training this LLM are generated by a generative flow matching model. This model has a specially designed structure with a hidden layer merged with an LLM to harness the linguistic cues of explanations into generating appropriate rewards. Experiments on both RL and LLM tasks demonstrate that our method can generate dense and effective rewards while saving on expensive human feedback; it thus enables effective explanations and even improves the accuracy of the decisions in original tasks.
2502.12531
GSCE: A Prompt Framework with Enhanced Reasoning for Reliable LLM-driven Drone Control
cs.RO cs.AI
The integration of Large Language Models (LLMs) into robotic control, including drones, has the potential to revolutionize autonomous systems. Research studies have demonstrated that LLMs can be leveraged to support robotic operations. However, when facing tasks with complex reasoning, concerns and challenges are raised about the reliability of solutions produced by LLMs. In this paper, we propose a prompt framework with enhanced reasoning to enable reliable LLM-driven control for drones. Our framework consists of novel technical components designed using Guidelines, Skill APIs, Constraints, and Examples, namely GSCE. GSCE is featured by its reliable and constraint-compliant code generation. We performed thorough experiments using GSCE for the control of drones with a wide level of task complexities. Our experiment results demonstrate that GSCE can significantly improve task success rates and completeness compared to baseline approaches, highlighting its potential for reliable LLM-driven autonomous drone systems.
2502.12532
CityEQA: A Hierarchical LLM Agent on Embodied Question Answering Benchmark in City Space
cs.AI
Embodied Question Answering (EQA) has primarily focused on indoor environments, leaving the complexities of urban settings - spanning environment, action, and perception - largely unexplored. To bridge this gap, we introduce CityEQA, a new task where an embodied agent answers open-vocabulary questions through active exploration in dynamic city spaces. To support this task, we present CityEQA-EC, the first benchmark dataset featuring 1,412 human-annotated tasks across six categories, grounded in a realistic 3D urban simulator. Moreover, we propose Planner-Manager-Actor (PMA), a novel agent tailored for CityEQA. PMA enables long-horizon planning and hierarchical task execution: the Planner breaks down the question answering into sub-tasks, the Manager maintains an object-centric cognitive map for spatial reasoning during the process control, and the specialized Actors handle navigation, exploration, and collection sub-tasks. Experiments demonstrate that PMA achieves 60.7% of human-level answering accuracy, significantly outperforming frontier-based baselines. While promising, the performance gap compared to humans highlights the need for enhanced visual reasoning in CityEQA. This work paves the way for future advancements in urban spatial intelligence. Dataset and code are available at https://github.com/BiluYong/CityEQA.git.
2502.12534
NoKSR: Kernel-Free Neural Surface Reconstruction via Point Cloud Serialization
cs.CV
We present a novel approach to large-scale point cloud surface reconstruction by developing an efficient framework that converts an irregular point cloud into a signed distance field (SDF). Our backbone builds upon recent transformer-based architectures (i.e., PointTransformerV3), that serializes the point cloud into a locality-preserving sequence of tokens. We efficiently predict the SDF value at a point by aggregating nearby tokens, where fast approximate neighbors can be retrieved thanks to the serialization. We serialize the point cloud at different levels/scales, and non-linearly aggregate a feature to predict the SDF value. We show that aggregating across multiple scales is critical to overcome the approximations introduced by the serialization (i.e. false negatives in the neighborhood). Our frameworks sets the new state-of-the-art in terms of accuracy and efficiency (better or similar performance with half the latency of the best prior method, coupled with a simpler implementation), particularly on outdoor datasets where sparse-grid methods have shown limited performance.
2502.12535
Learning Transformation-Isomorphic Latent Space for Accurate Hand Pose Estimation
cs.CV
Vision-based regression tasks, such as hand pose estimation, have achieved higher accuracy and faster convergence through representation learning. However, existing representation learning methods often encounter the following issues: the high semantic level of features extracted from images is inadequate for regressing low-level information, and the extracted features include task-irrelevant information, reducing their compactness and interfering with regression tasks. To address these challenges, we propose TI-Net, a highly versatile visual Network backbone designed to construct a Transformation Isomorphic latent space. Specifically, we employ linear transformations to model geometric transformations in the latent space and ensure that {\rm TI-Net} aligns them with those in the image space. This ensures that the latent features capture compact, low-level information beneficial for pose estimation tasks. We evaluated TI-Net on the hand pose estimation task to demonstrate the network's superiority. On the DexYCB dataset, TI-Net achieved a 10% improvement in the PA-MPJPE metric compared to specialized state-of-the-art (SOTA) hand pose estimation methods. Our code will be released in the future.
2502.12536
An Algorithm Board in Neural Decoding
cs.NE cs.AI
Understanding the mechanisms of neural encoding and decoding has always been a highly interesting research topic in fields such as neuroscience and cognitive intelligence. In prior studies, some researchers identified a symmetry in neural data decoded by unsupervised methods in motor scenarios and constructed a cognitive learning system based on this pattern (i.e., symmetry). Nevertheless, the distribution state of the data flow that significantly influences neural decoding positions still remains a mystery within the system, which further restricts the enhancement of the system's interpretability. Based on this, this paper mainly explores changes in the distribution state within the system from the machine learning and mathematical statistics perspectives. In the experiment, we assessed the correctness of this symmetry using various tools and indicators commonly utilized in mathematics and statistics. According to the experimental results, the normal distribution (or Gaussian distribution) plays a crucial role in the decoding of prediction positions within the system. Eventually, an algorithm board similar to the Galton board was built to serve as the mathematical foundation of the discovered symmetry.
2502.12537
Finding Optimal Trading History in Reinforcement Learning for Stock Market Trading
cs.LG cs.AI
This paper investigates the optimization of temporal windows in Financial Deep Reinforcement Learning (DRL) models using 2D Convolutional Neural Networks (CNNs). We introduce a novel approach to treating the temporal field as a hyperparameter and examine its impact on model performance across various datasets and feature arrangements. We introduce a new hyperparameter for the CNN policy, proposing that this temporal field can and should be treated as a hyperparameter for these models. We examine the significance of this temporal field by iteratively expanding the window of observations presented to the CNN policy during the deep reinforcement learning process. Our iterative process involves progressively increasing the observation period from two weeks to twelve weeks, allowing us to examine the effects of different temporal windows on the model's performance. This window expansion is implemented in two settings. In one setting, we rearrange the features in the dataset to group them by company, allowing the model to have a full view of company data in its observation window and CNN kernel. In the second setting, we do not group the features by company, and features are arranged by category. Our study reveals that shorter temporal windows are most effective when no feature rearrangement to group per company is in effect. However, the model will utilize longer temporal windows and yield better performance once we introduce the feature rearrangement. To examine the consistency of our findings, we repeated our experiment on two datasets containing the same thirty companies from the Dow Jones Index but with different features in each dataset and consistently observed the above-mentioned patterns. The result is a trading model significantly outperforming global financial services firms such as the Global X Guru by the established Mirae Asset.
2502.12539
Design and Implementation of a Dual Uncrewed Surface Vessel Platform for Bathymetry Research under High-flow Conditions
cs.RO cs.LG cs.SY eess.SY
Bathymetry, the study of underwater topography, relies on sonar mapping of submerged structures. These measurements, critical for infrastructure health monitoring, often require expensive instrumentation. The high financial risk associated with sensor damage or vessel loss creates a reluctance to deploy uncrewed surface vessels (USVs) for bathymetry. However, the crewed-boat bathymetry operations, are costly, pose hazards to personnel, and frequently fail to achieve the stable conditions necessary for bathymetry data collection, especially under high currents. Further research is essential to advance autonomous control, navigation, and data processing technologies, with a particular focus on bathymetry. There is a notable lack of accessible hardware platforms that allow for integrated research in both bathymetry-focused autonomous control and navigation, as well as data evaluation and processing. This paper addresses this gap through the design and implementation of two complementary USV systems tailored for uncrewed bathymetry research. This includes a low-cost USV for Navigation And Control research (NAC-USV) and a second, high-end USV equipped with a high-resolution multi-beam sonar and the associated hardware for Bathymetry data quality Evaluation and Post-processing research (BEP-USV). The NAC-USV facilitates the investigation of autonomous, fail-safe navigation and control, emphasizing the stability requirements for high-quality bathymetry data collection while minimizing the risk to equipment. The BEP-USV, which mirrors the NAC-USV hardware, is then used for additional control validation and in-depth exploration of bathymetry data evaluation and post-processing methodologies. We detail the design and implementation of both systems, and open source the design. Furthermore, we demonstrate the system's effectiveness in a range of operational scenarios.
2502.12541
When Segmentation Meets Hyperspectral Image: New Paradigm for Hyperspectral Image Classification
cs.CV
Hyperspectral image (HSI) classification is a cornerstone of remote sensing, enabling precise material and land-cover identification through rich spectral information. While deep learning has driven significant progress in this task, small patch-based classifiers, which account for over 90% of the progress, face limitations: (1) the small patch (e.g., 7x7, 9x9)-based sampling approach considers a limited receptive field, resulting in insufficient spatial structural information critical for object-level identification and noise-like misclassifications even within uniform regions; (2) undefined optimal patch sizes lead to coarse label predictions, which degrade performance; and (3) a lack of multi-shape awareness around objects. To address these challenges, we draw inspiration from large-scale image segmentation techniques, which excel at handling object boundaries-a capability essential for semantic labeling in HSI classification. However, their application remains under-explored in this task due to (1) the prevailing notion that larger patch sizes degrade performance, (2) the extensive unlabeled regions in HSI groundtruth, and (3) the misalignment of input shapes between HSI data and segmentation models. Thus, in this study, we propose a novel paradigm and baseline, HSIseg, for HSI classification that leverages segmentation techniques combined with a novel Dynamic Shifted Regional Transformer (DSRT) to overcome these challenges. We also introduce an intuitive progressive learning framework with adaptive pseudo-labeling to iteratively incorporate unlabeled regions into the training process, thereby advancing the application of segmentation techniques. Additionally, we incorporate auxiliary data through multi-source data collaboration, promoting better feature interaction. Validated on five public HSI datasets, our proposal outperforms state-of-the-art methods.
2502.12542
Computing Voting Rules with Improvement Feedback
cs.GT cs.AI
Aggregating preferences under incomplete or constrained feedback is a fundamental problem in social choice and related domains. While prior work has established strong impossibility results for pairwise comparisons, this paper extends the inquiry to improvement feedback, where voters express incremental adjustments rather than complete preferences. We provide a complete characterization of the positional scoring rules that can be computed given improvement feedback. Interestingly, while plurality is learnable under improvement feedback--unlike with pairwise feedback--strong impossibility results persist for many other positional scoring rules. Furthermore, we show that improvement feedback, unlike pairwise feedback, does not suffice for the computation of any Condorcet-consistent rule. We complement our theoretical findings with experimental results, providing further insights into the practical implications of improvement feedback for preference aggregation.
2502.12545
IM360: Textured Mesh Reconstruction for Large-scale Indoor Mapping with 360$^\circ$ Cameras
cs.CV
We present a novel 3D reconstruction pipeline for 360$^\circ$ cameras for 3D mapping and rendering of indoor environments. Traditional Structure-from-Motion (SfM) methods may not work well in large-scale indoor scenes due to the prevalence of textureless and repetitive regions. To overcome these challenges, our approach (IM360) leverages the wide field of view of omnidirectional images and integrates the spherical camera model into every core component of the SfM pipeline. In order to develop a comprehensive 3D reconstruction solution, we integrate a neural implicit surface reconstruction technique to generate high-quality surfaces from sparse input data. Additionally, we utilize a mesh-based neural rendering approach to refine texture maps and accurately capture view-dependent properties by combining diffuse and specular components. We evaluate our pipeline on large-scale indoor scenes from the Matterport3D and Stanford2D3D datasets. In practice, IM360 demonstrate superior performance in terms of textured mesh reconstruction over SOTA. We observe accuracy improvements in terms of camera localization and registration as well as rendering high frequency details.
2502.12546
Spatiotemporal Multi-Camera Calibration using Freely Moving People
cs.CV
We propose a novel method for spatiotemporal multi-camera calibration using freely moving people in multiview videos. Since calibrating multiple cameras and finding matches across their views are inherently interdependent, performing both in a unified framework poses a significant challenge. We address these issues as a single registration problem of matching two sets of 3D points, leveraging human motion in dynamic multi-person scenes. To this end, we utilize 3D human poses obtained from an off-the-shelf monocular 3D human pose estimator and transform them into 3D points on a unit sphere, to solve the rotation, time offset, and the association alternatingly. We employ a probabilistic approach that can jointly solve both problems of aligning spatiotemporal data and establishing correspondences through soft assignment between two views. The translation is determined by applying coplanarity constraints. The pairwise registration results are integrated into a multiview setup, and then a nonlinear optimization method is used to improve the accuracy of the camera poses, temporal offsets, and multi-person associations. Extensive experiments on synthetic and real data demonstrate the effectiveness and flexibility of the proposed method as a practical marker-free calibration tool.
2502.12548
Improving the Stability of GNN Force Field Models by Reducing Feature Correlation
cs.LG cs.AI
Recently, Graph Neural Network based Force Field (GNNFF) models are widely used in Molecular Dynamics (MD) simulation, which is one of the most cost-effective means in semiconductor material research. However, even such models provide high accuracy in energy and force Mean Absolute Error (MAE) over trained (in-distribution) datasets, they often become unstable during long-time MD simulation when used for out-of-distribution datasets. In this paper, we propose a feature correlation based method for GNNFF models to enhance the stability of MD simulation. We reveal the negative relationship between feature correlation and the stability of GNNFF models, and design a loss function with a dynamic loss coefficient scheduler to reduce edge feature correlation that can be applied in general GNNFF training. We also propose an empirical metric to evaluate the stability in MD simulation. Experiments show our method can significantly improve stability for GNNFF models especially in out-of-distribution data with less than 3% computational overhead. For example, we can ensure the stable MD simulation time from 0.03ps to 10ps for Allegro model.
2502.12552
LLM Safety for Children
cs.CY cs.AI
This paper analyzes the safety of Large Language Models (LLMs) in interactions with children below age of 18 years. Despite the transformative applications of LLMs in various aspects of children's lives such as education and therapy, there remains a significant gap in understanding and mitigating potential content harms specific to this demographic. The study acknowledges the diverse nature of children often overlooked by standard safety evaluations and proposes a comprehensive approach to evaluating LLM safety specifically for children. We list down potential risks that children may encounter when using LLM powered applications. Additionally we develop Child User Models that reflect the varied personalities and interests of children informed by literature in child care and psychology. These user models aim to bridge the existing gap in child safety literature across various fields. We utilize Child User Models to evaluate the safety of six state of the art LLMs. Our observations reveal significant safety gaps in LLMs particularly in categories harmful to children but not adults
2502.12555
Warm Starting of CMA-ES for Contextual Optimization Problems
cs.NE
Several practical applications of evolutionary computation possess objective functions that receive the design variables and externally given parameters. Such problems are termed contextual optimization problems. These problems require finding the optimal solutions corresponding to the given context vectors. Existing contextual optimization methods train a policy model to predict the optimal solution from context vectors. However, the performance of such models is limited by their representation ability. By contrast, warm starting methods have been used to initialize evolutionary algorithms on a given problem using the optimization results on similar problems. Because warm starting methods do not consider the context vectors, their performances can be improved on contextual optimization problems. Herein, we propose a covariance matrix adaptation evolution strategy with contextual warm starting (CMA-ES-CWS) to efficiently optimize the contextual optimization problem with a given context vector. The CMA-ES-CWS utilizes the optimization results of past context vectors to train the multivariate Gaussian process regression. Subsequently, the CMA-ES-CWS performs warm starting for a given context vector by initializing the search distribution using posterior distribution of the Gaussian process regression. The results of the numerical simulation suggest that CMA-ES-CWS outperforms the existing contextual optimization and warm starting methods.
2502.12556
From Maneuver to Mishap: A Systematic Literature Review on U-Turn Safety Risks
eess.SY cs.SY
Understanding the impacts of U-turn configurations on intersection safety and traffic operations is essential for developing effective strategies to enhance road safety and efficiency. Extensive research has been conducted to investigate the role of geometric designs, driver behavior, and advanced technologies in mitigating crash risks and improving traffic flow at U-turn facilities. By synthesizing this collective body of work through the guidelines of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), this paper provides a valuable resource for transportation professionals, policymakers, and researchers seeking evidence-based solutions. This systematic review draws on studies from diverse traffic environments and regional contexts, focusing on innovative design interventions, such as restricted crossing U-turns (RCUTs) and median U-turn intersections (MUTs), as well as integrated strategies leveraging technological advancements. By presenting a comprehensive analysis of U-turn-related challenges and opportunities, this review contributes to advancing transportation safety research and guiding the development of adaptive strategies tailored to varied traffic conditions and evolving technologies.
2502.12558
MomentSeeker: A Comprehensive Benchmark and A Strong Baseline For Moment Retrieval Within Long Videos
cs.CV cs.AI
Retrieval augmented generation (RAG) holds great promise in addressing challenges associated with long video understanding. These methods retrieve useful moments from long videos for their presented tasks, thereby enabling multimodal large language models (MLLMs) to generate high-quality answers in a cost-effective way. In this work, we present MomentSeeker, a comprehensive benchmark to evaluate retrieval models' performance in handling general long-video moment retrieval (LVMR) tasks. MomentSeeker offers three key advantages. First, it incorporates long videos of over 500 seconds on average, making it the first benchmark specialized for long-video moment retrieval. Second, it covers a wide range of task categories (including Moment Search, Caption Alignment, Image-conditioned Moment Search, and Video-conditioned Moment Search) and diverse application scenarios (e.g., sports, movies, cartoons, and ego), making it a comprehensive tool for assessing retrieval models' general LVMR performance. Additionally, the evaluation tasks are carefully curated through human annotation, ensuring the reliability of assessment. We further fine-tune an MLLM-based LVMR retriever on synthetic data, which demonstrates strong performance on our benchmark. We perform extensive experiments with various popular multimodal retrievers based on our benchmark, whose results highlight the challenges of LVMR and limitations for existing methods. Our created resources will be shared with community to advance future research in this field.
2502.12560
How does a Language-Specific Tokenizer affect LLMs?
cs.CL
The necessity of language-specific tokenizers intuitively appears crucial for effective natural language processing, yet empirical analyses on their significance and underlying reasons are lacking. This study explores how language-specific tokenizers influence the behavior of Large Language Models predominantly trained with English text data, through the case study of Korean. The research unfolds in two main stages: (1) the development of a Korean-specific extended tokenizer and (2) experiments to compare models with the basic tokenizer and the extended tokenizer through various Next Token Prediction tasks. Our in-depth analysis reveals that the extended tokenizer decreases confidence in incorrect predictions during generation and reduces cross-entropy in complex tasks, indicating a tendency to produce less nonsensical outputs. Consequently, the extended tokenizer provides stability during generation, potentially leading to higher performance in downstream tasks.
2502.12561
UXAgent: An LLM Agent-Based Usability Testing Framework for Web Design
cs.HC cs.CL
Usability testing is a fundamental yet challenging (e.g., inflexible to iterate the study design flaws and hard to recruit study participants) research method for user experience (UX) researchers to evaluate a web design. Recent advances in Large Language Model-simulated Agent (LLM-Agent) research inspired us to design UXAgent to support UX researchers in evaluating and reiterating their usability testing study design before they conduct the real human subject study. Our system features an LLM-Agent module and a universal browser connector module so that UX researchers can automatically generate thousands of simulated users to test the target website. The results are shown in qualitative (e.g., interviewing how an agent thinks ), quantitative (e.g., # of actions), and video recording formats for UX researchers to analyze. Through a heuristic user evaluation with five UX researchers, participants praised the innovation of our system but also expressed concerns about the future of LLM Agent-assisted UX study.
2502.12562
SEA: Low-Resource Safety Alignment for Multimodal Large Language Models via Synthetic Embeddings
cs.CL cs.CR cs.MM
Multimodal Large Language Models (MLLMs) have serious security vulnerabilities.While safety alignment using multimodal datasets consisting of text and data of additional modalities can effectively enhance MLLM's security, it is costly to construct these datasets. Existing low-resource security alignment methods, including textual alignment, have been found to struggle with the security risks posed by additional modalities. To address this, we propose Synthetic Embedding augmented safety Alignment (SEA), which optimizes embeddings of additional modality through gradient updates to expand textual datasets. This enables multimodal safety alignment training even when only textual data is available. Extensive experiments on image, video, and audio-based MLLMs demonstrate that SEA can synthesize a high-quality embedding on a single RTX3090 GPU within 24 seconds. SEA significantly improves the security of MLLMs when faced with threats from additional modalities. To assess the security risks introduced by video and audio, we also introduced a new benchmark called VA-SafetyBench. High attack success rates across multiple MLLMs validate its challenge. Our code and data will be available at https://github.com/ZeroNLP/SEA.
2502.12563
Evaluating Language Models on Grooming Risk Estimation Using Fuzzy Theory
cs.CL cs.AI cs.LG
Encoding implicit language presents a challenge for language models, especially in high-risk domains where maintaining high precision is important. Automated detection of online child grooming is one such critical domain, where predators manipulate victims using a combination of explicit and implicit language to convey harmful intentions. While recent studies have shown the potential of Transformer language models like SBERT for preemptive grooming detection, they primarily depend on surface-level features and approximate real victim grooming processes using vigilante and law enforcement conversations. The question of whether these features and approximations are reasonable has not been addressed thus far. In this paper, we address this gap and study whether SBERT can effectively discern varying degrees of grooming risk inherent in conversations, and evaluate its results across different participant groups. Our analysis reveals that while fine-tuning aids language models in learning to assign grooming scores, they show high variance in predictions, especially for contexts containing higher degrees of grooming risk. These errors appear in cases that 1) utilize indirect speech pathways to manipulate victims and 2) lack sexually explicit content. This finding underscores the necessity for robust modeling of indirect speech acts by language models, particularly those employed by predators.
2502.12564
Sample Efficient Omniprediction and Downstream Swap Regret for Non-Linear Losses
cs.LG cs.GT
We define "decision swap regret" which generalizes both prediction for downstream swap regret and omniprediction, and give algorithms for obtaining it for arbitrary multi-dimensional Lipschitz loss functions in online adversarial settings. We also give sample complexity bounds in the batch setting via an online-to-batch reduction. When applied to omniprediction, our algorithm gives the first polynomial sample-complexity bounds for Lipschitz loss functions -- prior bounds either applied only to linear loss (or binary outcomes) or scaled exponentially with the error parameter even under the assumption that the loss functions were convex. When applied to prediction for downstream regret, we give the first algorithm capable of guaranteeing swap regret bounds for all downstream agents with non-linear loss functions over a multi-dimensional outcome space: prior work applied only to linear loss functions, modeling risk neutral agents. Our general bounds scale exponentially with the dimension of the outcome space, but we give improved regret and sample complexity bounds for specific families of multidimensional functions of economic interest: constant elasticity of substitution (CES), Cobb-Douglas, and Leontief utility functions.
2502.12565
Self Iterative Label Refinement via Robust Unlabeled Learning
cs.CL
Recent advances in large language models (LLMs) have yielded impressive performance on various tasks, yet they often depend on high-quality feedback that can be costly. Self-refinement methods attempt to leverage LLMs' internal evaluation mechanisms with minimal human supervision; however, these approaches frequently suffer from inherent biases and overconfidence, especially in domains where the models lack sufficient internal knowledge, resulting in performance degradation. As an initial step toward enhancing self-refinement for broader applications, we introduce an iterative refinement pipeline that employs the Unlabeled-Unlabeled learning framework to improve LLM-generated pseudo-labels for classification tasks. By exploiting two unlabeled datasets with differing positive class ratios, our approach iteratively denoises and refines the initial pseudo-labels, thereby mitigating the adverse effects of internal biases with minimal human supervision. Evaluations on diverse datasets, including low-resource language corpora, patent classifications, and protein structure categorizations, demonstrate that our method consistently outperforms both initial LLM's classification performance and the self-refinement approaches by cutting-edge models (e.g., GPT-4o and DeepSeek-R1).
2502.12566
Exploring the Impact of Personality Traits on LLM Bias and Toxicity
cs.AI
With the different roles that AI is expected to play in human life, imbuing large language models (LLMs) with different personalities has attracted increasing research interests. While the "personification" enhances human experiences of interactivity and adaptability of LLMs, it gives rise to critical concerns about content safety, particularly regarding bias, sentiment and toxicity of LLM generation. This study explores how assigning different personality traits to LLMs affects the toxicity and biases of their outputs. Leveraging the widely accepted HEXACO personality framework developed in social psychology, we design experimentally sound prompts to test three LLMs' performance on three toxic and bias benchmarks. The findings demonstrate the sensitivity of all three models to HEXACO personality traits and, more importantly, a consistent variation in the biases, negative sentiment and toxicity of their output. In particular, adjusting the levels of several personality traits can effectively reduce bias and toxicity in model performance, similar to humans' correlations between personality traits and toxic behaviors. The findings highlight the additional need to examine content safety besides the efficiency of training or fine-tuning methods for LLM personification. They also suggest a potential for the adjustment of personalities to be a simple and low-cost method to conduct controlled text generation.
2502.12567
DeltaDiff: A Residual-Guided Diffusion Model for Enhanced Image Super-Resolution
cs.CV
Recently, the application of diffusion models in super-resolution tasks has become a popular research direction. Existing work is focused on fully migrating diffusion models to SR tasks. The diffusion model is proposed in the field of image generation, so in order to make the generated results diverse, the diffusion model combines random Gaussian noise and distributed sampling to increase the randomness of the model. However, the essence of super-resolution tasks requires the model to generate high-resolution images with fidelity. Excessive addition of random factors can result in the model generating detailed information that does not belong to the HR image. To address this issue, we propose a new diffusion model called Deltadiff, which uses only residuals between images for diffusion, making the entire diffusion process more stable. The experimental results show that our method surpasses state-of-the-art models and generates results with better fidelity. Our code and model are publicly available at https://github.com/continueyang/DeltaDiff
2502.12568
A Cognitive Writing Perspective for Constrained Long-Form Text Generation
cs.CL cs.AI
Like humans, Large Language Models (LLMs) struggle to generate high-quality long-form text that adheres to strict requirements in a single pass. This challenge is unsurprising, as successful human writing, according to the Cognitive Writing Theory, is a complex cognitive process involving iterative planning, translating, reviewing, and monitoring. Motivated by these cognitive principles, we aim to equip LLMs with human-like cognitive writing capabilities through CogWriter, a novel training-free framework that transforms LLM constrained long-form text generation into a systematic cognitive writing paradigm. Our framework consists of two key modules: (1) a Planning Agent that performs hierarchical planning to decompose the task, and (2) multiple Generation Agents that execute these plans in parallel. The system maintains quality via continuous monitoring and reviewing mechanisms, which evaluate outputs against specified requirements and trigger necessary revisions. CogWriter demonstrates exceptional performance on LongGenBench, a benchmark for complex constrained long-form text generation. Even when using Qwen-2.5-14B as its backbone, CogWriter surpasses GPT-4o by 22% in complex instruction completion accuracy while reliably generating texts exceeding 10,000 words. We hope this cognitive science-inspired approach provides a paradigm for LLM writing advancements: \href{https://github.com/KaiyangWan/CogWriter}{CogWriter}.
2502.12569
Maximizing Value in Challenge the Champ Tournaments
cs.DS cs.GT cs.MA
A tournament is a method to decide the winner in a competition, and describes the overall sequence in which matches between the players are held. While deciding a worthy winner is the primary goal of a tournament, a close second is to maximize the value generated for the matches played, with value for a match measured either in terms of tickets sold, television viewership, advertising revenue, or other means. Tournament organizers often seed the players -- i.e., decide which matches are played -- to increase this value. We study the value maximization objective in a particular tournament format called Challenge the Champ. This is a simple tournament format where an ordering of the players is decided. The first player in this order is the initial champion. The remaining players in order challenge the current champion; if a challenger wins, she replaces the current champion. We model the outcome of a match between two players using a complete directed graph, called a strength graph, with each player represented as a vertex, and the direction of an edge indicating the winner in a match. The value-maximization objective has been recently explored for knockout tournaments when the strength graph is a directed acyclic graph (DAG). We extend the investigation to Challenge the Champ tournaments and general strength graphs. We study different representations of the value of each match, and completely characterize the computational complexity of the problem.
2502.12570
GVTNet: Graph Vision Transformer For Face Super-Resolution
cs.CV
Recent advances in face super-resolution research have utilized the Transformer architecture. This method processes the input image into a series of small patches. However, because of the strong correlation between different facial components in facial images. When it comes to super-resolution of low-resolution images, existing algorithms cannot handle the relationships between patches well, resulting in distorted facial components in the super-resolution results. To solve the problem, we propose a transformer architecture based on graph neural networks called graph vision transformer network. We treat each patch as a graph node and establish an adjacency matrix based on the information between patches. In this way, the patch only interacts between neighboring patches, further processing the relationship of facial components. Quantitative and visualization experiments have underscored the superiority of our algorithm over state-of-the-art techniques. Through detailed comparisons, we have demonstrated that our algorithm possesses more advanced super-resolution capabilities, particularly in enhancing facial components. The PyTorch code is available at https://github.com/continueyang/GVTNet
2502.12571
A Novel Gain Modeling Technique for LLC Resonant Converters based on The Hybrid Deep-Learning/GMDH Neural Network
eess.SY cs.SY
This paper presents a novel hybrid approach for modeling the voltage gain of LLC resonant converters by combining deep-learning neural networks with the polynomial based Group Method of Data Handling (GMDH). While deep learning offers high accuracy in predicting nonlinear converter behavior, it produces complex network models. GMDH neural networks, in contrast, yield simpler algebraic equations that can be more convenient in converter design. By training a deep network on data from an FPGA based real time simulator and then using the network s predictions to train a GMDH model, the proposed hybrid method achieves both high accuracy and design friendly simplicity. Experimental results show significant improvements over traditional methods such as First Harmonic Approximation (FHA) and frequency domain corrections, particularly for wide operating ranges.
2502.12574
HeadInfer: Memory-Efficient LLM Inference by Head-wise Offloading
cs.LG cs.AI
Transformer-based large language models (LLMs) demonstrate impressive performance in long context generation. Extending the context length has disproportionately shifted the memory footprint of LLMs during inference to the key-value cache (KV cache). In this paper, we propose HEADINFER, which offloads the KV cache to CPU RAM while avoiding the need to fully store the KV cache for any transformer layer on the GPU. HEADINFER employs a fine-grained, head-wise offloading strategy, maintaining only selective attention heads KV cache on the GPU while computing attention output dynamically. Through roofline analysis, we demonstrate that HEADINFER maintains computational efficiency while significantly reducing memory footprint. We evaluate HEADINFER on the Llama-3-8B model with a 1-million-token sequence, reducing the GPU memory footprint of the KV cache from 128 GB to 1 GB and the total GPU memory usage from 207 GB to 17 GB, achieving a 92% reduction compared to BF16 baseline inference. Notably, HEADINFER enables 4-million-token inference with an 8B model on a single consumer GPU with 24GB memory (e.g., NVIDIA RTX 4090) without approximation methods.
2502.12575
DemonAgent: Dynamically Encrypted Multi-Backdoor Implantation Attack on LLM-based Agent
cs.CR cs.AI
As LLM-based agents become increasingly prevalent, backdoors can be implanted into agents through user queries or environment feedback, raising critical concerns regarding safety vulnerabilities. However, backdoor attacks are typically detectable by safety audits that analyze the reasoning process of agents. To this end, we propose a novel backdoor implantation strategy called \textbf{Dynamically Encrypted Multi-Backdoor Implantation Attack}. Specifically, we introduce dynamic encryption, which maps the backdoor into benign content, effectively circumventing safety audits. To enhance stealthiness, we further decompose the backdoor into multiple sub-backdoor fragments. Based on these advancements, backdoors are allowed to bypass safety audits significantly. Additionally, we present AgentBackdoorEval, a dataset designed for the comprehensive evaluation of agent backdoor attacks. Experimental results across multiple datasets demonstrate that our method achieves an attack success rate nearing 100\% while maintaining a detection rate of 0\%, illustrating its effectiveness in evading safety audits. Our findings highlight the limitations of existing safety mechanisms in detecting advanced attacks, underscoring the urgent need for more robust defenses against backdoor threats. Code and data are available at https://github.com/whfeLingYu/DemonAgent.
2502.12576
A Fuzzy Evaluation of Sentence Encoders on Grooming Risk Classification
cs.CL cs.AI cs.LG
With the advent of social media, children are becoming increasingly vulnerable to the risk of grooming in online settings. Detecting grooming instances in an online conversation poses a significant challenge as the interactions are not necessarily sexually explicit, since the predators take time to build trust and a relationship with their victim. Moreover, predators evade detection using indirect and coded language. While previous studies have fine-tuned Transformers to automatically identify grooming in chat conversations, they overlook the impact of coded and indirect language on model predictions, and how these align with human perceptions of grooming. In this paper, we address this gap and evaluate bi-encoders on the task of classifying different degrees of grooming risk in chat contexts, for three different participant groups, i.e. law enforcement officers, real victims, and decoys. Using a fuzzy-theoretic framework, we map human assessments of grooming behaviors to estimate the actual degree of grooming risk. Our analysis reveals that fine-tuned models fail to tag instances where the predator uses indirect speech pathways and coded language to evade detection. Further, we find that such instances are characterized by a higher presence of out-of-vocabulary (OOV) words in samples, causing the model to misclassify. Our findings highlight the need for more robust models to identify coded language from noisy chat inputs in grooming contexts.
2502.12579
CHATS: Combining Human-Aligned Optimization and Test-Time Sampling for Text-to-Image Generation
cs.CV
Diffusion models have emerged as a dominant approach for text-to-image generation. Key components such as the human preference alignment and classifier-free guidance play a crucial role in ensuring generation quality. However, their independent application in current text-to-image models continues to face significant challenges in achieving strong text-image alignment, high generation quality, and consistency with human aesthetic standards. In this work, we for the first time, explore facilitating the collaboration of human performance alignment and test-time sampling to unlock the potential of text-to-image models. Consequently, we introduce CHATS (Combining Human-Aligned optimization and Test-time Sampling), a novel generative framework that separately models the preferred and dispreferred distributions and employs a proxy-prompt-based sampling strategy to utilize the useful information contained in both distributions. We observe that CHATS exhibits exceptional data efficiency, achieving strong performance with only a small, high-quality funetuning dataset. Extensive experiments demonstrate that CHATS surpasses traditional preference alignment methods, setting new state-of-the-art across various standard benchmarks.
2502.12581
The Majority Vote Paradigm Shift: When Popular Meets Optimal
stat.ML cs.AI cs.LG
Reliably labelling data typically requires annotations from multiple human workers. However, humans are far from being perfect. Hence, it is a common practice to aggregate labels gathered from multiple annotators to make a more confident estimate of the true label. Among many aggregation methods, the simple and well known Majority Vote (MV) selects the class label polling the highest number of votes. However, despite its importance, the optimality of MV's label aggregation has not been extensively studied. We address this gap in our work by characterising the conditions under which MV achieves the theoretically optimal lower bound on label estimation error. Our results capture the tolerable limits on annotation noise under which MV can optimally recover labels for a given class distribution. This certificate of optimality provides a more principled approach to model selection for label aggregation as an alternative to otherwise inefficient practices that sometimes include higher experts, gold labels, etc., that are all marred by the same human uncertainty despite huge time and monetary costs. Experiments on both synthetic and real world data corroborate our theoretical findings.
2502.12582
Adaptive Prototype Model for Attribute-based Multi-label Few-shot Action Recognition
cs.CV
In real-world action recognition systems, incorporating more attributes helps achieve a more comprehensive understanding of human behavior. However, using a single model to simultaneously recognize multiple attributes can lead to a decrease in accuracy. In this work, we propose a novel method i.e. Adaptive Attribute Prototype Model (AAPM) for human action recognition, which captures rich action-relevant attribute information and strikes a balance between accuracy and robustness. Firstly, we introduce the Text-Constrain Module (TCM) to incorporate textual information from potential labels, and constrain the construction of different attributes prototype representations. In addition, we explore the Attribute Assignment Method (AAM) to address the issue of training bias and increase robustness during the training process.Furthermore, we construct a new video dataset with attribute-based multi-label called Multi-Kinetics for evaluation, which contains various attribute labels (e.g. action, scene, object, etc.) related to human behavior. Extensive experiments demonstrate that our AAPM achieves the state-of-the-art performance in both attribute-based multi-label few-shot action recognition and single-label few-shot action recognition. The project and dataset are available at an anonymous account https://github.com/theAAPM/AAPM
2502.12583
LongFaith: Enhancing Long-Context Reasoning in LLMs with Faithful Synthetic Data
cs.CL
Despite the growing development of long-context large language models (LLMs), data-centric approaches relying on synthetic data have been hindered by issues related to faithfulness, which limit their effectiveness in enhancing model performance on tasks such as long-context reasoning and question answering (QA). These challenges are often exacerbated by misinformation caused by lack of verification, reasoning without attribution, and potential knowledge conflicts. We propose LongFaith, a novel pipeline for synthesizing faithful long-context reasoning instruction datasets. By integrating ground truth and citation-based reasoning prompts, we eliminate distractions and improve the accuracy of reasoning chains, thus mitigating the need for costly verification processes. We open-source two synthesized datasets, LongFaith-SFT and LongFaith-PO, which systematically address multiple dimensions of faithfulness, including verified reasoning, attribution, and contextual grounding. Extensive experiments on multi-hop reasoning datasets and LongBench demonstrate that models fine-tuned on these datasets significantly improve performance. Our ablation studies highlight the scalability and adaptability of the LongFaith pipeline, showcasing its broad applicability in developing long-context LLMs.
2502.12584
Enhancing Semi-supervised Learning with Noisy Zero-shot Pseudolabels
cs.LG cs.AI
Semi-supervised learning (SSL) leverages limited labeled data alongside abundant unlabeled data to address labeling costs in machine learning. While recent foundation models enable zero-shot inference, attempts to integrate these capabilities into SSL through pseudo-labeling have shown mixed results due to unreliable zero-shot predictions. We present ZMT (Zero-Shot Multi-Task Learning), a framework that jointly optimizes zero-shot pseudo-labels and unsupervised representation learning objectives from contemporary SSL approaches. Our method introduces a multi-task learning-based mechanism that incorporates pseudo-labels while ensuring robustness to varying pseudo-label quality. Experiments across 8 datasets in vision, language, and audio domains demonstrate that ZMT reduces error by up to 56% compared to traditional SSL methods, with particularly compelling results when pseudo-labels are noisy and unreliable. ZMT represents a significant step toward making semi-supervised learning more effective and accessible in resource-constrained environments.
2502.12586
G-Refer: Graph Retrieval-Augmented Large Language Model for Explainable Recommendation
cs.IR cs.CL
Explainable recommendation has demonstrated significant advantages in informing users about the logic behind recommendations, thereby increasing system transparency, effectiveness, and trustworthiness. To provide personalized and interpretable explanations, existing works often combine the generation capabilities of large language models (LLMs) with collaborative filtering (CF) information. CF information extracted from the user-item interaction graph captures the user behaviors and preferences, which is crucial for providing informative explanations. However, due to the complexity of graph structure, effectively extracting the CF information from graphs still remains a challenge. Moreover, existing methods often struggle with the integration of extracted CF information with LLMs due to its implicit representation and the modality gap between graph structures and natural language explanations. To address these challenges, we propose G-Refer, a framework using graph retrieval-augmented large language models (LLMs) for explainable recommendation. Specifically, we first employ a hybrid graph retrieval mechanism to retrieve explicit CF signals from both structural and semantic perspectives. The retrieved CF information is explicitly formulated as human-understandable text by the proposed graph translation and accounts for the explanations generated by LLMs. To bridge the modality gap, we introduce knowledge pruning and retrieval-augmented fine-tuning to enhance the ability of LLMs to process and utilize the retrieved CF information to generate explanations. Extensive experiments show that G-Refer achieves superior performance compared with existing methods in both explainability and stability. Codes and data are available at https://github.com/Yuhan1i/G-Refer.
2502.12587
RSMLP: A light Sampled MLP Structure for Incomplete Utterance Rewrite
cs.CL cs.AI
The Incomplete Utterance Rewriting (IUR) task has garnered significant attention in recent years. Its goal is to reconstruct conversational utterances to better align with the current context, thereby enhancing comprehension. In this paper, we introduce a novel and versatile lightweight method, Rewritten-Sampled MLP (RSMLP). By employing an MLP based architecture with a carefully designed down-sampling strategy, RSMLP effectively extracts latent semantic information between utterances and makes appropriate edits to restore incomplete utterances. Due to its simple yet efficient structure, our method achieves competitive performance on public IUR datasets and in real-world applications.
2502.12589
RM-PoT: Reformulating Mathematical Problems and Solving via Program of Thoughts
cs.AI
Recently, substantial advancements have been made in training language models to carry out step-by-step reasoning for solving intricate numerical reasoning tasks. Beyond the methods used to solve these problems, the structure and formulation of the problems themselves also play a crucial role in determining the performance of large language models. We observe that even small changes in the surface form of mathematical problems can have a profound impact on both the answer distribution and solve rate. This highlights the vulnerability of LLMs to surface-level variations, revealing its limited robustness when reasoning through complex problems. In this paper, we propose RM-PoT, a three-stage framework that integrates problem reformulation (RM), code-aided reasoning (PoT), and domain-aware few-shot learning to address these limitations. Our approach first reformulates the input problem into diverse surface forms to reduce structural bias, then retrieves five semantically aligned examples from a pre-constructed domain-specific question bank to provide contextual guidance, and finally generates executable Python code for precise computation.
2502.12591
CutPaste&Find: Efficient Multimodal Hallucination Detector with Visual-aid Knowledge Base
cs.CV cs.CL
Large Vision-Language Models (LVLMs) have demonstrated impressive multimodal reasoning capabilities, but they remain susceptible to hallucination, particularly object hallucination where non-existent objects or incorrect attributes are fabricated in generated descriptions. Existing detection methods achieve strong performance but rely heavily on expensive API calls and iterative LVLM-based validation, making them impractical for large-scale or offline use. To address these limitations, we propose CutPaste\&Find, a lightweight and training-free framework for detecting hallucinations in LVLM-generated outputs. Our approach leverages off-the-shelf visual and linguistic modules to perform multi-step verification efficiently without requiring LVLM inference. At the core of our framework is a Visual-aid Knowledge Base that encodes rich entity-attribute relationships and associated image representations. We introduce a scaling factor to refine similarity scores, mitigating the issue of suboptimal alignment values even for ground-truth image-text pairs. Comprehensive evaluations on benchmark datasets, including POPE and R-Bench, demonstrate that CutPaste\&Find achieves competitive hallucination detection performance while being significantly more efficient and cost-effective than previous methods.
2502.12594
PASER: Post-Training Data Selection for Efficient Pruned Large Language Model Recovery
cs.CL
Model pruning is an effective approach for compressing large language models. However, this process often leads to significant degradation of model capabilities. While post-training techniques such as instruction tuning are commonly employed to recover model performance, existing methods often overlook the uneven deterioration of model capabilities and incur high computational costs. Moreover, some instruction data irrelevant to model capability recovery may introduce negative effects. To address these challenges, we propose the \textbf{P}ost-training d\textbf{A}ta \textbf{S}election method for \textbf{E}fficient pruned large language model \textbf{R}ecovery (\textbf{PASER}). PASER aims to identify instructions where model capabilities are most severely compromised within a certain recovery data budget. Our approach first applies manifold learning and spectral clustering to group recovery data in the semantic space, revealing capability-specific instruction sets. We then adaptively allocate the data budget to different clusters based on the degrees of model capability degradation. In each cluster, we prioritize data samples where model performance has declined dramatically. To mitigate potential negative transfer, we also detect and filter out conflicting or irrelevant recovery data. Extensive experiments demonstrate that PASER significantly outperforms conventional baselines, effectively recovering the general capabilities of pruned LLMs while utilizing merely 4\%-20\% of the original post-training data.
2502.12598
Bring Your Own Knowledge: A Survey of Methods for LLM Knowledge Expansion
cs.CL
Adapting large language models (LLMs) to new and diverse knowledge is essential for their lasting effectiveness in real-world applications. This survey provides an overview of state-of-the-art methods for expanding the knowledge of LLMs, focusing on integrating various knowledge types, including factual information, domain expertise, language proficiency, and user preferences. We explore techniques, such as continual learning, model editing, and retrieval-based explicit adaptation, while discussing challenges like knowledge consistency and scalability. Designed as a guide for researchers and practitioners, this survey sheds light on opportunities for advancing LLMs as adaptable and robust knowledge systems.
2502.12599
Learning a High-quality Robotic Wiping Policy Using Systematic Reward Analysis and Visual-Language Model Based Curriculum
cs.RO cs.LG
Autonomous robotic wiping is an important task in various industries, ranging from industrial manufacturing to sanitization in healthcare. Deep reinforcement learning (Deep RL) has emerged as a promising algorithm, however, it often suffers from a high demand for repetitive reward engineering. Instead of relying on manual tuning, we first analyze the convergence of quality-critical robotic wiping, which requires both high-quality wiping and fast task completion, to show the poor convergence of the problem and propose a new bounded reward formulation to make the problem feasible. Then, we further improve the learning process by proposing a novel visual-language model (VLM) based curriculum, which actively monitors the progress and suggests hyperparameter tuning. We demonstrate that the combined method can find a desirable wiping policy on surfaces with various curvatures, frictions, and waypoints, which cannot be learned with the baseline formulation. The demo of this project can be found at: https://sites.google.com/view/highqualitywiping.
2502.12600
Revisiting the Generalization Problem of Low-level Vision Models Through the Lens of Image Deraining
cs.CV
Generalization remains a significant challenge for low-level vision models, which often struggle with unseen degradations in real-world scenarios despite their success in controlled benchmarks. In this paper, we revisit the generalization problem in low-level vision models. Image deraining is selected as a case study due to its well-defined and easily decoupled structure, allowing for more effective observation and analysis. Through comprehensive experiments, we reveal that the generalization issue is not primarily due to limited network capacity but rather the failure of existing training strategies, which leads networks to overfit specific degradation patterns. Our findings show that guiding networks to focus on learning the underlying image content, rather than the degradation patterns, is key to improving generalization. We demonstrate that balancing the complexity of background images and degradations in the training data helps networks better fit the image distribution. Furthermore, incorporating content priors from pre-trained generative models significantly enhances generalization. Experiments on both image deraining and image denoising validate the proposed strategies. We believe the insights and solutions will inspire further research and improve the generalization of low-level vision models.
2502.12601
COPU: Conformal Prediction for Uncertainty Quantification in Natural Language Generation
cs.CL
Uncertainty Quantification (UQ) for Natural Language Generation (NLG) is crucial for assessing the performance of Large Language Models (LLMs), as it reveals confidence in predictions, identifies failure modes, and gauges output reliability. Conformal Prediction (CP), a model-agnostic method that generates prediction sets with a specified error rate, has been adopted for UQ in classification tasks, where the size of the prediction set indicates the model's uncertainty. However, when adapting CP to NLG, the sampling-based method for generating candidate outputs cannot guarantee the inclusion of the ground truth, limiting its applicability across a wide range of error rates. To address this, we propose \ourmethod, a method that explicitly adds the ground truth to the candidate outputs and uses logit scores to measure nonconformity. Our experiments with six LLMs on four NLG tasks show that \ourmethod outperforms baseline methods in calibrating error rates and empirical cover rates, offering accurate UQ across a wide range of user-specified error rates.
2502.12602
Learning-based Dynamic Robot-to-Human Handover
cs.RO
This paper presents a novel learning-based approach to dynamic robot-to-human handover, addressing the challenges of delivering objects to a moving receiver. We hypothesize that dynamic handover, where the robot adjusts to the receiver's movements, results in more efficient and comfortable interaction compared to static handover, where the receiver is assumed to be stationary. To validate this, we developed a nonparametric method for generating continuous handover motion, conditioned on the receiver's movements, and trained the model using a dataset of 1,000 human-to-human handover demonstrations. We integrated preference learning for improved handover effectiveness and applied impedance control to ensure user safety and adaptiveness. The approach was evaluated in both simulation and real-world settings, with user studies demonstrating that dynamic handover significantly reduces handover time and improves user comfort compared to static methods. Videos and demonstrations of our approach are available at https://zerotohero7886.github.io/dyn-r2h-handover .
2502.12603
Disentangling Long-Short Term State Under Unknown Interventions for Online Time Series Forecasting
cs.LG cs.AI
Current methods for time series forecasting struggle in the online scenario, since it is difficult to preserve long-term dependency while adapting short-term changes when data are arriving sequentially. Although some recent methods solve this problem by controlling the updates of latent states, they cannot disentangle the long/short-term states, leading to the inability to effectively adapt to nonstationary. To tackle this challenge, we propose a general framework to disentangle long/short-term states for online time series forecasting. Our idea is inspired by the observations where short-term changes can be led by unknown interventions like abrupt policies in the stock market. Based on this insight, we formalize a data generation process with unknown interventions on short-term states. Under mild assumptions, we further leverage the independence of short-term states led by unknown interventions to establish the identification theory to achieve the disentanglement of long/short-term states. Built on this theory, we develop a long short-term disentanglement model (LSTD) to extract the long/short-term states with long/short-term encoders, respectively. Furthermore, the LSTD model incorporates a smooth constraint to preserve the long-term dependencies and an interrupted dependency constraint to enforce the forgetting of short-term dependencies, together boosting the disentanglement of long/short-term states. Experimental results on several benchmark datasets show that our \textbf{LSTD} model outperforms existing methods for online time series forecasting, validating its efficacy in real-world applications.
2502.12604
S2C: Learning Noise-Resistant Differences for Unsupervised Change Detection in Multimodal Remote Sensing Images
cs.CV
Unsupervised Change Detection (UCD) in multimodal Remote Sensing (RS) images remains a difficult challenge due to the inherent spatio-temporal complexity within data, and the heterogeneity arising from different imaging sensors. Inspired by recent advancements in Visual Foundation Models (VFMs) and Contrastive Learning (CL) methodologies, this research aims to develop CL methodologies to translate implicit knowledge in VFM into change representations, thus eliminating the need for explicit supervision. To this end, we introduce a Semantic-to-Change (S2C) learning framework for UCD in both homogeneous and multimodal RS images. Differently from existing CL methodologies that typically focus on learning multi-temporal similarities, we introduce a novel triplet learning strategy that explicitly models temporal differences, which are crucial to the CD task. Furthermore, random spatial and spectral perturbations are introduced during the training to enhance robustness to temporal noise. In addition, a grid sparsity regularization is defined to suppress insignificant changes, and an IoU-matching algorithm is developed to refine the CD results. Experiments on four benchmark CD datasets demonstrate that the proposed S2C learning framework achieves significant improvements in accuracy, surpassing current state-of-the-art by over 31\%, 9\%, 23\%, and 15\%, respectively. It also demonstrates robustness and sample efficiency, suitable for training and adaptation of various Visual Foundation Models (VFMs) or backbone neural networks. The relevant code will be available at: github.com/DingLei14/S2C.
2502.12605
Hypernetwork-based approach for optimal composition design in partially controlled multi-agent systems
cs.MA cs.LG
Partially Controlled Multi-Agent Systems (PCMAS) are comprised of controllable agents, managed by a system designer, and uncontrollable agents, operating autonomously. This study addresses an optimal composition design problem in PCMAS, which involves the system designer's problem, determining the optimal number and policies of controllable agents, and the uncontrollable agents' problem, identifying their best-response policies. Solving this bi-level optimization problem is computationally intensive, as it requires repeatedly solving multi-agent reinforcement learning problems under various compositions for both types of agents. To address these challenges, we propose a novel hypernetwork-based framework that jointly optimizes the system's composition and agent policies. Unlike traditional methods that train separate policy networks for each composition, the proposed framework generates policies for both controllable and uncontrollable agents through a unified hypernetwork. This approach enables efficient information sharing across similar configurations, thereby reducing computational overhead. Additional improvements are achieved by incorporating reward parameter optimization and mean action networks. Using real-world New York City taxi data, we demonstrate that our framework outperforms existing methods in approximating equilibrium policies. Our experimental results show significant improvements in key performance metrics, such as order response rate and served demand, highlighting the practical utility of controlling agents and their potential to enhance decision-making in PCMAS.
2502.12607
Generalized Kernel Inducing Points by Duality Gap for Dataset Distillation
stat.ML cs.LG
We propose Duality Gap KIP (DGKIP), an extension of the Kernel Inducing Points (KIP) method for dataset distillation. While existing dataset distillation methods often rely on bi-level optimization, DGKIP eliminates the need for such optimization by leveraging duality theory in convex programming. The KIP method has been introduced as a way to avoid bi-level optimization; however, it is limited to the squared loss and does not support other loss functions (e.g., cross-entropy or hinge loss) that are more suitable for classification tasks. DGKIP addresses this limitation by exploiting an upper bound on parameter changes after dataset distillation using the duality gap, enabling its application to a wider range of loss functions. We also characterize theoretical properties of DGKIP by providing upper bounds on the test error and prediction consistency after dataset distillation. Experimental results on standard benchmarks such as MNIST and CIFAR-10 demonstrate that DGKIP retains the efficiency of KIP while offering broader applicability and robust performance.
2502.12608
Unveiling Mode Connectivity in Graph Neural Networks
cs.LG cs.AI
A fundamental challenge in understanding graph neural networks (GNNs) lies in characterizing their optimization dynamics and loss landscape geometry, critical for improving interpretability and robustness. While mode connectivity, a lens for analyzing geometric properties of loss landscapes has proven insightful for other deep learning architectures, its implications for GNNs remain unexplored. This work presents the first investigation of mode connectivity in GNNs. We uncover that GNNs exhibit distinct non-linear mode connectivity, diverging from patterns observed in fully-connected networks or CNNs. Crucially, we demonstrate that graph structure, rather than model architecture, dominates this behavior, with graph properties like homophily correlating with mode connectivity patterns. We further establish a link between mode connectivity and generalization, proposing a generalization bound based on loss barriers and revealing its utility as a diagnostic tool. Our findings further bridge theoretical insights with practical implications: they rationalize domain alignment strategies in graph learning and provide a foundation for refining GNN training paradigms.
2502.12611
Who Writes What: Unveiling the Impact of Author Roles on AI-generated Text Detection
cs.CL
The rise of Large Language Models (LLMs) necessitates accurate AI-generated text detection. However, current approaches largely overlook the influence of author characteristics. We investigate how sociolinguistic attributes-gender, CEFR proficiency, academic field, and language environment-impact state-of-the-art AI text detectors. Using the ICNALE corpus of human-authored texts and parallel AI-generated texts from diverse LLMs, we conduct a rigorous evaluation employing multi-factor ANOVA and weighted least squares (WLS). Our results reveal significant biases: CEFR proficiency and language environment consistently affected detector accuracy, while gender and academic field showed detector-dependent effects. These findings highlight the crucial need for socially aware AI text detection to avoid unfairly penalizing specific demographic groups. We offer novel empirical evidence, a robust statistical framework, and actionable insights for developing more equitable and reliable detection systems in real-world, out-of-domain contexts. This work paves the way for future research on bias mitigation, inclusive evaluation benchmarks, and socially responsible LLM detectors.
2502.12614
Label Drop for Multi-Aspect Relation Modeling in Universal Information Extraction
cs.CL cs.AI
Universal Information Extraction (UIE) has garnered significant attention due to its ability to address model explosion problems effectively. Extractive UIE can achieve strong performance using a relatively small model, making it widely adopted. Extractive UIEs generally rely on task instructions for different tasks, including single-target instructions and multiple-target instructions. Single-target instruction UIE enables the extraction of only one type of relation at a time, limiting its ability to model correlations between relations and thus restricting its capability to extract complex relations. While multiple-target instruction UIE allows for the extraction of multiple relations simultaneously, the inclusion of irrelevant relations introduces decision complexity and impacts extraction accuracy. Therefore, for multi-relation extraction, we propose LDNet, which incorporates multi-aspect relation modeling and a label drop mechanism. By assigning different relations to different levels for understanding and decision-making, we reduce decision confusion. Additionally, the label drop mechanism effectively mitigates the impact of irrelevant relations. Experiments show that LDNet outperforms or achieves competitive performance with state-of-the-art systems on 9 tasks, 33 datasets, in both single-modal and multi-modal, few-shot and zero-shot settings.\footnote{https://github.com/Lu-Yang666/LDNet}
2502.12616
Improving Chain-of-Thought Reasoning via Quasi-Symbolic Abstractions
cs.CL
Chain-of-Though (CoT) represents a common strategy for reasoning in Large Language Models (LLMs) by decomposing complex tasks into intermediate inference steps. However, explanations generated via CoT are susceptible to content biases that negatively affect their robustness and faithfulness. To mitigate existing limitations, recent work has proposed using logical formalisms coupled with external symbolic solvers. However, fully symbolic approaches possess the bottleneck of requiring a complete translation from natural language to formal languages, a process that affects efficiency and flexibility. To achieve a trade-off, this paper investigates methods to disentangle content from logical reasoning without a complete formalisation. In particular, we present QuaSAR (for Quasi-Symbolic Abstract Reasoning), a variation of CoT that guides LLMs to operate at a higher level of abstraction via quasi-symbolic explanations. Our framework leverages the capability of LLMs to formalise only relevant variables and predicates, enabling the coexistence of symbolic elements with natural language. We show the impact of QuaSAR for in-context learning and for constructing demonstrations to improve the reasoning capabilities of smaller models. Our experiments show that quasi-symbolic abstractions can improve CoT-based methods by up to 8% accuracy, enhancing robustness and consistency on challenging adversarial variations on both natural language (i.e. MMLU-Redux) and symbolic reasoning tasks (i.e., GSM-Symbolic).
2502.12617
A Graph-Enhanced Deep-Reinforcement Learning Framework for the Aircraft Landing Problem
cs.LG cs.AI cs.SY eess.SY
The Aircraft Landing Problem (ALP) is one of the challenging problems in aircraft transportation and management. The challenge is to schedule the arriving aircraft in a sequence so that the cost and delays are optimized. There are various solution approaches to solving this problem, most of which are based on operations research algorithms and meta-heuristics. Although traditional methods perform better on one or the other factors, there remains a problem of solving real-time rescheduling and computational scalability altogether. This paper presents a novel deep reinforcement learning (DRL) framework that combines graph neural networks with actor-critic architectures to address the ALP. This paper introduces three key contributions: A graph-based state representation that efficiently captures temporal and spatial relationships between aircraft, a specialized actor-critic architecture designed to handle multiple competing objectives in landing scheduling, and a runway balance strategy that ensures efficient resource utilization while maintaining safety constraints. The results show that the trained algorithm can be tested on different problem sets and the results are competitive to operation research algorithms. The experimental results on standard benchmark data sets demonstrate a 99.95 reduction in computational time compared to Mixed Integer Programming (MIP) and 38 higher runway throughput over First Come First Serve (FCFS) approaches. Therefore, the proposed solution is competitive to traditional approaches and achieves substantial advancements. Notably, it does not require retraining, making it particularly suitable for industrial deployment. The frameworks capability to generate solutions within 1 second enables real-time rescheduling, addressing critical requirements of air traffic management.
2502.12618
Uncertainty-Aware Graph Structure Learning
cs.LG
Graph Neural Networks (GNNs) have become a prominent approach for learning from graph-structured data. However, their effectiveness can be significantly compromised when the graph structure is suboptimal. To address this issue, Graph Structure Learning (GSL) has emerged as a promising technique that refines node connections adaptively. Nevertheless, we identify two key limitations in existing GSL methods: 1) Most methods primarily focus on node similarity to construct relationships, while overlooking the quality of node information. Blindly connecting low-quality nodes and aggregating their ambiguous information can degrade the performance of other nodes. 2) The constructed graph structures are often constrained to be symmetric, which may limit the model's flexibility and effectiveness. To overcome these limitations, we propose an Uncertainty-aware Graph Structure Learning (UnGSL) strategy. UnGSL estimates the uncertainty of node information and utilizes it to adjust the strength of directional connections, where the influence of nodes with high uncertainty is adaptively reduced. Importantly, UnGSL serves as a plug-in module that can be seamlessly integrated into existing GSL methods with minimal additional computational cost. In our experiments, we implement UnGSL into six representative GSL methods, demonstrating consistent performance improvements.
2502.12623
DeepResonance: Enhancing Multimodal Music Understanding via Music-centric Multi-way Instruction Tuning
cs.SD cs.AI cs.CL cs.MM eess.AS
Recent advancements in music large language models (LLMs) have significantly improved music understanding tasks, which involve the model's ability to analyze and interpret various musical elements. These improvements primarily focused on integrating both music and text inputs. However, the potential of incorporating additional modalities such as images, videos and textual music features to enhance music understanding remains unexplored. To bridge this gap, we propose DeepResonance, a multimodal music understanding LLM fine-tuned via multi-way instruction tuning with multi-way aligned music, text, image, and video data. To this end, we construct Music4way-MI2T, Music4way-MV2T, and Music4way-Any2T, three 4-way training and evaluation datasets designed to enable DeepResonance to integrate both visual and textual music feature content. We also introduce multi-sampled ImageBind embeddings and a pre-alignment Transformer to enhance modality fusion prior to input into text LLMs, tailoring DeepResonance for multi-way instruction tuning. Our model achieves state-of-the-art performances across six music understanding tasks, highlighting the benefits of the auxiliary modalities and the structural superiority of DeepResonance. We plan to open-source the models and the newly constructed datasets.
2502.12624
Implicit Repair with Reinforcement Learning in Emergent Communication
cs.LG cs.MA
Conversational repair is a mechanism used to detect and resolve miscommunication and misinformation problems when two or more agents interact. One particular and underexplored form of repair in emergent communication is the implicit repair mechanism, where the interlocutor purposely conveys the desired information in such a way as to prevent misinformation from any other interlocutor. This work explores how redundancy can modify the emergent communication protocol to continue conveying the necessary information to complete the underlying task, even with additional external environmental pressures such as noise. We focus on extending the signaling game, called the Lewis Game, by adding noise in the communication channel and inputs received by the agents. Our analysis shows that agents add redundancy to the transmitted messages as an outcome to prevent the negative impact of noise on the task success. Additionally, we observe that the emerging communication protocol's generalization capabilities remain equivalent to architectures employed in simpler games that are entirely deterministic. Additionally, our method is the only one suitable for producing robust communication protocols that can handle cases with and without noise while maintaining increased generalization performance levels.
2502.12627
DAMamba: Vision State Space Model with Dynamic Adaptive Scan
cs.CV
State space models (SSMs) have recently garnered significant attention in computer vision. However, due to the unique characteristics of image data, adapting SSMs from natural language processing to computer vision has not outperformed the state-of-the-art convolutional neural networks (CNNs) and Vision Transformers (ViTs). Existing vision SSMs primarily leverage manually designed scans to flatten image patches into sequences locally or globally. This approach disrupts the original semantic spatial adjacency of the image and lacks flexibility, making it difficult to capture complex image structures. To address this limitation, we propose Dynamic Adaptive Scan (DAS), a data-driven method that adaptively allocates scanning orders and regions. This enables more flexible modeling capabilities while maintaining linear computational complexity and global modeling capacity. Based on DAS, we further propose the vision backbone DAMamba, which significantly outperforms current state-of-the-art vision Mamba models in vision tasks such as image classification, object detection, instance segmentation, and semantic segmentation. Notably, it surpasses some of the latest state-of-the-art CNNs and ViTs. Code will be available at https://github.com/ltzovo/DAMamba.
2502.12629
Rate Maximization for Downlink Pinching-Antenna Systems
cs.IT eess.SP math.IT
In this letter, we consider a new type of flexible-antenna system, termed pinching-antenna, where multiple low-cost pinching antennas, realized by activating small dielectric particles on a dielectric waveguide, are jointly used to serve a single-antenna user. Our goal is to maximize the downlink transmission rate by optimizing the locations of the pinching antennas. However, these locations affect both the path losses and the phase shifts of the user's effective channel gain, making the problem challenging to solve. To address this challenge and solve the problem in a low complexity manner, a relaxed optimization problem is developed that minimizes the impact of path loss while ensuring that the received signals at the user are constructive. This approach leads to a two-stage algorithm: in the first stage, the locations of the pinching antennas are optimized to minimize the large-scale path loss; in the second stage, the antenna locations are refined to maximize the received signal strength. Simulation results show that pinching-antenna systems significantly outperform conventional fixed-location antenna systems, and the proposed algorithm achieves nearly the same performance as the highly complex exhaustive search-based benchmark.
2502.12630
Automating Prompt Leakage Attacks on Large Language Models Using Agentic Approach
cs.CR cs.AI
This paper presents a novel approach to evaluating the security of large language models (LLMs) against prompt leakage-the exposure of system-level prompts or proprietary configurations. We define prompt leakage as a critical threat to secure LLM deployment and introduce a framework for testing the robustness of LLMs using agentic teams. Leveraging AG2 (formerly AutoGen), we implement a multi-agent system where cooperative agents are tasked with probing and exploiting the target LLM to elicit its prompt. Guided by traditional definitions of security in cryptography, we further define a prompt leakage-safe system as one in which an attacker cannot distinguish between two agents: one initialized with an original prompt and the other with a prompt stripped of all sensitive information. In a safe system, the agents' outputs will be indistinguishable to the attacker, ensuring that sensitive information remains secure. This cryptographically inspired framework provides a rigorous standard for evaluating and designing secure LLMs. This work establishes a systematic methodology for adversarial testing of prompt leakage, bridging the gap between automated threat modeling and practical LLM security. You can find the implementation of our prompt leakage probing on GitHub.
2502.12631
Score-Based Diffusion Policy Compatible with Reinforcement Learning via Optimal Transport
cs.LG cs.AI
Diffusion policies have shown promise in learning complex behaviors from demonstrations, particularly for tasks requiring precise control and long-term planning. However, they face challenges in robustness when encountering distribution shifts. This paper explores improving diffusion-based imitation learning models through online interactions with the environment. We propose OTPR (Optimal Transport-guided score-based diffusion Policy for Reinforcement learning fine-tuning), a novel method that integrates diffusion policies with RL using optimal transport theory. OTPR leverages the Q-function as a transport cost and views the policy as an optimal transport map, enabling efficient and stable fine-tuning. Moreover, we introduce masked optimal transport to guide state-action matching using expert keypoints and a compatibility-based resampling strategy to enhance training stability. Experiments on three simulation tasks demonstrate OTPR's superior performance and robustness compared to existing methods, especially in complex and sparse-reward environments. In sum, OTPR provides an effective framework for combining IL and RL, achieving versatile and reliable policy learning. The code will be released at https://github.com/Sunmmyy/OTPR.git.
2502.12632
MALT Diffusion: Memory-Augmented Latent Transformers for Any-Length Video Generation
cs.CV cs.LG
Diffusion models are successful for synthesizing high-quality videos but are limited to generating short clips (e.g., 2-10 seconds). Synthesizing sustained footage (e.g. over minutes) still remains an open research question. In this paper, we propose MALT Diffusion (using Memory-Augmented Latent Transformers), a new diffusion model specialized for long video generation. MALT Diffusion (or just MALT) handles long videos by subdividing them into short segments and doing segment-level autoregressive generation. To achieve this, we first propose recurrent attention layers that encode multiple segments into a compact memory latent vector; by maintaining this memory vector over time, MALT is able to condition on it and continuously generate new footage based on a long temporal context. We also present several training techniques that enable the model to generate frames over a long horizon with consistent quality and minimal degradation. We validate the effectiveness of MALT through experiments on long video benchmarks. We first perform extensive analysis of MALT in long-contextual understanding capability and stability using popular long video benchmarks. For example, MALT achieves an FVD score of 220.4 on 128-frame video generation on UCF-101, outperforming the previous state-of-the-art of 648.4. Finally, we explore MALT's capabilities in a text-to-video generation setting and show that it can produce long videos compared with recent techniques for long text-to-video generation.
2502.12633
One Size doesn't Fit All: A Personalized Conversational Tutoring Agent for Mathematics Instruction
cs.CL cs.AI
Large language models (LLMs) have been increasingly employed in various intelligent educational systems, simulating human tutors to facilitate effective human-machine interaction. However, previous studies often overlook the significance of recognizing and adapting to individual learner characteristics. Such adaptation is crucial for enhancing student engagement and learning efficiency, particularly in mathematics instruction, where diverse learning styles require personalized strategies to promote comprehension and enthusiasm. In this paper, we propose a \textbf{P}erson\textbf{A}lized \textbf{C}onversational tutoring ag\textbf{E}nt (PACE) for mathematics instruction. PACE simulates students' learning styles based on the Felder and Silverman learning style model, aligning with each student's persona. In this way, our PACE can effectively assess the personality of students, allowing to develop individualized teaching strategies that resonate with their unique learning styles. To further enhance students' comprehension, PACE employs the Socratic teaching method to provide instant feedback and encourage deep thinking. By constructing personalized teaching data and training models, PACE demonstrates the ability to identify and adapt to the unique needs of each student, significantly improving the overall learning experience and outcomes. Moreover, we establish multi-aspect evaluation criteria and conduct extensive analysis to assess the performance of personalized teaching. Experimental results demonstrate the superiority of our model in personalizing the educational experience and motivating students compared to existing methods.
2502.12634
Introducing Context Information in Lifelong Sequential Modeling using Temporal Convolutional Networks
cs.IR
The importance of lifelong sequential modeling (LSM) is growing in the realm of social media recommendation systems. A key component in this process is the attention module, which derives interest representations with respect to candidate items from the sequence. Typically, attention modules function in a point-wise fashion, concentrating only on the relevance of individual items in the sequence to the candidate item. However, the context information in the neighboring items that is useful for more accurately evaluating the significance of each item has not been taken into account. In this study, we introduce a novel network which employs the Temporal Convolutional Network (TCN) to generate context-aware representations for each item throughout the lifelong sequence. These improved representations are then utilized in the attention module to produce context-aware interest representations. Expanding on this TCN framework, we present a enhancement module which includes multiple TCN layers and their respective attention modules to capture interest representations across different context scopes. Additionally, we also incorporate a lightweight sub-network to create convolution filters based on users' basic profile features. These personalized filters are then applied in the TCN layers instead of the original global filters to produce more user-specific representations. We performed experiments on both a public dataset and a proprietary dataset. The findings indicate that the proposed network surpasses existing methods in terms of prediction accuracy and online performance metrics.
2502.12635
Corrupted but Not Broken: Rethinking the Impact of Corrupted Data in Visual Instruction Tuning
cs.CV
Visual Instruction Tuning (VIT) enhances Multimodal Large Language Models (MLLMs) but it is hindered by corrupted datasets containing hallucinated content, incorrect responses, and poor OCR quality. While prior works focus on dataset refinement through high-quality data collection or rule-based filtering, they are costly or limited to specific types of corruption. To deeply understand how corrupted data affects MLLMs, in this paper, we systematically investigate this issue and find that while corrupted data degrades the performance of MLLMs, its effects are largely superficial in that the performance of MLLMs can be largely restored by either disabling a small subset of parameters or post-training with a small amount of clean data. Additionally, corrupted MLLMs exhibit improved ability to distinguish clean samples from corrupted ones, enabling the dataset cleaning without external help. Based on those insights, we propose a corruption-robust training paradigm combining self-validation and post-training, which significantly outperforms existing corruption mitigation strategies.
2502.12638
NExT-Mol: 3D Diffusion Meets 1D Language Modeling for 3D Molecule Generation
q-bio.QM cs.LG q-bio.BM
3D molecule generation is crucial for drug discovery and material design. While prior efforts focus on 3D diffusion models for their benefits in modeling continuous 3D conformers, they overlook the advantages of 1D SELFIES-based Language Models (LMs), which can generate 100% valid molecules and leverage the billion-scale 1D molecule datasets. To combine these advantages for 3D molecule generation, we propose a foundation model -- NExT-Mol: 3D Diffusion Meets 1D Language Modeling for 3D Molecule Generation. NExT-Mol uses an extensively pretrained molecule LM for 1D molecule generation, and subsequently predicts the generated molecule's 3D conformers with a 3D diffusion model. We enhance NExT-Mol's performance by scaling up the LM's model size, refining the diffusion neural architecture, and applying 1D to 3D transfer learning. Notably, our 1D molecule LM significantly outperforms baselines in distributional similarity while ensuring validity, and our 3D diffusion model achieves leading performances in conformer prediction. Given these improvements in 1D and 3D modeling, NExT-Mol achieves a 26% relative improvement in 3D FCD for de novo 3D generation on GEOM-DRUGS, and a 13% average relative gain for conditional 3D generation on QM9-2014. Our codes and pretrained checkpoints are available at https://github.com/acharkq/NExT-Mol.
2502.12640
RecDreamer: Consistent Text-to-3D Generation via Uniform Score Distillation
cs.CV
Current text-to-3D generation methods based on score distillation often suffer from geometric inconsistencies, leading to repeated patterns across different poses of 3D assets. This issue, known as the Multi-Face Janus problem, arises because existing methods struggle to maintain consistency across varying poses and are biased toward a canonical pose. While recent work has improved pose control and approximation, these efforts are still limited by this inherent bias, which skews the guidance during generation. To address this, we propose a solution called RecDreamer, which reshapes the underlying data distribution to achieve a more consistent pose representation. The core idea behind our method is to rectify the prior distribution, ensuring that pose variation is uniformly distributed rather than biased toward a canonical form. By modifying the prescribed distribution through an auxiliary function, we can reconstruct the density of the distribution to ensure compliance with specific marginal constraints. In particular, we ensure that the marginal distribution of poses follows a uniform distribution, thereby eliminating the biases introduced by the prior knowledge. We incorporate this rectified data distribution into existing score distillation algorithms, a process we refer to as uniform score distillation. To efficiently compute the posterior distribution required for the auxiliary function, RecDreamer introduces a training-free classifier that estimates pose categories in a plug-and-play manner. Additionally, we utilize various approximation techniques for noisy states, significantly improving system performance. Our experimental results demonstrate that RecDreamer effectively mitigates the Multi-Face Janus problem, leading to more consistent 3D asset generation across different poses.
2502.12654
Free Energy and Network Structure: Breaking Scale-Free Behaviour Through Information Processing Constraints
cs.SI physics.soc-ph
In this paper we show how The Free Energy Principle (FEP) can provide an explanation for why real-world networks deviate from scale-free behaviour, and how these characteristic deviations can emerge from constraints on information processing. We propose a minimal FEP model for node behaviour reveals three distinct regimes: when detection noise dominates, agents seek better information, reducing isolated agents compared to expectations from classical preferential attachment. In the optimal detection regime, super-linear growth emerges from compounded improvements in detection, belief, and action, which produce a preferred cluster scale. Finally, saturation effects occur as limits on the agent's information processing capabilities prevent indefinite cluster growth. These regimes produce the knee-shaped degree distributions observed in real networks, explaining them as signatures of agents with optimal information processing under constraints. We show that agents evolving under FEP principles provides a mechanism for preferential attachment, connecting agent psychology with the macroscopic network features that underpin the structure of real-world networks.
2502.12655
LiMo-Calib: On-Site Fast LiDAR-Motor Calibration for Quadruped Robot-Based Panoramic 3D Sensing System
cs.RO
Conventional single LiDAR systems are inherently constrained by their limited field of view (FoV), leading to blind spots and incomplete environmental awareness, particularly on robotic platforms with strict payload limitations. Integrating a motorized LiDAR offers a practical solution by significantly expanding the sensor's FoV and enabling adaptive panoramic 3D sensing. However, the high-frequency vibrations of the quadruped robot introduce calibration challenges, causing variations in the LiDAR-motor transformation that degrade sensing accuracy. Existing calibration methods that use artificial targets or dense feature extraction lack feasibility for on-site applications and real-time implementation. To overcome these limitations, we propose LiMo-Calib, an efficient on-site calibration method that eliminates the need for external targets by leveraging geometric features directly from raw LiDAR scans. LiMo-Calib optimizes feature selection based on normal distribution to accelerate convergence while maintaining accuracy and incorporates a reweighting mechanism that evaluates local plane fitting quality to enhance robustness. We integrate and validate the proposed method on a motorized LiDAR system mounted on a quadruped robot, demonstrating significant improvements in calibration efficiency and 3D sensing accuracy, making LiMo-Calib well-suited for real-world robotic applications. The demo video is available at: https://youtu.be/FMINa-sap7g
2502.12658
R.R.: Unveiling LLM Training Privacy through Recollection and Ranking
cs.CL
Large Language Models (LLMs) pose significant privacy risks, potentially leaking training data due to implicit memorization. Existing privacy attacks primarily focus on membership inference attacks (MIAs) or data extraction attacks, but reconstructing specific personally identifiable information (PII) in LLM's training data remains challenging. In this paper, we propose R.R. (Recollect and Rank), a novel two-step privacy stealing attack that enables attackers to reconstruct PII entities from scrubbed training data where the PII entities have been masked. In the first stage, we introduce a prompt paradigm named recollection, which instructs the LLM to repeat a masked text but fill in masks. Then we can use PII identifiers to extract recollected PII candidates. In the second stage, we design a new criterion to score each PII candidate and rank them. Motivated by membership inference, we leverage the reference model as a calibration to our criterion. Experiments across three popular PII datasets demonstrate that the R.R. achieves better PII identical performance compared to baselines. These results highlight the vulnerability of LLMs to PII leakage even when training data has been scrubbed. We release the replicate package of R.R. at a link.
2502.12659
The Hidden Risks of Large Reasoning Models: A Safety Assessment of R1
cs.CY cs.AI
The rapid development of large reasoning models, such as OpenAI-o3 and DeepSeek-R1, has led to significant improvements in complex reasoning over non-reasoning large language models~(LLMs). However, their enhanced capabilities, combined with the open-source access of models like DeepSeek-R1, raise serious safety concerns, particularly regarding their potential for misuse. In this work, we present a comprehensive safety assessment of these reasoning models, leveraging established safety benchmarks to evaluate their compliance with safety regulations. Furthermore, we investigate their susceptibility to adversarial attacks, such as jailbreaking and prompt injection, to assess their robustness in real-world applications. Through our multi-faceted analysis, we uncover four key findings: (1) There is a significant safety gap between the open-source R1 models and the o3-mini model, on both safety benchmark and attack, suggesting more safety effort on R1 is needed. (2) The distilled reasoning model shows poorer safety performance compared to its safety-aligned base models. (3) The stronger the model's reasoning ability, the greater the potential harm it may cause when answering unsafe questions. (4) The thinking process in R1 models pose greater safety concerns than their final answers. Our study provides insights into the security implications of reasoning models and highlights the need for further advancements in R1 models' safety to close the gap.
2502.12663
Demystifying Multilingual Chain-of-Thought in Process Reward Modeling
cs.CL
Large language models (LLMs) are designed to perform a wide range of tasks. To improve their ability to solve complex problems requiring multi-step reasoning, recent research leverages process reward modeling to provide fine-grained feedback at each step of the reasoning process for reinforcement learning (RL), but it predominantly focuses on English. In this paper, we tackle the critical challenge of extending process reward models (PRMs) to multilingual settings. To achieve this, we train multilingual PRMs on a dataset spanning seven languages, which is translated from English. Through comprehensive evaluations on two widely used reasoning benchmarks across 11 languages, we demonstrate that multilingual PRMs not only improve average accuracy but also reduce early-stage reasoning errors. Furthermore, our results highlight the sensitivity of multilingual PRMs to both the number of training languages and the volume of English data, while also uncovering the benefits arising from more candidate responses and trainable parameters. This work opens promising avenues for robust multilingual applications in complex, multi-step reasoning tasks. In addition, we release the code to foster research along this line.
2502.12665
A$^2$ATS: Retrieval-Based KV Cache Reduction via Windowed Rotary Position Embedding and Query-Aware Vector Quantization
cs.CL
Long context large language models (LLMs) pose significant challenges for efficient serving due to the large memory footprint and high access overhead of KV cache. Retrieval-based KV cache reduction methods can mitigate these challenges, typically by offloading the complete KV cache to CPU and retrieving necessary tokens on demand during inference. However, these methods still suffer from unsatisfactory accuracy degradation and extra retrieval overhead. To address these limitations, this paper proposes A$^2$ATS, a novel retrieval-based KV cache reduction method. A$^2$ATS aims to obtain an accurate approximation of attention scores by applying the vector quantization technique to key states, thereby enabling efficient and precise retrieval of the top-K tokens. First, we propose Windowed Rotary Position Embedding, which decouples the positional dependency from query and key states after position embedding. Then, we propose query-aware vector quantization that optimizes the objective of attention score approximation directly. Finally, we design the heterogeneous inference architecture for KV cache offloading, enabling long context serving with larger batch sizes. Experimental results demonstrate that A$^2$ATS can achieve a lower performance degradation with similar or lower overhead compared to existing methods, thereby increasing long context serving throughput by up to $2.7 \times$.
2502.12668
Evaluation of Best-of-N Sampling Strategies for Language Model Alignment
cs.CL
Best-of-N (BoN) sampling with a reward model has been shown to be an effective strategy for aligning Large Language Models (LLMs) with human preferences at the time of decoding. BoN sampling is susceptible to a problem known as reward hacking. Since the reward model is an imperfect proxy for the true objective, an excessive focus on optimizing its value can lead to a compromise of its performance on the true objective. Previous work proposes Regularized BoN sampling (RBoN), a BoN sampling with regularization to the objective, and shows that it outperforms BoN sampling so that it mitigates reward hacking and empirically (Jinnai et al., 2024). However, Jinnai et al. (2024) introduce RBoN based on a heuristic and they lack the analysis of why such regularization strategy improves the performance of BoN sampling. The aim of this study is to analyze the effect of BoN sampling on regularization strategies. Using the regularization strategies corresponds to robust optimization, which maximizes the worst case over a set of possible perturbations in the proxy reward. Although the theoretical guarantees are not directly applicable to RBoN, RBoN corresponds to a practical implementation. This paper proposes an extension of the RBoN framework, called Stochastic RBoN sampling (SRBoN), which is a theoretically guaranteed approach to worst-case RBoN in proxy reward. We then perform an empirical evaluation using the AlpacaFarm and Anthropic's hh-rlhf datasets to evaluate which factors of the regularization strategies contribute to the improvement of the true proxy reward. In addition, we also propose another simple RBoN method, the Sentence Length Regularized BoN, which has a better performance in the experiment as compared to the previous methods.
2502.12669
Perovskite-LLM: Knowledge-Enhanced Large Language Models for Perovskite Solar Cell Research
cs.AI
The rapid advancement of perovskite solar cells (PSCs) has led to an exponential growth in research publications, creating an urgent need for efficient knowledge management and reasoning systems in this domain. We present a comprehensive knowledge-enhanced system for PSCs that integrates three key components. First, we develop Perovskite-KG, a domain-specific knowledge graph constructed from 1,517 research papers, containing 23,789 entities and 22,272 relationships. Second, we create two complementary datasets: Perovskite-Chat, comprising 55,101 high-quality question-answer pairs generated through a novel multi-agent framework, and Perovskite-Reasoning, containing 2,217 carefully curated materials science problems. Third, we introduce two specialized large language models: Perovskite-Chat-LLM for domain-specific knowledge assistance and Perovskite-Reasoning-LLM for scientific reasoning tasks. Experimental results demonstrate that our system significantly outperforms existing models in both domain-specific knowledge retrieval and scientific reasoning tasks, providing researchers with effective tools for literature review, experimental design, and complex problem-solving in PSC research.
2502.12671
Baichuan-M1: Pushing the Medical Capability of Large Language Models
cs.CL
The current generation of large language models (LLMs) is typically designed for broad, general-purpose applications, while domain-specific LLMs, especially in vertical fields like medicine, remain relatively scarce. In particular, the development of highly efficient and practical LLMs for the medical domain is challenging due to the complexity of medical knowledge and the limited availability of high-quality data. To bridge this gap, we introduce Baichuan-M1, a series of large language models specifically optimized for medical applications. Unlike traditional approaches that simply continue pretraining on existing models or apply post-training to a general base model, Baichuan-M1 is trained from scratch with a dedicated focus on enhancing medical capabilities. Our model is trained on 20 trillion tokens and incorporates a range of effective training methods that strike a balance between general capabilities and medical expertise. As a result, Baichuan-M1 not only performs strongly across general domains such as mathematics and coding but also excels in specialized medical fields. We have open-sourced Baichuan-M1-14B, a mini version of our model, which can be accessed through the following links.
2502.12672
Speech-FT: A Fine-tuning Strategy for Enhancing Speech Representation Models Without Compromising Generalization Ability
cs.CL cs.AI
Speech representation models are highly effective at extracting general features for various tasks. While fine-tuning can enhance these representations for specific applications, it often compromises their generalization ability. To address this challenge, we propose Speech-FT, a fine-tuning strategy for speech representation models that leverages model merging to preserve generalization ability while still benefiting from fine-tuning. Speech-FT is effective across different fine-tuning scenarios and is compatible with various types of speech representation models, providing a versatile solution. Speech-FT offers an efficient and practical approach to further improving general speech representations after pre-training.
2502.12673
ROI-NeRFs: Hi-Fi Visualization of Objects of Interest within a Scene by NeRFs Composition
cs.CV cs.GR
Efficient and accurate 3D reconstruction is essential for applications in cultural heritage. This study addresses the challenge of visualizing objects within large-scale scenes at a high level of detail (LOD) using Neural Radiance Fields (NeRFs). The aim is to improve the visual fidelity of chosen objects while maintaining the efficiency of the computations by focusing on details only for relevant content. The proposed ROI-NeRFs framework divides the scene into a Scene NeRF, which represents the overall scene at moderate detail, and multiple ROI NeRFs that focus on user-defined objects of interest. An object-focused camera selection module automatically groups relevant cameras for each NeRF training during the decomposition phase. In the composition phase, a Ray-level Compositional Rendering technique combines information from the Scene NeRF and ROI NeRFs, allowing simultaneous multi-object rendering composition. Quantitative and qualitative experiments conducted on two real-world datasets, including one on a complex eighteen's century cultural heritage room, demonstrate superior performance compared to baseline methods, improving LOD for object regions, minimizing artifacts, and without significantly increasing inference time.