paperlist.ai

AI 顶会论文

AI 顶会论文库

检索 30+ 个顶会的论文，包括 NeurIPS、ICLR、ICML、CVPR、ACL、AAAI 等。输入关键词筛选，打开任意一篇查看摘要、作者和链接。

arXiv

Semantics-Aware Hierarchical Consensus Learning for Remote Sensing Image Classification

Giulio Weikmann, Gianmarco Perantoni, Lorenzo Bruzzone

Deep learning has become increasingly important in remote sensing image classification due to its ability to extract semantic information from complex data. Classification tasks often include predefined label hierarchies that represent the semantic relationships among classes. However, these hierarchies are frequently overlooked, and most approaches focus only on fine- grained classification schemes. In this paper, we present a novel Semantics-Aware Hierarchical Consensus (SAHC) approach that integrates hierarchical-level-specific classification heads within a deep network architecture and combines their output through cross-level probability projectors. Direct and projected predictions are fused into a geometric consensus distribution, which is used for self-consistent training and optional hierarchy-aware inference. This mechanism acts as a geometric ensemble that leverages the inherent structure of the hierarchical classification task. The projectors are initialized from the user-defined taxonomy (i.e. the hierarchical label structure), and can be adaptively refined during optimization. The proposed SAHC method is evaluated on two benchmark datasets with different degrees of hierarchical complexity on different tasks, considering varying spectral and spatial resolutions. Experimental results show both the effectiveness of the proposed approach in guiding network learning and the robustness of the hierarchical consensus for remote sensing image classification tasks. The source code will be released at https://github.com/rslab-unitrento/sahc.

arXiv

Forge4D: Feed-Forward 4D Human Reconstruction and Interpolation from Uncalibrated Sparse-view Videos

Yingdong Hu, Yisheng He, Jinnan Chen, Weihao Yuan, Kejie Qiu, Zehong Lin, Siyu Zhu, Zilong Dong, Steven Hoi, Jun Zhang

Instant reconstruction of dynamic 3D humans from uncalibrated sparse-view videos is critical for numerous downstream applications. Existing methods, however, are either limited by the slow reconstruction speeds or incapable of generating novel-time representations. To address these challenges, we propose Forge4D, a feed-forward 4D human reconstruction and interpolation model that efficiently reconstructs temporally aligned representations from uncalibrated sparse-view videos, enabling both novel view and novel time synthesis. Our model simplifies the 4D reconstruction and interpolation problem as a joint task of streaming 3D Gaussian reconstruction and dense motion prediction. For the task of streaming 3D Gaussian reconstruction, we first reconstruct static 3D Gaussians from uncalibrated sparse-view images and then introduce learnable state tokens to enforce temporal consistency in a memory-friendly manner by interactively updating shared information across different timestamps. For novel time synthesis, we design a novel motion prediction module to predict dense motions for each 3D Gaussian between two adjacent frames, coupled with an occlusion-aware Gaussian fusion process to interpolate 3D Gaussians at arbitrary timestamps. To overcome the lack of the ground truth for dense motion supervision, we formulate dense motion prediction as a dense point matching task and introduce a self-supervised retargeting loss to optimize this module. An additional occlusion-aware optical flow loss is introduced to ensure motion consistency with plausible human movement, providing stronger regularization. Extensive experiments demonstrate the effectiveness of our model on both in-domain and out-of-domain datasets. Project page and code at: https://zhenliuzju.github.io/huyingdong/Forge4D.

arXiv

Towards Fine-Grained Text-to-3D Quality Assessment: A Benchmark and A Two-Stage Rank-Learning Metric

Bingyang Cui, Yujie Zhang, Qi Yang, Zhu Li, Yiling Xu

Recent advances in Text-to-3D (T23D) generative models have enabled the synthesis of diverse, high-fidelity 3D assets from textual prompts. However, existing challenges restrict the development of reliable T23D quality assessment (T23DQA). First, existing benchmarks are outdated, fragmented, and coarse-grained, making fine-grained metric training infeasible. Moreover, current objective metrics exhibit inherent design limitations, resulting in non-representative feature extraction and diminished metric robustness. To address these limitations, we introduce T23D-CompBench, a comprehensive benchmark for compositional T23D generation. We define five components with twelve sub-components for compositional prompts, which are used to generate 3,600 textured meshes from ten state-of-the-art generative models. A large-scale subjective experiment is conducted to collect 129,600 reliable human ratings across different perspectives. Based on T23D-CompBench, we further propose Rank2Score, an effective evaluator with two-stage training for T23DQA. Rank2Score enhances pairwise training via supervised contrastive regression and curriculum learning in the first stage, and subsequently refines predictions using mean opinion scores to achieve closer alignment with human judgments in the second stage. Extensive experiments and downstream applications demonstrate that Rank2Score consistently outperforms existing metrics across multiple dimensions and can additionally serve as a reward function to optimize generative models. The project is available at https://cbysjtu.github.io/Rank2Score/.

arXiv

Learning to Detect Unseen Jailbreak Attacks in Large Vision-Language Models

Shuang Liang, Zhihao Xu, Jiaqi Weng, Jialing Tao, Hui Xue, Xiting Wang

Despite extensive alignment efforts, Large Vision-Language Models (LVLMs) remain vulnerable to jailbreak attacks. To mitigate these risks, existing detection methods are essential, yet they face two major challenges: generalization and accuracy. While learning-based methods trained on specific attacks fail to generalize to unseen attacks, learning-free methods based on hand-crafted heuristics suffer from limited accuracy and reduced efficiency. To address these limitations, we propose Learning to Detect (LoD), a learnable framework that eliminates the need for any attack data or hand-crafted heuristics. LoD operates by first extracting layer-wise safety representations directly from the model's internal activations using Multi-modal Safety Concept Activation Vectors classifiers, and then converting the high-dimensional representations into a one-dimensional anomaly score for detection via a Safety Pattern Auto-Encoder. Extensive experiments demonstrate that LoD consistently achieves state-of-the-art detection performance (AUROC) across diverse unseen jailbreak attacks on multiple LVLMs, while also significantly improving efficiency. Code is available at https://github.com/ShuangLiangX/Learning-to-Detect.

arXiv

3DGAA: Realistic and Robust 3D Gaussian-based Adversarial Attack for Autonomous Driving

Yixun Zhang, Lizhi Wang, Junjun Zhao, Wending Zhao, Feng Zhou, Yonghao Dang, Jianqin Yin

Camera-based perception in connected and autonomous vehicles remains exposed to physical adversarial attacks. Prior attacks often either optimize image-plane textures, weakening cross-view consistency, or rely on shape modifications that are difficult to fabricate and deploy. Both cases limit their value for security and safety assessment of camera-based perception. This paper introduces a fabrication-first framework that learns view-consistent, geometry-preserving adversarial wraps for vehicles. The method performs three-dimensional multi-view optimization with a 3D Gaussian splatting surrogate to enforce consistency under an expectation over viewpoint, illumination, and occlusion, while the final artifact is constrained to print-only textures that keep vehicle geometry unchanged. We evaluate the framework through matched CARLA simulations, controlled miniature-vehicle physical evidence, and tests on multiple modern detectors. The wraps produce substantial and stable reductions in detection confidence and average precision across views while maintaining perceptual realism. Ablation studies isolate the effects of physical filtering, physical augmentation, and shape-consistency regularization, while efficiency analyses report convergence time and memory profiles. Additional experiments report ASR, mIoU-based segmentation degradation, cross-detector transfer against physical attack baselines, perturbation-area fairness diagnostics, and robustness under common input preprocessing defenses. By coupling three-dimensional optimization with a geometry-preserving realization, the study supports a practical pathway to systematically stress-test camera perception in safety-critical autonomous driving systems and provides evidence that can inform defense design and evaluation protocols in information forensics and security.

arXiv

BRIDLE: Generalized Self-supervised Learning with Quantization

Hoang M. Nguyen, Satya N. Shukla, Qiang Zhang, Hanchao Yu, Sreya D. Roy, Dipesh Tamboli, Taipeng Tian, Lingjiong Zhu, Yuchen Liu

Self-supervised learning has been a powerful approach for learning meaningful representations from unlabeled data across various domains, reducing the reliance on large labeled datasets. Inspired by BERT's success in capturing deep bidirectional contexts in natural language processing, similar frameworks have been adapted to other modalities such as audio, with models like BEATs extending the bidirectional training paradigm to audio signals using vector quantization (VQ). However, these frameworks face challenges, notably their dependence on a single codebook for quantization, which may not capture the complex, multifaceted nature of signals. In addition, inefficiencies in codebook utilization lead to underutilized code vectors. To address these limitations, we introduce BRIDLE (Bidirectional Residual Quantization Interleaved Discrete Learning Encoder), a self-supervised encoder pretraining framework that incorporates residual quantization (RQ) into the bidirectional training process, and is generalized for pretraining with audio, image, and video. Using multiple hierarchical codebooks, RQ enables fine-grained discretization in the latent space, enhancing representation quality. BRIDLE involves an interleaved training procedure between the encoder and tokenizer. We evaluate BRIDLE on audio understanding tasks using classification benchmarks, achieving state-of-the-art results, and demonstrate competitive performance on image classification and video classification tasks, showing consistent improvements over traditional VQ methods in downstream performance.

arXiv

SEG-SAM: Semantic-Guided SAM for Unified Medical Image Segmentation

Shuangping Huang, Hao Liang, Qingfeng Wang, Chulong Zhong, Zijian Zhou, Miaojing Shi

Recently, developing unified medical image segmentation models gains increasing attention, especially with the advent of the Segment Anything Model (SAM). SAM has shown promising binary segmentation performance in natural domains, however, transferring it to the medical domain remains challenging, as medical images often possess substantial inter-category overlaps. To address this, we propose the SEmantic-Guided SAM (SEG-SAM), a unified medical segmentation model that incorporates semantic medical knowledge to enhance medical segmentation performance. First, to avoid the potential conflict between binary and semantic predictions, we introduce a semantic-aware decoder independent of SAM's original decoder, specialized for both semantic segmentation on the prompted object and classification on unprompted objects in images. To further enhance the model's semantic understanding, we solicit key characteristics of medical categories from large language models and incorporate them into SEG-SAM through a text-to-vision semantic module, adaptively transferring the language information into the visual segmentation task. In the end, we introduce the cross-mask spatial alignment strategy to encourage greater overlap between the predicted masks from SEG-SAM's two decoders, thereby benefiting both predictions. Extensive experiments demonstrate that SEG-SAM outperforms state-of-the-art SAM-based methods in unified binary medical segmentation and task-specific methods in semantic medical segmentation, showcasing promising results and potential for broader medical applications.

arXiv

A Comprehensive Study for Edge Detection: Ideal-Prior Guidance Learning, High-Precision Architecture, and Better Distinguishability Evaluation Benchmark

Hao Shu

Image edge detection (ED) requires specialized architectures, reliable supervision, and rigorous evaluation criteria to ensure accurate localization. In this work, we present a framework for high-precision ED that jointly addresses architectural design, data supervision, and evaluation consistency. We propose SDPED, a compact ED model built upon Cascaded Skipping Density Blocks (CSDB), motivated by a task-adaptive architectural transfer from image super-resolution. By re-engineering texture-oriented structures for ED, SDPED effectively differentiates textures from edges while preserving fine spatial precision. Extensive experiments on four benchmark datasets (BRIND, UDED, MDBD, and BIPED2) demonstrate consistent performance improvements, particularly in Average Precision (AP), with gains of up to 22.5% on MDBD and 11.8% on BIPED2. In addition, we introduce an ideal-prior guidance strategy that incorporates noiseless data into training by treating labels as noise-free samples, providing a practical means to mitigate the subjectivity and noise inherent in human annotations. To enable better distinguishability and resolution-independent evaluation, we further study the criterion for assessing ED models, which suggests a stricter 1-pixel error tolerance distance. Overall, this work offers a coherent solution for high-precision ED and provides insights applicable to precision-oriented modeling in low-level and soft-computing-based vision tasks. Codes can be found on https://github.com/Hao-B-Shu/SDPED.

arXiv

Advancements in Content-Based Image Retrieval: A Comprehensive Survey of Relevance Feedback Techniques

Hamed Qazanfari, Mohammad M. AlyanNezhadi, Zohreh Nozari Khoshdaregi

Content-based image retrieval (CBIR) systems have emerged as crucial tools in the field of computer vision, allowing for image search based on visual content rather than relying solely on metadata. This survey paper presents a comprehensive overview of CBIR, emphasizing its role in object detection and its potential to identify and retrieve visually similar images based on content features. Challenges faced by CBIR systems, including the semantic gap and scalability, are discussed, along with potential solutions. It elaborates on the semantic gap, which arises from the disparity between low-level features and high-level semantic concepts, and explores approaches to bridge this gap. One notable solution is the integration of relevance feedback (RF), empowering users to provide feedback on retrieved images and refine search results iteratively. The survey encompasses long-term and short-term learning approaches that leverage RF for enhanced CBIR accuracy and relevance. These methods focus on weight optimization and the utilization of active learning algorithms to select samples for training classifiers. Furthermore, the paper investigates machine learning techniques and the utilization of deep learning and convolutional neural networks to enhance CBIR performance. This survey paper plays a significant role in advancing the understanding of CBIR and RF techniques. It guides researchers and practitioners in comprehending existing methodologies, challenges, and potential solutions while fostering knowledge dissemination and identifying research gaps. By addressing future research directions, it sets the stage for advancements in CBIR that will enhance retrieval accuracy, usability, and effectiveness in various application domains.

arXiv

Compositional Generalization via Structural Identification in a Category-Theoretic Framework

Akihiro Maeda, Thomas Seiller, Yohei Oseki

Compositional generalization is usually evaluated through model accuracy. We instead ask which structural or lexical identifications make held-out COGS examples admissible from the structures observed in training. Sentences are represented as functors from syntactic addresses to lexical tokens, and selective collapses induce Kan extensions that propagate observed associations. Across 21 COGS generalization types, admissibility follows distinct identification profiles, while residual failures separate unsupported structural templates. These data-side diagnoses characterize what the training corpus licenses under specified identifications, without training a predictive model.

arXiv

Diff Mining: Logit Differences Reveal Finetuning Objectives

Greg Kocher, Robert West, Clément Dumas, Julian Minder

Finetuning has become the gold standard for refining existing behaviors and inducing new ones in language models, yet it often remains unclear exactly which behaviors emerge during this process. As models grow ever more capable, understanding finetuning better becomes increasingly important, particularly since unwanted behaviors may arise during finetuning. In this paper, we introduce Diff Mining, a simple yet effective framework for identifying what a finetuned model has learned by comparing its logits to those of its base model. Diff Mining effectively surfaces salient tokens that are amplified in the finetuned model, serving as a fingerprint of its training -- even on text unrelated to the finetuning domain. Unlike many existing model diffing methods which require model internals, Diff Mining only needs access to output logits and scales to large models. The framework consists of two modular stages: (i) extracting per-context logit differences between the finetuned and base models on a reference corpus, and (ii) aggregating the resulting signals to construct an interpretable token set representing the finetune. For aggregation, we explore both a simple Top-K frequency method and a Non-negative Matrix Factorization (NMF)-based approach for disentangling multiple finetuning objectives into distinct token clusters. Empirically, Diff Mining succeeds across diverse settings: on finetune domain detection, it significantly outperforms state-of-the-art model diffing methods both in identifying relevant tokens and in downstream performance when an interpretability agent is given access to the extracted token set; on models with injected biases, it identifies more than one third of the biases without targeted probing. Overall, our framework shows promise in developing auditing tools to detect finetuning objectives.

arXiv

Vowel Signs Are Not Letters: A Pre-tokenization Ceiling on Multilingual Tokenizer Fertility

Sajal Regmi, Siddhartha Pudasaini, Chetan Phakami Pun

Byte-level BPE tokenizers that use the HuggingFace ByteLevel pre-tokenizer inherit GPT-2's word regex, where a word is defined as \p{L}+, one or more Unicode letters. In abugida scripts, vowels are written as combining marks; this pattern therefore splits each word at every vowel sign. Since BPE merges only within a pre-token, those splits persist through training regardless of vocabulary size or corpus composition. We formalise this effect as a training-free lower bound on fertility. Across 26 languages from a parallel corpus, every one of the 17 abugidas is affected, ranging from 1.47x (Tibetan) to 9.02x (Thai), whereas Latin, Cyrillic, Hangul, and Han show exactly 1.00x. For 5 languages, matched tokenizer pairs that differ only in this character class fall within 2.2% of the predicted floor, scoring 4.78 versus 1.58 tokens per word on Nepali. When the Nepali share of the training corpus is swept from 5% to 95%, the broken tokenizer barely shifts at all (1.7%) while the fixed one shifts 33.9%, which separates a structural ceiling from a data shortage without needing to inspect any code. We train three 268M models that differ only in their tokenizer; the fixed variant achieves 4.43% lower held-out Nepali bits per byte at equal compute, and it still leads when given the same bytes with 1.59x the compute. A census of 3,479 HuggingFace repositories finds the letters-only word class present in 63.3% of the most-downloaded text-generation models, accounting for 72.5% of their downloads. GPT-4o's o200k pattern already uses a mark-aware word class, making the repair itself prior art. We quantify its value, show how to recognise its absence from symptoms alone, map which scripts it reaches, measure how widely it is deployed, and release a 65,536-entry Nepali-English tokenizer with a harness that regenerates every number here from public data on a laptop.

arXiv

Don't Overthink, Don't Underthink: Toward Adaptive Reasoning in Agentic AI

Md Jueal Mia, M. Hadi Amini

Recent advances in Large Language Models (LLMs) have shown that increased inference-time reasoning can improve performance on complex tasks. However, many existing approaches rely on fixed or preallocated reasoning controls, such as fixed token budgets, pre-execution difficulty estimates, or activation-space interventions, and are often evaluated on standalone reasoning benchmarks rather than full agentic workflows. These assumptions may not hold in agentic AI systems, where reasoning requirements evolve dynamically through planning, tool use, memory retrieval, and agent-to-agent interactions. Consequently, reasoning can become either excessive or insufficient, resulting in unnecessary computation, increased latency, planning drift, excessive tool use, or incomplete solutions. We argue that a major challenge for next-generation agentic AI is not merely how much reasoning a language model should perform, but how it should allocate reasoning according to evolving task demands. We characterize over-reasoning and under-reasoning as recurring failure modes of misallocated reasoning and evaluate them on MATH-500 and the GAIA public validation benchmark. Using tool-decision latency, token consumption, token-limit exhaustion, and answer correctness, our results suggest that cases classified as over-reasoning are associated with higher computational cost without proportional accuracy gains, whereas cases classified as under-reasoning are consistently associated with incorrect or incomplete solutions. These findings motivate future research on adaptive reasoning mechanisms for agentic AI.

arXiv

AfriSwitch: A Benchmark for In-the-Wild African Code-Switched Speech Recognition

Gabrial Zencha Ashungafac, Busayo Awobade, Tobi Olatunji

Code-switching is pervasive in bilingual African conversation, yet most ASR systems assume monolingual input and are evaluated on curated monolingual benchmarks. We present AfriSwitch, a 61.36-hour human-transcribed benchmark of in-the-wild code-switched speech spanning 16 African languages and language varieties, released with switch-level English span tags, perutterance Code-Mixing Index (CMI), and switch-point counts. Corpus statistics show that mixing behaviour varies widely across African languages along two largely independent axes: how often speakers alternate, and how balanced the mixture is. No single scalar captures how code-switched a language is. Benchmarking five open and commercial multilingual ASR systems zero-shot yields word error rates far above published monolingual figures for the same languages, with the best system averaging 35.93% WER and no system falling below 24% on any language. Africa-targeted training, not model scale or nominal language coverage, best predicts performance.

arXiv

SpeechGym: An Audio-Native Gym for Training Voice Agents via Reinforcement Learning

Jiajun Fan, Jingyuan Li, Prashanth Gurunath Shivakumar, Jia-Hong Huang, Qi Luo, M. Maruf, Ivan Bulyko, Ge Liu, Roger Ren

Voice agents must call tools and hold multi-turn dialogue entirely through speech, yet the dominant paradigm trains them in text. Existing frameworks either cascade TTS and ASR around a proprietary voice API, where gradients cannot flow and per-call cost makes on-policy reinforcement learning prohibitive, or stay in text: they measure voice agents but cannot improve them. We present SpeechGym, an audio-native agentic environment in which two omni-modal models converse in native audio, with no external ASR or TTS and no API boundary, over the unmodified tasks, tools and success check of an established text agentic benchmark, so that the interaction modality is the only variable and the loop stays local and trainable end to end. Audio agentic capability does not follow from audio understanding. The failures speech introduces are perceptual rather than reasoning deficits: the agent picks the right tool and the right argument slot but fills it with a value misheard from the waveform, and that single error cascades into a failed call, a retry of the same call, and a wasted step budget. A second failure is behavioural: under an insistent caller the agent performs an unauthorised write and ends the episode believing it helped. Both are trainable, because the environment labels them for free: a call with a misheard argument fails against the database while a correct one succeeds. The obstacle is sparsity, not signal. Outcome-only GRPO is gradient-starved here, since almost every rollout group fails identically, while a per-turn process reward crediting each successful tool call restores variance to nearly every group. Trained this way, the agent transfers with no further tuning to an independently implemented voice benchmark, more than doubling task success and carrying an open-weights model from last place to second on that leaderboard, while using fewer turns and tokens than before training.

arXiv

The Latent Diagnostic Taxonomy: A Framework for Constructing Classifiers and Diagnosing Their Decisions, Applied to Prompt Injection Detection

Jaturong Kongmanee, Smile Thanapattheerakul

This paper proposes a framework for constructing a classifier as a safeguard layer, and for developing a complementary diagnostic that identifies which of the classifier's confident decisions can be trusted. This framework, the Latent Diagnostic Taxonomy, consists of (i) constructing a dimensionality-optimized classifier, in which the embedding dimensionality is empirically selected via cross-validated performance rather than fixed a priori, (ii) locating a relatively small set of latent support vectors (~ 29% of total training examples) representing influential prompts for identifying tokens that alter the classifier's predicted labels, and (iii) utilizing such tokens and their associated attack magnitudes for constructing a diagnostic taxonomy. This diagnostic taxonomy provides an end-to-end guideline for flagging prompts that require different treatments: rely Safely on the classifier's decision; flag Heuristic Bias and Heuristic Override cases; route Insufficient Context cases for further human/safety review. Applying the framework to a classifier trained on a public prompt injection dataset, we find that a substantial fraction of its confident decisions (~ 77%) are not robust to removing a single token, and that this brittleness separates into two distinct failure patterns: a confidence calibration failure and a genuinely exploitable shortcut. For each zone of the taxonomy, we also recommend strategies for remediating diagnosed prompts. We illustrate the framework as a series of steps, demonstrating how each step operates.

arXiv

Case2Flow: Bridging Patient Cases and Guideline Flowcharts through Multimodal Retrieval

Jiale Wei, Yufan Chen, Alexander Jaus, Zdravko Marinov, Julian Friedrich, Simon Reiß, Jens Kleesiek, Rainer Stiefelhagen

Medical guidelines encode rich, evidence-based decision logic, yet the specific decision artifact a clinician needs is hard to locate within a guideline, let alone across guidelines covering plausible diseases and treatments. While guideline passages have supported end-to-end question answering, flowcharts remain largely underused in decision support despite their ability to encode actionable clinical pathways. We therefore introduce Case2Flow, a task designed to retrieve the most relevant guideline flowchart for a given patient case from a collection of guideline documents. To support it, we construct FlowAtlas, a curated corpus of 202 flowcharts extracted from 2,080 medical guidelines, together with a pipeline that synthesises 1,911 aligned case-flowchart pairs. Our evaluation of multimodal retrieval methods reveals systematic failure modes, including overreliance on keywords and spurious token-patch matches induced by uninformative background regions in flowcharts. Motivated by this, we propose CRISP, a training-free scoring method that sharpens late-interaction retrieval by suppressing uninformative patches, discounting ambiguous token matches, and incorporating bidirectional query-image alignment. CRISP improves Recall@1 by up to 18.71 percentage points, while a blinded physician assessment on published case narratives provides preliminary feasibility evidence beyond synthetic queries.

arXiv

LowRankArena: A Standardized Evaluation Platform for SVD-Based LLM Compression

Zishan Shao, Lixun Zhang, Kangning Cui, Wenhao Wu, Jinhee Kim, Yixiao Wang, Ting Jiang, Hancheng Ye, Qinsi Wang, Fan Yang, Danyang Zhuo, Yiran Chen, Hai Li

SVD-based low-rank compression has become a fast-growing direction for reducing the memory and computational cost of large language models (LLMs). However, meaningful comparison across existing studies remains difficult as prior evaluations use varied benchmarks, inconsistent ratios, and diverse setups, often failing to isolate low-rank effects from auxiliary techniques. As a result, it remains unclear whether reported gains reflect method-level improvements or differences in evaluation protocol. This lack of comparability highlights the need for a unified, reproducible evaluation platform. To address this problem, we present LowRankArena, a standardized evaluation platform for SVD-based LLM compression. LowRankArena unifies task versions, uniform-precision compression budgets, comparison regimes, and inference measurements, and provides a reproducible pipeline with over 3 TiB released compressed checkpoints. Using LowRankArena, our aligned audit of five representative SVD methods reveals that prior findings are highly conditional under standardized protocols: clear leaders and performance tiers shift across backbones and keep ratios, multiple-choice accuracy can hide large perplexity degradation, and nominal low-rank savings yield workload-dependent and often limited end-to-end speedups. Our code is available at: https://github.com/Zishan-Shao/lowrankarena.git.

arXiv

Co-Evolving Structured Knowledge and Reasoning in Language Models

Ryan Thomas Noonan, Linxi Zhao, Menghan Xu, Akanksha Sarkar, Mihir Mishra, Dongyoung Go, Kilian Q Weinberger, Yoav Artzi, Jennifer J. Sun

Retrieval-augmented methods improve factual accuracy by grounding language models in external knowledge, but retrieving over unstructured text often introduces irrelevant context and offers limited control over the retrieved information. Structured knowledge bases offer a more controllable alternative, yet they are expensive to construct and often brittle to reason over. To address these limitations, we propose KBevo: a co-evolving framework that jointly learns to construct a structured knowledge base and reason over it for knowledge-intensive question answering. By optimizing both components end-to-end with QA outcome rewards, our method enables reasoning success to directly improve the quality of the constructed knowledge base. This leads to larger, better-connected knowledge structures with higher answer reachability, while also improving compositional factual reasoning and controllability compared to standard retrieval baselines.

arXiv

Why RAGs Hallucinate: Penalty-Aware Evaluation of Retrieval-Augmented Generation Systems with Knowledge-Gap Canaries

Alden Do Rosario, Hussein Younes, Felipe Pires

Volume-based accuracy rewards retrieval-augmented generation (RAG) systems for guessing: a system that answers everything outscores one that declines when its knowledge base cannot support an answer. Building on the confidence-target analysis of Kalai et al. (2025), we present a penalty-aware evaluation framework for deployed RAG products, combining (i) asymmetric scoring (correct +1, wrong -4, abstain 0), (ii) knowledge-gap canaries, questions whose answers are verifiably absent from the knowledge base, so that any answer constitutes ungrounded generation from parametric memory, and (iii) a failure-attribution pipeline that separates retrieval, generation, and abstention-policy failures. Applying the framework to three commercial RAG systems and a no-retrieval baseline on SimpleQA-Verified (1,000 questions x 3 repeats, graded blind by a cross-family three-judge panel with 98.9% unanimity), we find that accuracy when answering is closely clustered across systems (97.0-98.0%), while canary violation rates differ roughly sixfold (16.7% vs. 98.1%). The systems are separated less by what they answer correctly than by whether they answer at all when they should not, and penalty-aware scoring reorders the volume-based ranking accordingly; the reordering is stable across penalty settings from k=1 to k=9. All code, configurations, transcripts, and judge votes are released for independent audit.

继续浏览

更多 AI 顶会论文

30+ 个顶会，按会议、年份和主题浏览。

浏览全部会议

进入任一会议和年份，查看完整论文列表。

全部会议 →

搜索论文

按标题、作者或关键词，在 21 万+ 篇论文里搜索。

搜索论文 →