arxiv_id
float64
1.5k
2.51k
title
stringlengths
9
178
authors
stringlengths
2
22.8k
categories
stringlengths
4
146
summary
stringlengths
103
1.92k
published
stringdate
2015-02-06 10:44:00
2025-07-10 17:59:58
comments
stringlengths
2
417
journal_ref
stringclasses
321 values
doi
stringclasses
398 values
ss_title
stringlengths
8
159
ss_authors
stringlengths
11
8.38k
ss_year
float64
2.02k
2.03k
ss_venue
stringclasses
281 values
ss_citationCount
float64
0
134k
ss_referenceCount
float64
0
429
ss_fieldsOfStudy
stringclasses
47 values
2,505.14674
Reward Reasoning Model
['Jiaxin Guo', 'Zewen Chi', 'Li Dong', 'Qingxiu Dong', 'Xun Wu', 'Shaohan Huang', 'Furu Wei']
['cs.CL']
Reward models play a critical role in guiding large language models toward outputs that align with human expectations. However, an open challenge remains in effectively utilizing test-time compute to enhance reward model performance. In this work, we introduce Reward Reasoning Models (RRMs), which are specifically designed to execute a deliberate reasoning process before generating final rewards. Through chain-of-thought reasoning, RRMs leverage additional test-time compute for complex queries where appropriate rewards are not immediately apparent. To develop RRMs, we implement a reinforcement learning framework that fosters self-evolved reward reasoning capabilities without requiring explicit reasoning traces as training data. Experimental results demonstrate that RRMs achieve superior performance on reward modeling benchmarks across diverse domains. Notably, we show that RRMs can adaptively exploit test-time compute to further improve reward accuracy. The pretrained reward reasoning models are available at https://huggingface.co/Reward-Reasoning.
2025-05-20T17:58:03Z
null
null
null
Reward Reasoning Model
['Jiaxin Guo', 'Zewen Chi', 'Li Dong', 'Qingxiu Dong', 'Xun Wu', 'Shaohan Huang', 'Furu Wei']
2,025
arXiv.org
1
74
['Computer Science']
2,505.14677
Visionary-R1: Mitigating Shortcuts in Visual Reasoning with Reinforcement Learning
['Jiaer Xia', 'Yuhang Zang', 'Peng Gao', 'Yixuan Li', 'Kaiyang Zhou']
['cs.CV']
Learning general-purpose reasoning capabilities has long been a challenging problem in AI. Recent research in large language models (LLMs), such as DeepSeek-R1, has shown that reinforcement learning techniques like GRPO can enable pre-trained LLMs to develop reasoning capabilities using simple question-answer pairs. In this paper, we aim to train visual language models (VLMs) to perform reasoning on image data through reinforcement learning and visual question-answer pairs, without any explicit chain-of-thought (CoT) supervision. Our findings indicate that simply applying reinforcement learning to a VLM -- by prompting the model to produce a reasoning chain before providing an answer -- can lead the model to develop shortcuts from easy questions, thereby reducing its ability to generalize across unseen data distributions. We argue that the key to mitigating shortcut learning is to encourage the model to interpret images prior to reasoning. Therefore, we train the model to adhere to a caption-reason-answer output format: initially generating a detailed caption for an image, followed by constructing an extensive reasoning chain. When trained on 273K CoT-free visual question-answer pairs and using only reinforcement learning, our model, named Visionary-R1, outperforms strong multimodal models, such as GPT-4o, Claude3.5-Sonnet, and Gemini-1.5-Pro, on multiple visual reasoning benchmarks.
2025-05-20T17:58:35Z
null
null
null
Visionary-R1: Mitigating Shortcuts in Visual Reasoning with Reinforcement Learning
['Jiaer Xia', 'Y.-F. Zang', 'Peng Gao', 'Yixuan Li', 'Kaiyang Zhou']
2,025
arXiv.org
0
47
['Computer Science']
2,505.14683
Emerging Properties in Unified Multimodal Pretraining
['Chaorui Deng', 'Deyao Zhu', 'Kunchang Li', 'Chenhui Gou', 'Feng Li', 'Zeyu Wang', 'Shu Zhong', 'Weihao Yu', 'Xiaonan Nie', 'Ziang Song', 'Guang Shi', 'Haoqi Fan']
['cs.CV']
Unifying multimodal understanding and generation has shown impressive capabilities in cutting-edge proprietary systems. In this work, we introduce BAGEL, an open-source foundational model that natively supports multimodal understanding and generation. BAGEL is a unified, decoder-only model pretrained on trillions of tokens curated from large-scale interleaved text, image, video, and web data. When scaled with such diverse multimodal interleaved data, BAGEL exhibits emerging capabilities in complex multimodal reasoning. As a result, it significantly outperforms open-source unified models in both multimodal generation and understanding across standard benchmarks, while exhibiting advanced multimodal reasoning abilities such as free-form image manipulation, future frame prediction, 3D manipulation, and world navigation. In the hope of facilitating further opportunities for multimodal research, we share the key findings, pretraining details, data creation protocal, and release our code and checkpoints to the community. The project page is at https://bagel-ai.org/
2025-05-20T17:59:30Z
37 pages, 17 figures
null
null
null
null
null
null
null
null
null
2,505.14684
Mind the Gap: Bridging Thought Leap for Improved Chain-of-Thought Tuning
['Haolei Xu', 'Yuchen Yan', 'Yongliang Shen', 'Wenqi Zhang', 'Guiyang Hou', 'Shengpei Jiang', 'Kaitao Song', 'Weiming Lu', 'Jun Xiao', 'Yueting Zhuang']
['cs.CL', 'cs.AI']
Large language models (LLMs) have achieved remarkable progress on mathematical tasks through Chain-of-Thought (CoT) reasoning. However, existing mathematical CoT datasets often suffer from Thought Leaps due to experts omitting intermediate steps, which negatively impacts model learning and generalization. We propose the CoT Thought Leap Bridge Task, which aims to automatically detect leaps and generate missing intermediate reasoning steps to restore the completeness and coherence of CoT. To facilitate this, we constructed a specialized training dataset called ScaleQM+, based on the structured ScaleQuestMath dataset, and trained CoT-Bridge to bridge thought leaps. Through comprehensive experiments on mathematical reasoning benchmarks, we demonstrate that models fine-tuned on bridged datasets consistently outperform those trained on original datasets, with improvements of up to +5.87% on NuminaMath. Our approach effectively enhances distilled data (+3.02%) and provides better starting points for reinforcement learning (+3.1%), functioning as a plug-and-play module compatible with existing optimization techniques. Furthermore, CoT-Bridge demonstrate improved generalization to out-of-domain logical reasoning tasks, confirming that enhancing reasoning completeness yields broadly applicable benefits.
2025-05-20T17:59:31Z
Project: https://zju-real.github.io/CoT-Bridge/
null
null
null
null
null
null
null
null
null
2,505.14766
This Time is Different: An Observability Perspective on Time Series Foundation Models
['Ben Cohen', 'Emaad Khwaja', 'Youssef Doubli', 'Salahidine Lemaachi', 'Chris Lettieri', 'Charles Masson', 'Hugo Miccinilli', 'Elise Ramé', 'Qiqi Ren', 'Afshin Rostamizadeh', 'Jean Ogier du Terrail', 'Anna-Monica Toon', 'Kan Wang', 'Stephan Xie', 'Zongzhe Xu', 'Viktoriya Zhukova', 'David Asker', 'Ameet Talwalkar', 'Othmane Abou-Amal']
['cs.LG', 'cs.AI']
We introduce Toto, a time series forecasting foundation model with 151 million parameters. Toto uses a modern decoder-only architecture coupled with architectural innovations designed to account for specific challenges found in multivariate observability time series data. Toto's pre-training corpus is a mixture of observability data, open datasets, and synthetic data, and is 4-10$\times$ larger than those of leading time series foundation models. Additionally, we introduce BOOM, a large-scale benchmark consisting of 350 million observations across 2,807 real-world time series. For both Toto and BOOM, we source observability data exclusively from Datadog's own telemetry and internal observability metrics. Extensive evaluations demonstrate that Toto achieves state-of-the-art performance on both BOOM and on established general purpose time series forecasting benchmarks. Toto's model weights, inference code, and evaluation scripts, as well as BOOM's data and evaluation code, are all available as open source under the Apache 2.0 License available at https://huggingface.co/Datadog/Toto-Open-Base-1.0 and https://github.com/DataDog/toto.
2025-05-20T17:48:13Z
null
null
null
null
null
null
null
null
null
null
2,505.1481
Scaling Reasoning, Losing Control: Evaluating Instruction Following in Large Reasoning Models
['Tingchen Fu', 'Jiawei Gu', 'Yafu Li', 'Xiaoye Qu', 'Yu Cheng']
['cs.CL', 'cs.AI']
Instruction-following is essential for aligning large language models (LLMs) with user intent. While recent reasoning-oriented models exhibit impressive performance on complex mathematical problems, their ability to adhere to natural language instructions remains underexplored. In this work, we introduce MathIF, a dedicated benchmark for evaluating instruction-following in mathematical reasoning tasks. Our empirical analysis reveals a consistent tension between scaling up reasoning capacity and maintaining controllability, as models that reason more effectively often struggle to comply with user directives. We find that models tuned on distilled long chains-of-thought or trained with reasoning-oriented reinforcement learning often degrade in instruction adherence, especially when generation length increases. Furthermore, we show that even simple interventions can partially recover obedience, though at the cost of reasoning performance. These findings highlight a fundamental tension in current LLM training paradigms and motivate the need for more instruction-aware reasoning models. We release the code and data at https://github.com/TingchenFu/MathIF.
2025-05-20T18:18:01Z
null
null
null
Scaling Reasoning, Losing Control: Evaluating Instruction Following in Large Reasoning Models
['Ting Fu', 'Jiawei Gu', 'Yafu Li', 'Xiaoye Qu', 'Yu Cheng']
2,025
arXiv.org
1
44
['Computer Science']
2,505.14884
Polar Sparsity: High Throughput Batched LLM Inferencing with Scalable Contextual Sparsity
['Susav Shrestha', 'Brad Settlemyer', 'Nikoli Dryden', 'Narasimha Reddy']
['cs.LG', 'cs.AI']
Accelerating large language model (LLM) inference is critical for real-world deployments requiring high throughput and low latency. Contextual sparsity, where each token dynamically activates only a small subset of the model parameters, shows promise but does not scale to large batch sizes due to union of active neurons quickly approaching dense computation. We introduce Polar Sparsity, highlighting a key shift in sparsity importance from MLP to Attention layers as we scale batch size and sequence length. While MLP layers become more compute-efficient under batching, their sparsity vanishes. In contrast, attention becomes increasingly more expensive at scale, while their head sparsity remains stable and batch-invariant. We develop hardware-efficient, sparsity-aware GPU kernels for selective MLP and Attention computations, delivering up to \(2.2\times\) end-to-end speedups for models like OPT, LLaMA-2 \& 3, across various batch sizes and sequence lengths without compromising accuracy. To our knowledge, this is the first work to demonstrate that contextual sparsity can scale effectively to large batch sizes, delivering substantial inference acceleration with minimal changes, making Polar Sparsity practical for large-scale, high-throughput LLM deployment systems. Our code is available at: https://github.com/susavlsh10/Polar-Sparsity.
2025-05-20T20:15:42Z
null
null
null
null
null
null
null
null
null
null
2,505.14969
STree: Speculative Tree Decoding for Hybrid State-Space Models
['Yangchao Wu', 'Zongyue Qin', 'Alex Wong', 'Stefano Soatto']
['cs.LG', 'cs.AI']
Speculative decoding is a technique to leverage hardware concurrency to improve the efficiency of large-scale autoregressive (AR) Transformer models by enabling multiple steps of token generation in a single forward pass. State-space models (SSMs) are already more efficient than AR Transformers, since their state summarizes all past data with no need to cache or re-process tokens in the sliding window context. However, their state can also comprise thousands of tokens; so, speculative decoding has recently been extended to SSMs. Existing approaches, however, do not leverage the tree-based verification methods, since current SSMs lack the means to compute a token tree efficiently. We propose the first scalable algorithm to perform tree-based speculative decoding in state-space models (SSMs) and hybrid architectures of SSMs and Transformer layers. We exploit the structure of accumulated state transition matrices to facilitate tree-based speculative decoding with minimal overhead to current SSM state update implementations. With the algorithm, we describe a hardware-aware implementation that improves naive application of AR Transformer tree-based speculative decoding methods to SSMs. Furthermore, we outperform vanilla speculative decoding with SSMs even with a baseline drafting model and tree structure on three different benchmarks, opening up opportunities for further speed up with SSM and hybrid model inference. Code will be released upon paper acceptance.
2025-05-20T23:12:16Z
null
null
null
null
null
null
null
null
null
null
2,505.15093
Steering Generative Models with Experimental Data for Protein Fitness Optimization
['Jason Yang', 'Wenda Chu', 'Daniel Khalil', 'Raul Astudillo', 'Bruce J. Wittmann', 'Frances H. Arnold', 'Yisong Yue']
['q-bio.BM', 'cs.LG']
Protein fitness optimization involves finding a protein sequence that maximizes desired quantitative properties in a combinatorially large design space of possible sequences. Recent developments in steering protein generative models (e.g diffusion models, language models) offer a promising approach. However, by and large, past studies have optimized surrogate rewards and/or utilized large amounts of labeled data for steering, making it unclear how well existing methods perform and compare to each other in real-world optimization campaigns where fitness is measured by low-throughput wet-lab assays. In this study, we explore fitness optimization using small amounts (hundreds) of labeled sequence-fitness pairs and comprehensively evaluate strategies such as classifier guidance and posterior sampling for guiding generation from different discrete diffusion models of protein sequences. We also demonstrate how guidance can be integrated into adaptive sequence selection akin to Thompson sampling in Bayesian optimization, showing that plug-and-play guidance strategies offer advantages compared to alternatives such as reinforcement learning with protein language models.
2025-05-21T04:30:48Z
null
null
null
null
null
null
null
null
null
null
2,505.15263
gen2seg: Generative Models Enable Generalizable Instance Segmentation
['Om Khangaonkar', 'Hamed Pirsiavash']
['cs.CV', 'cs.LG']
By pretraining to synthesize coherent images from perturbed inputs, generative models inherently learn to understand object boundaries and scene compositions. How can we repurpose these generative representations for general-purpose perceptual organization? We finetune Stable Diffusion and MAE (encoder+decoder) for category-agnostic instance segmentation using our instance coloring loss exclusively on a narrow set of object types (indoor furnishings and cars). Surprisingly, our models exhibit strong zero-shot generalization, accurately segmenting objects of types and styles unseen in finetuning (and in many cases, MAE's ImageNet-1K pretraining too). Our best-performing models closely approach the heavily supervised SAM when evaluated on unseen object types and styles, and outperform it when segmenting fine structures and ambiguous boundaries. In contrast, existing promptable segmentation architectures or discriminatively pretrained models fail to generalize. This suggests that generative models learn an inherent grouping mechanism that transfers across categories and domains, even without internet-scale pretraining. Code, pretrained models, and demos are available on our website.
2025-05-21T08:42:05Z
Website: https://reachomk.github.io/gen2seg/
null
null
gen2seg: Generative Models Enable Generalizable Instance Segmentation
['Om Khangaonkar', 'Hamed Pirsiavash']
2,025
arXiv.org
0
65
['Computer Science']
2,505.1527
Scaling Diffusion Transformers Efficiently via $μ$P
['Chenyu Zheng', 'Xinyu Zhang', 'Rongzhen Wang', 'Wei Huang', 'Zhi Tian', 'Weilin Huang', 'Jun Zhu', 'Chongxuan Li']
['cs.LG', 'cs.AI', 'cs.CV']
Diffusion Transformers have emerged as the foundation for vision generative models, but their scalability is limited by the high cost of hyperparameter (HP) tuning at large scales. Recently, Maximal Update Parametrization ($\mu$P) was proposed for vanilla Transformers, which enables stable HP transfer from small to large language models, and dramatically reduces tuning costs. However, it remains unclear whether $\mu$P of vanilla Transformers extends to diffusion Transformers, which differ architecturally and objectively. In this work, we generalize standard $\mu$P to diffusion Transformers and validate its effectiveness through large-scale experiments. First, we rigorously prove that $\mu$P of mainstream diffusion Transformers, including DiT, U-ViT, PixArt-$\alpha$, and MMDiT, aligns with that of the vanilla Transformer, enabling the direct application of existing $\mu$P methodologies. Leveraging this result, we systematically demonstrate that DiT-$\mu$P enjoys robust HP transferability. Notably, DiT-XL-2-$\mu$P with transferred learning rate achieves 2.9 times faster convergence than the original DiT-XL-2. Finally, we validate the effectiveness of $\mu$P on text-to-image generation by scaling PixArt-$\alpha$ from 0.04B to 0.61B and MMDiT from 0.18B to 18B. In both cases, models under $\mu$P outperform their respective baselines while requiring small tuning cost, only 5.5% of one training run for PixArt-$\alpha$ and 3% of consumption by human experts for MMDiT-18B. These results establish $\mu$P as a principled and efficient framework for scaling diffusion Transformers.
2025-05-21T08:49:03Z
35 pages, 10 figures, 15 tables
null
null
null
null
null
null
null
null
null
2,505.15277
Web-Shepherd: Advancing PRMs for Reinforcing Web Agents
['Hyungjoo Chae', 'Sunghwan Kim', 'Junhee Cho', 'Seungone Kim', 'Seungjun Moon', 'Gyeom Hwangbo', 'Dongha Lim', 'Minjin Kim', 'Yeonjun Hwang', 'Minju Gwak', 'Dongwook Choi', 'Minseok Kang', 'Gwanhoon Im', 'ByeongUng Cho', 'Hyojun Kim', 'Jun Hee Han', 'Taeyoon Kwon', 'Minju Kim', 'Beong-woo Kwak', 'Dongjin Kang', 'Jinyoung Yeo']
['cs.CL']
Web navigation is a unique domain that can automate many repetitive real-life tasks and is challenging as it requires long-horizon sequential decision making beyond typical multimodal large language model (MLLM) tasks. Yet, specialized reward models for web navigation that can be utilized during both training and test-time have been absent until now. Despite the importance of speed and cost-effectiveness, prior works have utilized MLLMs as reward models, which poses significant constraints for real-world deployment. To address this, in this work, we propose the first process reward model (PRM) called Web-Shepherd which could assess web navigation trajectories in a step-level. To achieve this, we first construct the WebPRM Collection, a large-scale dataset with 40K step-level preference pairs and annotated checklists spanning diverse domains and difficulty levels. Next, we also introduce the WebRewardBench, the first meta-evaluation benchmark for evaluating PRMs. In our experiments, we observe that our Web-Shepherd achieves about 30 points better accuracy compared to using GPT-4o on WebRewardBench. Furthermore, when testing on WebArena-lite by using GPT-4o-mini as the policy and Web-Shepherd as the verifier, we achieve 10.9 points better performance, in 10 less cost compared to using GPT-4o-mini as the verifier. Our model, dataset, and code are publicly available at LINK.
2025-05-21T08:56:55Z
Work in progress
null
null
null
null
null
null
null
null
null
2,505.15379
The P$^3$ dataset: Pixels, Points and Polygons for Multimodal Building Vectorization
['Raphael Sulzer', 'Liuyun Duan', 'Nicolas Girard', 'Florent Lafarge']
['cs.CV']
We present the P$^3$ dataset, a large-scale multimodal benchmark for building vectorization, constructed from aerial LiDAR point clouds, high-resolution aerial imagery, and vectorized 2D building outlines, collected across three continents. The dataset contains over 10 billion LiDAR points with decimeter-level accuracy and RGB images at a ground sampling distance of 25 centimeter. While many existing datasets primarily focus on the image modality, P$^3$ offers a complementary perspective by also incorporating dense 3D information. We demonstrate that LiDAR point clouds serve as a robust modality for predicting building polygons, both in hybrid and end-to-end learning frameworks. Moreover, fusing aerial LiDAR and imagery further improves accuracy and geometric quality of predicted polygons. The P$^3$ dataset is publicly available, along with code and pretrained weights of three state-of-the-art models for building polygon prediction at https://github.com/raphaelsulzer/PixelsPointsPolygons .
2025-05-21T11:16:29Z
null
null
null
The P3 dataset: Pixels, Points and Polygons for Multimodal Building Vectorization
['Raphael Sulzer', 'Liuyun Duan', 'Nicolas Girard', 'Florent Lafarge']
2,025
arXiv.org
0
39
['Computer Science']
2,505.15425
On the Robustness of Medical Vision-Language Models: Are they Truly Generalizable?
['Raza Imam', 'Rufael Marew', 'Mohammad Yaqub']
['cs.CV']
Medical Vision-Language Models (MVLMs) have achieved par excellence generalization in medical image analysis, yet their performance under noisy, corrupted conditions remains largely untested. Clinical imaging is inherently susceptible to acquisition artifacts and noise; however, existing evaluations predominantly assess generally clean datasets, overlooking robustness -- i.e., the model's ability to perform under real-world distortions. To address this gap, we first introduce MediMeta-C, a corruption benchmark that systematically applies several perturbations across multiple medical imaging datasets. Combined with MedMNIST-C, this establishes a comprehensive robustness evaluation framework for MVLMs. We further propose RobustMedCLIP, a visual encoder adaptation of a pretrained MVLM that incorporates few-shot tuning to enhance resilience against corruptions. Through extensive experiments, we benchmark 5 major MVLMs across 5 medical imaging modalities, revealing that existing models exhibit severe degradation under corruption and struggle with domain-modality tradeoffs. Our findings highlight the necessity of diverse training and robust adaptation strategies, demonstrating that efficient low-rank adaptation when paired with few-shot tuning, improves robustness while preserving generalization across modalities.
2025-05-21T12:08:31Z
Dataset and Code is available at https://github.com/BioMedIA-MBZUAI/RobustMedCLIP Accepted at: Medical Image Understanding and Analysis (MIUA) 2025
null
null
null
null
null
null
null
null
null
2,505.15436
Chain-of-Focus: Adaptive Visual Search and Zooming for Multimodal Reasoning via RL
['Xintong Zhang', 'Zhi Gao', 'Bofei Zhang', 'Pengxiang Li', 'Xiaowen Zhang', 'Yang Liu', 'Tao Yuan', 'Yuwei Wu', 'Yunde Jia', 'Song-Chun Zhu', 'Qing Li']
['cs.CV']
Vision language models (VLMs) have achieved impressive performance across a variety of computer vision tasks. However, the multimodal reasoning capability has not been fully explored in existing models. In this paper, we propose a Chain-of-Focus (CoF) method that allows VLMs to perform adaptive focusing and zooming in on key image regions based on obtained visual cues and the given questions, achieving efficient multimodal reasoning. To enable this CoF capability, we present a two-stage training pipeline, including supervised fine-tuning (SFT) and reinforcement learning (RL). In the SFT stage, we construct the MM-CoF dataset, comprising 3K samples derived from a visual agent designed to adaptively identify key regions to solve visual tasks with different image resolutions and questions. We use MM-CoF to fine-tune the Qwen2.5-VL model for cold start. In the RL stage, we leverage the outcome accuracies and formats as rewards to update the Qwen2.5-VL model, enabling further refining the search and reasoning strategy of models without human priors. Our model achieves significant improvements on multiple benchmarks. On the V* benchmark that requires strong visual reasoning capability, our model outperforms existing VLMs by 5% among 8 image resolutions ranging from 224 to 4K, demonstrating the effectiveness of the proposed CoF method and facilitating the more efficient deployment of VLMs in practical applications.
2025-05-21T12:18:15Z
null
null
null
Chain-of-Focus: Adaptive Visual Search and Zooming for Multimodal Reasoning via RL
['Xintong Zhang', 'Zhi Gao', 'Bofei Zhang', 'Pengxiang Li', 'Xiaowen Zhang', 'Yang Liu', 'Tao Yuan', 'Yuwei Wu', 'Yunde Jia', 'Song-Chun Zhu', 'Qing Li']
2,025
arXiv.org
0
49
['Computer Science']
2,505.15607
From Problem-Solving to Teaching Problem-Solving: Aligning LLMs with Pedagogy using Reinforcement Learning
['David Dinucu-Jianu', 'Jakub Macina', 'Nico Daheim', 'Ido Hakimi', 'Iryna Gurevych', 'Mrinmaya Sachan']
['cs.CL', 'cs.AI']
Large language models (LLMs) can transform education, but their optimization for direct question-answering often undermines effective pedagogy which requires strategically withholding answers. To mitigate this, we propose an online reinforcement learning (RL)-based alignment framework that can quickly adapt LLMs into effective tutors using simulated student-tutor interactions by emphasizing pedagogical quality and guided problem-solving over simply giving away answers. We use our method to train a 7B parameter tutor model without human annotations which reaches similar performance to larger proprietary models like LearnLM. We introduce a controllable reward weighting to balance pedagogical support and student solving accuracy, allowing us to trace the Pareto frontier between these two objectives. Our models better preserve reasoning capabilities than single-turn SFT baselines and can optionally enhance interpretability through thinking tags that expose the model's instructional planning.
2025-05-21T15:00:07Z
David Dinucu-Jianu and Jakub Macina contributed equally. Code available: https://github.com/eth-lre/PedagogicalRL
null
null
null
null
null
null
null
null
null
2,505.15776
ConvSearch-R1: Enhancing Query Reformulation for Conversational Search with Reasoning via Reinforcement Learning
['Changtai Zhu', 'Siyin Wang', 'Ruijun Feng', 'Kai Song', 'Xipeng Qiu']
['cs.CL', 'cs.IR']
Conversational search systems require effective handling of context-dependent queries that often contain ambiguity, omission, and coreference. Conversational Query Reformulation (CQR) addresses this challenge by transforming these queries into self-contained forms suitable for off-the-shelf retrievers. However, existing CQR approaches suffer from two critical constraints: high dependency on costly external supervision from human annotations or large language models, and insufficient alignment between the rewriting model and downstream retrievers. We present ConvSearch-R1, the first self-driven framework that completely eliminates dependency on external rewrite supervision by leveraging reinforcement learning to optimize reformulation directly through retrieval signals. Our novel two-stage approach combines Self-Driven Policy Warm-Up to address the cold-start problem through retrieval-guided self-distillation, followed by Retrieval-Guided Reinforcement Learning with a specially designed rank-incentive reward shaping mechanism that addresses the sparsity issue in conventional retrieval metrics. Extensive experiments on TopiOCQA and QReCC datasets demonstrate that ConvSearch-R1 significantly outperforms previous state-of-the-art methods, achieving over 10% improvement on the challenging TopiOCQA dataset while using smaller 3B parameter models without any external supervision.
2025-05-21T17:27:42Z
null
null
null
null
null
null
null
null
null
null
2,505.15801
VerifyBench: Benchmarking Reference-based Reward Systems for Large Language Models
['Yuchen Yan', 'Jin Jiang', 'Zhenbang Ren', 'Yijun Li', 'Xudong Cai', 'Yang Liu', 'Xin Xu', 'Mengdi Zhang', 'Jian Shao', 'Yongliang Shen', 'Jun Xiao', 'Yueting Zhuang']
['cs.CL', 'cs.AI']
Large reasoning models such as OpenAI o1 and DeepSeek-R1 have achieved remarkable performance in the domain of reasoning. A key component of their training is the incorporation of verifiable rewards within reinforcement learning (RL). However, existing reward benchmarks do not evaluate reference-based reward systems, leaving researchers with limited understanding of the accuracy of verifiers used in RL. In this paper, we introduce two benchmarks, VerifyBench and VerifyBench-Hard, designed to assess the performance of reference-based reward systems. These benchmarks are constructed through meticulous data collection and curation, followed by careful human annotation to ensure high quality. Current models still show considerable room for improvement on both VerifyBench and VerifyBench-Hard, especially smaller-scale models. Furthermore, we conduct a thorough and comprehensive analysis of evaluation results, offering insights for understanding and developing reference-based reward systems. Our proposed benchmarks serve as effective tools for guiding the development of verifier accuracy and the reasoning capabilities of models trained via RL in reasoning tasks.
2025-05-21T17:54:43Z
Project Page: https://zju-real.github.io/VerifyBench Dataset: https://huggingface.co/datasets/ZJU-REAL/VerifyBench Code: https://github.com/ZJU-REAL/VerifyBench
null
null
VerifyBench: Benchmarking Reference-based Reward Systems for Large Language Models
['Yuchen Yan', 'Jin Jiang', 'Zhenbang Ren', 'Yijun Li', 'Xudong Cai', 'Yang Liu', 'Xin Xu', 'Mengdi Zhang', 'Jian Shao', 'Yongliang Shen', 'Jun Xiao', 'Yueting Zhuang']
2,025
arXiv.org
0
58
['Computer Science']
2,505.15809
MMaDA: Multimodal Large Diffusion Language Models
['Ling Yang', 'Ye Tian', 'Bowen Li', 'Xinchen Zhang', 'Ke Shen', 'Yunhai Tong', 'Mengdi Wang']
['cs.CV']
We introduce MMaDA, a novel class of multimodal diffusion foundation models designed to achieve superior performance across diverse domains such as textual reasoning, multimodal understanding, and text-to-image generation. The approach is distinguished by three key innovations: (i) MMaDA adopts a unified diffusion architecture with a shared probabilistic formulation and a modality-agnostic design, eliminating the need for modality-specific components. This architecture ensures seamless integration and processing across different data types. (ii) We implement a mixed long chain-of-thought (CoT) fine-tuning strategy that curates a unified CoT format across modalities. By aligning reasoning processes between textual and visual domains, this strategy facilitates cold-start training for the final reinforcement learning (RL) stage, thereby enhancing the model's ability to handle complex tasks from the outset. (iii) We propose UniGRPO, a unified policy-gradient-based RL algorithm specifically tailored for diffusion foundation models. Utilizing diversified reward modeling, UniGRPO unifies post-training across both reasoning and generation tasks, ensuring consistent performance improvements. Experimental results demonstrate that MMaDA-8B exhibits strong generalization capabilities as a unified multimodal foundation model. It surpasses powerful models like LLaMA-3-7B and Qwen2-7B in textual reasoning, outperforms Show-o and SEED-X in multimodal understanding, and excels over SDXL and Janus in text-to-image generation. These achievements highlight MMaDA's effectiveness in bridging the gap between pretraining and post-training within unified diffusion architectures, providing a comprehensive framework for future research and development. We open-source our code and trained models at: https://github.com/Gen-Verse/MMaDA
2025-05-21T17:59:05Z
Project: https://github.com/Gen-Verse/MMaDA
null
null
MMaDA: Multimodal Large Diffusion Language Models
['Ling Yang', 'Ye Tian', 'Bowen Li', 'Xinchen Zhang', 'Ke Shen', 'Yunhai Tong', 'Mengdi Wang']
2,025
arXiv.org
6
98
['Computer Science']
2,505.1596
Training Step-Level Reasoning Verifiers with Formal Verification Tools
['Ryo Kamoi', 'Yusen Zhang', 'Nan Zhang', 'Sarkar Snigdha Sarathi Das', 'Rui Zhang']
['cs.CL']
Process Reward Models (PRMs), which provide step-by-step feedback on the reasoning generated by Large Language Models (LLMs), are receiving increasing attention. However, two key research gaps remain: collecting accurate step-level error labels for training typically requires costly human annotation, and existing PRMs are limited to math reasoning problems. In response to these gaps, this paper aims to address the challenges of automatic dataset creation and the generalization of PRMs to diverse reasoning tasks. To achieve this goal, we propose FoVer, an approach for training PRMs on step-level error labels automatically annotated by formal verification tools, such as Z3 for formal logic and Isabelle for theorem proof, which provide automatic and accurate verification for symbolic tasks. Using this approach, we synthesize a training dataset with error labels on LLM responses for formal logic and theorem proof tasks without human annotation. Although this data synthesis is feasible only for tasks compatible with formal verification, we observe that LLM-based PRMs trained on our dataset exhibit cross-task generalization, improving verification across diverse reasoning tasks. Specifically, PRMs trained with FoVer significantly outperform baseline PRMs based on the original LLMs and achieve competitive or superior results compared to state-of-the-art PRMs trained on labels annotated by humans or stronger models, as measured by step-level verification on ProcessBench and Best-of-K performance across 12 reasoning benchmarks, including MATH, AIME, ANLI, MMLU, and BBH. The datasets, models, and code are provided at https://github.com/psunlpgroup/FoVer.
2025-05-21T19:23:45Z
Datasets, models, and code are provided at https://github.com/psunlpgroup/FoVer. Please also refer to our project website at https://fover-prm.github.io/
null
null
Training Step-Level Reasoning Verifiers with Formal Verification Tools
['Ryo Kamoi', 'Yusen Zhang', 'Nan Zhang', 'Sarkar Snigdha Sarathi Das', 'Rui Zhang']
2,025
arXiv.org
0
65
['Computer Science']
2,505.15966
Pixel Reasoner: Incentivizing Pixel-Space Reasoning with Curiosity-Driven Reinforcement Learning
['Alex Su', 'Haozhe Wang', 'Weiming Ren', 'Fangzhen Lin', 'Wenhu Chen']
['cs.CV', 'cs.AI', 'cs.CL']
Chain-of-thought reasoning has significantly improved the performance of Large Language Models (LLMs) across various domains. However, this reasoning process has been confined exclusively to textual space, limiting its effectiveness in visually intensive tasks. To address this limitation, we introduce the concept of reasoning in the pixel-space. Within this novel framework, Vision-Language Models (VLMs) are equipped with a suite of visual reasoning operations, such as zoom-in and select-frame. These operations enable VLMs to directly inspect, interrogate, and infer from visual evidences, thereby enhancing reasoning fidelity for visual tasks. Cultivating such pixel-space reasoning capabilities in VLMs presents notable challenges, including the model's initially imbalanced competence and its reluctance to adopt the newly introduced pixel-space operations. We address these challenges through a two-phase training approach. The first phase employs instruction tuning on synthesized reasoning traces to familiarize the model with the novel visual operations. Following this, a reinforcement learning (RL) phase leverages a curiosity-driven reward scheme to balance exploration between pixel-space reasoning and textual reasoning. With these visual operations, VLMs can interact with complex visual inputs, such as information-rich images or videos to proactively gather necessary information. We demonstrate that this approach significantly improves VLM performance across diverse visual reasoning benchmarks. Our 7B model, \model, achieves 84\% on V* bench, 74\% on TallyQA-Complex, and 84\% on InfographicsVQA, marking the highest accuracy achieved by any open-source model to date. These results highlight the importance of pixel-space reasoning and the effectiveness of our framework.
2025-05-21T19:35:08Z
Project Page: https://tiger-ai-lab.github.io/Pixel-Reasoner/, Hands-on Demo: https://huggingface.co/spaces/TIGER-Lab/Pixel-Reasoner
null
null
null
null
null
null
null
null
null
2,505.16
Leveraging Online Data to Enhance Medical Knowledge in a Small Persian Language Model
['Mehrdad Ghassabi', 'Pedram Rostami', 'Hamidreza Baradaran Kashani', 'Amirhossein Poursina', 'Zahra Kazemi', 'Milad Tavakoli']
['cs.CL', 'cs.AI']
The rapid advancement of language models has demonstrated the potential of artificial intelligence in the healthcare industry. However, small language models struggle with specialized domains in low-resource languages like Persian. While numerous medical-domain websites exist in Persian, no curated dataset or corpus has been available making ours the first of its kind. This study explores the enhancement of medical knowledge in a small language model by leveraging accessible online data, including a crawled corpus from medical magazines and a dataset of real doctor-patient QA pairs. We fine-tuned a baseline model using our curated data to improve its medical knowledge. Benchmark evaluations demonstrate that the fine-tuned model achieves improved accuracy in medical question answering and provides better responses compared to its baseline. This work highlights the potential of leveraging open-access online data to enrich small language models in medical fields, providing a novel solution for Persian medical AI applications suitable for resource-constrained environments.
2025-05-21T20:30:47Z
6 pages, 4 figures
null
null
null
null
null
null
null
null
null
2,505.1616
EduBench: A Comprehensive Benchmarking Dataset for Evaluating Large Language Models in Diverse Educational Scenarios
['Bin Xu', 'Yu Bai', 'Huashan Sun', 'Yiguan Lin', 'Siming Liu', 'Xinyue Liang', 'Yaolin Li', 'Yang Gao', 'Heyan Huang']
['cs.CL']
As large language models continue to advance, their application in educational contexts remains underexplored and under-optimized. In this paper, we address this gap by introducing the first diverse benchmark tailored for educational scenarios, incorporating synthetic data containing 9 major scenarios and over 4,000 distinct educational contexts. To enable comprehensive assessment, we propose a set of multi-dimensional evaluation metrics that cover 12 critical aspects relevant to both teachers and students. We further apply human annotation to ensure the effectiveness of the model-generated evaluation responses. Additionally, we succeed to train a relatively small-scale model on our constructed dataset and demonstrate that it can achieve performance comparable to state-of-the-art large models (e.g., Deepseek V3, Qwen Max) on the test set. Overall, this work provides a practical foundation for the development and evaluation of education-oriented language models. Code and data are released at https://github.com/ybai-nlp/EduBench.
2025-05-22T03:01:28Z
null
null
null
null
null
null
null
null
null
null
2,505.16186
SafeKey: Amplifying Aha-Moment Insights for Safety Reasoning
['Kaiwen Zhou', 'Xuandong Zhao', 'Gaowen Liu', 'Jayanth Srinivasa', 'Aosong Feng', 'Dawn Song', 'Xin Eric Wang']
['cs.AI', 'cs.CL', 'cs.CR']
Large Reasoning Models (LRMs) introduce a new generation paradigm of explicitly reasoning before answering, leading to remarkable improvements in complex tasks. However, they pose great safety risks against harmful queries and adversarial attacks. While recent mainstream safety efforts on LRMs, supervised fine-tuning (SFT), improve safety performance, we find that SFT-aligned models struggle to generalize to unseen jailbreak prompts. After thorough investigation of LRMs' generation, we identify a safety aha moment that can activate safety reasoning and lead to a safe response. This aha moment typically appears in the `key sentence', which follows models' query understanding process and can indicate whether the model will proceed safely. Based on these insights, we propose SafeKey, including two complementary objectives to better activate the safety aha moment in the key sentence: (1) a Dual-Path Safety Head to enhance the safety signal in the model's internal representations before the key sentence, and (2) a Query-Mask Modeling objective to improve the models' attention on its query understanding, which has important safety hints. Experiments across multiple safety benchmarks demonstrate that our methods significantly improve safety generalization to a wide range of jailbreak attacks and out-of-distribution harmful prompts, lowering the average harmfulness rate by 9.6\%, while maintaining general abilities. Our analysis reveals how SafeKey enhances safety by reshaping internal attention and improving the quality of hidden representations.
2025-05-22T03:46:03Z
null
null
null
SafeKey: Amplifying Aha-Moment Insights for Safety Reasoning
['KAI-QING Zhou', 'Xuandong Zhao', 'Gaowen Liu', 'Jayanth Srinivasa', 'Aosong Feng', 'D. Song', 'Xin Eric Wang']
2,025
arXiv.org
0
35
['Computer Science']
2,505.16239
DOVE: Efficient One-Step Diffusion Model for Real-World Video Super-Resolution
['Zheng Chen', 'Zichen Zou', 'Kewei Zhang', 'Xiongfei Su', 'Xin Yuan', 'Yong Guo', 'Yulun Zhang']
['cs.CV']
Diffusion models have demonstrated promising performance in real-world video super-resolution (VSR). However, the dozens of sampling steps they require, make inference extremely slow. Sampling acceleration techniques, particularly single-step, provide a potential solution. Nonetheless, achieving one step in VSR remains challenging, due to the high training overhead on video data and stringent fidelity demands. To tackle the above issues, we propose DOVE, an efficient one-step diffusion model for real-world VSR. DOVE is obtained by fine-tuning a pretrained video diffusion model (*i.e.*, CogVideoX). To effectively train DOVE, we introduce the latent-pixel training strategy. The strategy employs a two-stage scheme to gradually adapt the model to the video super-resolution task. Meanwhile, we design a video processing pipeline to construct a high-quality dataset tailored for VSR, termed HQ-VSR. Fine-tuning on this dataset further enhances the restoration capability of DOVE. Extensive experiments show that DOVE exhibits comparable or superior performance to multi-step diffusion-based VSR methods. It also offers outstanding inference efficiency, achieving up to a **28$\times$** speed-up over existing methods such as MGLD-VSR. Code is available at: https://github.com/zhengchen1999/DOVE.
2025-05-22T05:16:45Z
Code is available at: https://github.com/zhengchen1999/DOVE
null
null
null
null
null
null
null
null
null
2,505.16368
SATURN: SAT-based Reinforcement Learning to Unleash Language Model Reasoning
['Huanyu Liu', 'Jia Li', 'Hao Zhu', 'Kechi Zhang', 'Yihong Dong', 'Ge Li']
['cs.LG', 'cs.AI']
How to design reinforcement learning (RL) tasks that effectively unleash the reasoning capability of large language models (LLMs) remains an open question. Existing RL tasks (e.g., math, programming, and constructing reasoning tasks) suffer from three key limitations: (1) Scalability. They rely heavily on human annotation or expensive LLM synthesis to generate sufficient training data. (2) Verifiability. LLMs' outputs are hard to verify automatically and reliably. (3) Controllable Difficulty. Most tasks lack fine-grained difficulty control, making it hard to train LLMs to develop reasoning ability from easy to hard. To address these limitations, we propose Saturn, a SAT-based RL framework that uses Boolean Satisfiability (SAT) problems to train and evaluate LLM reasoning. Saturn enables scalable task construction, rule-based verification, and precise difficulty control. Saturn designs a curriculum learning pipeline that continuously improves LLMs' reasoning capability by constructing SAT tasks of increasing difficulty and training LLMs from easy to hard. To ensure stable training, we design a principled mechanism to control difficulty transitions. We introduce Saturn-2.6k, a dataset of 2,660 SAT problems with varying difficulty. It supports the evaluation of how LLM reasoning changes with problem difficulty. We apply Saturn to DeepSeek-R1-Distill-Qwen and obtain Saturn-1.5B and Saturn-7B. We achieve several notable results: (1) On SAT problems, Saturn-1.5B and Saturn-7B achieve average pass@3 improvements of +14.0 and +28.1, respectively. (2) On math and programming tasks, Saturn-1.5B and Saturn-7B improve average scores by +4.9 and +1.8 on benchmarks (e.g., AIME, LiveCodeBench). (3) Compared to the state-of-the-art (SOTA) approach in constructing RL tasks, Saturn achieves further improvements of +8.8%. We release the source code, data, and models to support future research.
2025-05-22T08:23:10Z
null
null
null
null
null
null
null
null
null
null
2,505.164
AceReason-Nemotron: Advancing Math and Code Reasoning through Reinforcement Learning
['Yang Chen', 'Zhuolin Yang', 'Zihan Liu', 'Chankyu Lee', 'Peng Xu', 'Mohammad Shoeybi', 'Bryan Catanzaro', 'Wei Ping']
['cs.LG', 'cs.AI', 'cs.CL']
Despite recent progress in large-scale reinforcement learning (RL) for reasoning, the training recipe for building high-performing reasoning models remains elusive. Key implementation details of frontier models, such as DeepSeek-R1, including data curation strategies and RL training recipe, are often omitted. Moreover, recent research indicates distillation remains more effective than RL for smaller models. In this work, we demonstrate that large-scale RL can significantly enhance the reasoning capabilities of strong, small- and mid-sized models, achieving results that surpass those of state-of-the-art distillation-based models. We systematically study the RL training process through extensive ablations and propose a simple yet effective approach: first training on math-only prompts, then on code-only prompts. Notably, we find that math-only RL not only significantly enhances the performance of strong distilled models on math benchmarks (e.g., +14.6% / +17.2% on AIME 2025 for the 7B / 14B models), but also code reasoning tasks (e.g., +6.8% / +5.8% on LiveCodeBench for the 7B / 14B models). In addition, extended code-only RL iterations further improve performance on code benchmarks with minimal or no degradation in math results. We develop a robust data curation pipeline to collect challenging prompts with high-quality, verifiable answers and test cases to enable verification-based RL across both domains. Finally, we identify key experimental insights, including curriculum learning with progressively increasing response lengths and the stabilizing effect of on-policy parameter updates. We find that RL not only elicits the foundational reasoning capabilities acquired during pretraining and supervised fine-tuning (e.g., distillation), but also pushes the limits of the model's reasoning ability, enabling it to solve problems that were previously unsolvable.
2025-05-22T08:50:47Z
Add pass@1024 evaluation results for LiveCodeBench v6. We release the models at: https://huggingface.co/collections/nvidia/acereason-682f4e1261dc22f697fd1485
null
null
null
null
null
null
null
null
null
2,505.1641
Tool-Star: Empowering LLM-Brained Multi-Tool Reasoner via Reinforcement Learning
['Guanting Dong', 'Yifei Chen', 'Xiaoxi Li', 'Jiajie Jin', 'Hongjin Qian', 'Yutao Zhu', 'Hangyu Mao', 'Guorui Zhou', 'Zhicheng Dou', 'Ji-Rong Wen']
['cs.CL', 'cs.AI', 'cs.LG']
Recently, large language models (LLMs) have shown remarkable reasoning capabilities via large-scale reinforcement learning (RL). However, leveraging the RL algorithm to empower effective multi-tool collaborative reasoning in LLMs remains an open challenge. In this paper, we introduce Tool-Star, an RL-based framework designed to empower LLMs to autonomously invoke multiple external tools during stepwise reasoning. Tool-Star integrates six types of tools and incorporates systematic designs in both data synthesis and training. To address the scarcity of tool-use data, we propose a general tool-integrated reasoning data synthesis pipeline, which combines tool-integrated prompting with hint-based sampling to automatically and scalably generate tool-use trajectories. A subsequent quality normalization and difficulty-aware classification process filters out low-quality samples and organizes the dataset from easy to hard. Furthermore, we propose a two-stage training framework to enhance multi-tool collaborative reasoning by: (1) cold-start fine-tuning, which guides LLMs to explore reasoning patterns via tool-invocation feedback; and (2) a multi-tool self-critic RL algorithm with hierarchical reward design, which reinforces reward understanding and promotes effective tool collaboration. Experimental analyses on over 10 challenging reasoning benchmarks highlight the effectiveness and efficiency of Tool-Star. The code is available at https://github.com/dongguanting/Tool-Star.
2025-05-22T09:00:19Z
Working in progress
null
null
null
null
null
null
null
null
null
2,505.16495
ALTo: Adaptive-Length Tokenizer for Autoregressive Mask Generation
['Lingfeng Wang', 'Hualing Lin', 'Senda Chen', 'Tao Wang', 'Changxu Cheng', 'Yangyang Zhong', 'Dong Zheng', 'Wuyue Zhao']
['cs.CV']
While humans effortlessly draw visual objects and shapes by adaptively allocating attention based on their complexity, existing multimodal large language models (MLLMs) remain constrained by rigid token representations. Bridging this gap, we propose ALTo, an adaptive length tokenizer for autoregressive mask generation. To achieve this, a novel token length predictor is designed, along with a length regularization term and a differentiable token chunking strategy. We further build ALToLLM that seamlessly integrates ALTo into MLLM. Preferences on the trade-offs between mask quality and efficiency is implemented by group relative policy optimization (GRPO). Experiments demonstrate that ALToLLM achieves state-of-the-art performance with adaptive token cost on popular segmentation benchmarks. Code and models are released at https://github.com/yayafengzi/ALToLLM.
2025-05-22T10:26:51Z
null
null
null
ALTo: Adaptive-Length Tokenizer for Autoregressive Mask Generation
['Lingfeng Wang', 'Hualing Lin', 'Senda Chen', 'Tao Wang', 'Changxu Cheng', 'Yangyang Zhong', 'Dong Zheng', 'Wuyue Zhao']
2,025
arXiv.org
0
64
['Computer Science']
2,505.16637
SSR-Zero: Simple Self-Rewarding Reinforcement Learning for Machine Translation
['Wenjie Yang', 'Mao Zheng', 'Mingyang Song', 'Zheng Li', 'Sitong Wang']
['cs.CL', 'cs.AI', 'cs.LG']
Large language models (LLMs) have recently demonstrated remarkable capabilities in machine translation (MT). However, most advanced MT-specific LLMs heavily rely on external supervision signals during training, such as human-annotated reference data or trained reward models (RMs), which are often expensive to obtain and challenging to scale. To overcome this limitation, we propose a Simple Self-Rewarding (SSR) Reinforcement Learning (RL) framework for MT that is reference-free, fully online, and relies solely on self-judging rewards. Training with SSR using 13K monolingual examples and Qwen-2.5-7B as the backbone, our model SSR-Zero-7B outperforms existing MT-specific LLMs, e.g., TowerInstruct-13B and GemmaX-28-9B, as well as larger general LLMs like Qwen2.5-32B-Instruct in English $\leftrightarrow$ Chinese translation tasks from WMT23, WMT24, and Flores200 benchmarks. Furthermore, by augmenting SSR with external supervision from COMET, our strongest model, SSR-X-Zero-7B, achieves state-of-the-art performance in English $\leftrightarrow$ Chinese translation, surpassing all existing open-source models under 72B parameters and even outperforming closed-source models, e.g., GPT-4o and Gemini 1.5 Pro. Our analysis highlights the effectiveness of the self-rewarding mechanism compared to the external LLM-as-a-judge approach in MT and demonstrates its complementary benefits when combined with trained RMs. Our findings provide valuable insight into the potential of self-improving RL methods. We have publicly released our code, data and models.
2025-05-22T13:08:25Z
null
null
null
null
null
null
null
null
null
null
2,505.16647
Point, Detect, Count: Multi-Task Medical Image Understanding with Instruction-Tuned Vision-Language Models
['Sushant Gautam', 'Michael A. Riegler', 'Pål Halvorsen']
['cs.CV', 'cs.AI', '68T45, 68T07', 'I.2.10; I.4.8']
We investigate fine-tuning Vision-Language Models (VLMs) for multi-task medical image understanding, focusing on detection, localization, and counting of findings in medical images. Our objective is to evaluate whether instruction-tuned VLMs can simultaneously improve these tasks, with the goal of enhancing diagnostic accuracy and efficiency. Using MedMultiPoints, a multimodal dataset with annotations from endoscopy (polyps and instruments) and microscopy (sperm cells), we reformulate each task into instruction-based prompts suitable for vision-language reasoning. We fine-tune Qwen2.5-VL-7B-Instruct using Low-Rank Adaptation (LoRA) across multiple task combinations. Results show that multi-task training improves robustness and accuracy. For example, it reduces the Count Mean Absolute Error (MAE) and increases Matching Accuracy in the Counting + Pointing task. However, trade-offs emerge, such as more zero-case point predictions, indicating reduced reliability in edge cases despite overall performance gains. Our study highlights the potential of adapting general-purpose VLMs to specialized medical tasks via prompt-driven fine-tuning. This approach mirrors clinical workflows, where radiologists simultaneously localize, count, and describe findings - demonstrating how VLMs can learn composite diagnostic reasoning patterns. The model produces interpretable, structured outputs, offering a promising step toward explainable and versatile medical AI. Code, model weights, and scripts will be released for reproducibility at https://github.com/simula/PointDetectCount.
2025-05-22T13:18:44Z
Accepted as a full paper at the 38th IEEE International Symposium on Computer-Based Medical Systems (CBMS) 2025
null
null
null
null
null
null
null
null
null
2,505.16661
A Japanese Language Model and Three New Evaluation Benchmarks for Pharmaceutical NLP
['Issey Sukeda', 'Takuro Fujii', 'Kosei Buma', 'Shunsuke Sasaki', 'Shinnosuke Ono']
['cs.CL']
We present a Japanese domain-specific language model for the pharmaceutical field, developed through continual pretraining on 2 billion Japanese pharmaceutical tokens and 8 billion English biomedical tokens. To enable rigorous evaluation, we introduce three new benchmarks: YakugakuQA, based on national pharmacist licensing exams; NayoseQA, which tests cross-lingual synonym and terminology normalization; and SogoCheck, a novel task designed to assess consistency reasoning between paired statements. We evaluate our model against both open-source medical LLMs and commercial models, including GPT-4o. Results show that our domain-specific model outperforms existing open models and achieves competitive performance with commercial ones, particularly on terminology-heavy and knowledge-based tasks. Interestingly, even GPT-4o performs poorly on SogoCheck, suggesting that cross-sentence consistency reasoning remains an open challenge. Our benchmark suite offers a broader diagnostic lens for pharmaceutical NLP, covering factual recall, lexical variation, and logical consistency. This work demonstrates the feasibility of building practical, secure, and cost-effective language models for Japanese domain-specific applications, and provides reusable evaluation resources for future research in pharmaceutical and healthcare NLP. Our model, codes, and datasets are released at https://github.com/EQUES-Inc/pharma-LLM-eval.
2025-05-22T13:27:37Z
15 pages, 9 tables, 5 figures
null
null
null
null
null
null
null
null
null
2,505.16826
KTAE: A Model-Free Algorithm to Key-Tokens Advantage Estimation in Mathematical Reasoning
['Wei Sun', 'Wen Yang', 'Pu Jian', 'Qianlong Du', 'Fuwei Cui', 'Shuo Ren', 'Jiajun Zhang']
['cs.AI', 'cs.CL']
Recent advances have demonstrated that integrating reinforcement learning with rule-based rewards can significantly enhance the reasoning capabilities of large language models, even without supervised fine-tuning. However, prevalent reinforcement learning algorithms such as GRPO and its variants like DAPO, suffer from a coarse granularity issue when computing the advantage. Specifically, they compute rollout-level advantages that assign identical values to every token within a sequence, failing to capture token-specific contributions and hindering effective learning. To address this limitation, we propose Key-token Advantage Estimation (KTAE) - a novel algorithm that estimates fine-grained, token-level advantages without introducing additional models. KTAE leverages the correctness of sampled rollouts and applies statistical analysis to quantify the importance of individual tokens within a sequence to the final outcome. This quantified token-level importance is then combined with the rollout-level advantage to obtain a more fine-grained token-level advantage estimation. Empirical results show that models trained with GRPO+KTAE and DAPO+KTAE outperform baseline methods across five mathematical reasoning benchmarks. Notably, they achieve higher accuracy with shorter responses and even surpass R1-Distill-Qwen-1.5B using the same base model.
2025-05-22T16:00:33Z
null
null
null
null
null
null
null
null
null
null
2,505.16839
LaViDa: A Large Diffusion Language Model for Multimodal Understanding
['Shufan Li', 'Konstantinos Kallidromitis', 'Hritik Bansal', 'Akash Gokul', 'Yusuke Kato', 'Kazuki Kozuka', 'Jason Kuen', 'Zhe Lin', 'Kai-Wei Chang', 'Aditya Grover']
['cs.CV']
Modern Vision-Language Models (VLMs) can solve a wide range of tasks requiring visual reasoning. In real-world scenarios, desirable properties for VLMs include fast inference and controllable generation (e.g., constraining outputs to adhere to a desired format). However, existing autoregressive (AR) VLMs like LLaVA struggle in these aspects. Discrete diffusion models (DMs) offer a promising alternative, enabling parallel decoding for faster inference and bidirectional context for controllable generation through text-infilling. While effective in language-only settings, DMs' potential for multimodal tasks is underexplored. We introduce LaViDa, a family of VLMs built on DMs. We build LaViDa by equipping DMs with a vision encoder and jointly fine-tune the combined parts for multimodal instruction following. To address challenges encountered, LaViDa incorporates novel techniques such as complementary masking for effective training, prefix KV cache for efficient inference, and timestep shifting for high-quality sampling. Experiments show that LaViDa achieves competitive or superior performance to AR VLMs on multi-modal benchmarks such as MMMU, while offering unique advantages of DMs, including flexible speed-quality tradeoff, controllability, and bidirectional reasoning. On COCO captioning, LaViDa surpasses Open-LLaVa-Next-8B by +4.1 CIDEr with 1.92x speedup. On bidirectional tasks, it achieves +59% improvement on Constrained Poem Completion. These results demonstrate LaViDa as a strong alternative to AR VLMs. Code and models will be released in the camera-ready version.
2025-05-22T16:07:12Z
26 pages, 8 figures
null
null
null
null
null
null
null
null
null
2,505.16854
Think or Not? Selective Reasoning via Reinforcement Learning for Vision-Language Models
['Jiaqi Wang', 'Kevin Qinghong Lin', 'James Cheng', 'Mike Zheng Shou']
['cs.AI', 'cs.CV']
Reinforcement Learning (RL) has proven to be an effective post-training strategy for enhancing reasoning in vision-language models (VLMs). Group Relative Policy Optimization (GRPO) is a recent prominent method that encourages models to generate complete reasoning traces before answering, leading to increased token usage and computational cost. Inspired by the human-like thinking process-where people skip reasoning for easy questions but think carefully when needed-we explore how to enable VLMs to first decide when reasoning is necessary. To realize this, we propose TON, a two-stage training strategy: (i) a supervised fine-tuning (SFT) stage with a simple yet effective 'thought dropout' operation, where reasoning traces are randomly replaced with empty thoughts. This introduces a think-or-not format that serves as a cold start for selective reasoning; (ii) a GRPO stage that enables the model to freely explore when to think or not, while maximizing task-aware outcome rewards. Experimental results show that TON can reduce the completion length by up to 90% compared to vanilla GRPO, without sacrificing performance or even improving it. Further evaluations across diverse vision-language tasks-covering a range of reasoning difficulties under both 3B and 7B models-consistently reveal that the model progressively learns to bypass unnecessary reasoning steps as training advances. These findings shed light on the path toward human-like reasoning patterns in reinforcement learning approaches. Our code is available at https://github.com/kokolerk/TON.
2025-05-22T16:13:29Z
update more examples in appendix
null
null
Think or Not? Selective Reasoning via Reinforcement Learning for Vision-Language Models
['Jiaqi Wang', 'Kevin Qinghong Lin', 'James Cheng', 'Mike Zheng Shou']
2,025
arXiv.org
0
37
['Computer Science']
2,505.16901
Code Graph Model (CGM): A Graph-Integrated Large Language Model for Repository-Level Software Engineering Tasks
['Hongyuan Tao', 'Ying Zhang', 'Zhenhao Tang', 'Hongen Peng', 'Xukun Zhu', 'Bingchang Liu', 'Yingguang Yang', 'Ziyin Zhang', 'Zhaogui Xu', 'Haipeng Zhang', 'Linchao Zhu', 'Rui Wang', 'Hang Yu', 'Jianguo Li', 'Peng Di']
['cs.SE', 'cs.LG']
Recent advances in Large Language Models (LLMs) have shown promise in function-level code generation, yet repository-level software engineering tasks remain challenging. Current solutions predominantly rely on proprietary LLM agents, which introduce unpredictability and limit accessibility, raising concerns about data privacy and model customization. This paper investigates whether open-source LLMs can effectively address repository-level tasks without requiring agent-based approaches. We demonstrate this is possible by enabling LLMs to comprehend functions and files within codebases through their semantic information and structural dependencies. To this end, we introduce Code Graph Models (CGMs), which integrate repository code graph structures into the LLM's attention mechanism and map node attributes to the LLM's input space using a specialized adapter. When combined with an agentless graph RAG framework, our approach achieves a 43.00% resolution rate on the SWE-bench Lite benchmark using the open-source Qwen2.5-72B model. This performance ranks first among open weight models, second among methods with open-source systems, and eighth overall, surpassing the previous best open-source model-based method by 12.33%.
2025-05-22T17:00:55Z
35 pages, 10 figures
null
null
null
null
null
null
null
null
null
2,505.16933
LLaDA-V: Large Language Diffusion Models with Visual Instruction Tuning
['Zebin You', 'Shen Nie', 'Xiaolu Zhang', 'Jun Hu', 'Jun Zhou', 'Zhiwu Lu', 'Ji-Rong Wen', 'Chongxuan Li']
['cs.LG', 'cs.CL', 'cs.CV']
In this work, we introduce LLaDA-V, a purely diffusion-based Multimodal Large Language Model (MLLM) that integrates visual instruction tuning with masked diffusion models, representing a departure from the autoregressive paradigms dominant in current multimodal approaches. Built upon LLaDA, a representative large language diffusion model, LLaDA-V incorporates a vision encoder and MLP connector that projects visual features into the language embedding space, enabling effective multimodal alignment. Our empirical investigation reveals several intriguing results: First, LLaDA-V demonstrates promising multimodal performance despite its language model being weaker on purely textual tasks than counterparts like LLaMA3-8B and Qwen2-7B. When trained on the same instruction data, LLaDA-V is highly competitive to LLaMA3-V across multimodal tasks with better data scalability. It also narrows the performance gap to Qwen2-VL, suggesting the effectiveness of its architecture for multimodal tasks. Second, LLaDA-V achieves state-of-the-art performance in multimodal understanding compared to existing hybrid autoregressive-diffusion and purely diffusion-based MLLMs. Our findings suggest that large language diffusion models show promise in multimodal contexts and warrant further investigation in future research. Project page and codes: https://ml-gsai.github.io/LLaDA-V-demo/.
2025-05-22T17:23:26Z
Project page and codes: \url{https://ml-gsai.github.io/LLaDA-V-demo/}
null
null
LLaDA-V: Large Language Diffusion Models with Visual Instruction Tuning
['Zebin You', 'Shen Nie', 'Xiaolu Zhang', 'Jun Hu', 'Jun Zhou', 'Zhiwu Lu', 'Ji-Rong Wen', 'Chongxuan Li']
2,025
arXiv.org
2
114
['Computer Science']
2,505.16938
NovelSeek: When Agent Becomes the Scientist -- Building Closed-Loop System from Hypothesis to Verification
['NovelSeek Team', 'Bo Zhang', 'Shiyang Feng', 'Xiangchao Yan', 'Jiakang Yuan', 'Zhiyin Yu', 'Xiaohan He', 'Songtao Huang', 'Shaowei Hou', 'Zheng Nie', 'Zhilong Wang', 'Jinyao Liu', 'Runmin Ma', 'Tianshuo Peng', 'Peng Ye', 'Dongzhan Zhou', 'Shufei Zhang', 'Xiaosong Wang', 'Yilan Zhang', 'Meng Li', 'Zhongying Tu', 'Xiangyu Yue', 'Wangli Ouyang', 'Bowen Zhou', 'Lei Bai']
['cs.AI', 'cs.CL', 'cs.CV']
Artificial Intelligence (AI) is accelerating the transformation of scientific research paradigms, not only enhancing research efficiency but also driving innovation. We introduce NovelSeek, a unified closed-loop multi-agent framework to conduct Autonomous Scientific Research (ASR) across various scientific research fields, enabling researchers to tackle complicated problems in these fields with unprecedented speed and precision. NovelSeek highlights three key advantages: 1) Scalability: NovelSeek has demonstrated its versatility across 12 scientific research tasks, capable of generating innovative ideas to enhance the performance of baseline code. 2) Interactivity: NovelSeek provides an interface for human expert feedback and multi-agent interaction in automated end-to-end processes, allowing for the seamless integration of domain expert knowledge. 3) Efficiency: NovelSeek has achieved promising performance gains in several scientific fields with significantly less time cost compared to human efforts. For instance, in reaction yield prediction, it increased from 27.6% to 35.4% in just 12 hours; in enhancer activity prediction, accuracy rose from 0.65 to 0.79 with only 4 hours of processing; and in 2D semantic segmentation, precision advanced from 78.8% to 81.0% in a mere 30 hours.
2025-05-22T17:27:43Z
HomePage: https://alpha-innovator.github.io/NovelSeek-project-page
null
null
NovelSeek: When Agent Becomes the Scientist - Building Closed-Loop System from Hypothesis to Verification
['NovelSeek Team Bo Zhang', 'Shi Feng', 'Xiangchao Yan', 'Jiakang Yuan', 'Zhiyin Yu', 'Xiaohan He', 'Songtao Huang', 'Shaowei Hou', 'Zheng Nie', 'Zhilong Wang', 'Jinyao Liu', 'Runmin Ma', 'Tianshuo Peng', 'Peng Ye', 'Dongzhan Zhou', 'Shufei Zhang', 'Xiaosong Wang', 'Yilan Zhang', 'Meng Li', 'Zhongying Tu', 'Xiangyu Yue', 'Wangli Ouyang', 'Bowen Zhou', 'Lei Bai']
2,025
arXiv.org
2
51
['Computer Science']
2,505.16947
MixAT: Combining Continuous and Discrete Adversarial Training for LLMs
['Csaba Dékány', 'Stefan Balauca', 'Robin Staab', 'Dimitar I. Dimitrov', 'Martin Vechev']
['cs.LG', 'cs.AI', 'I.2.7; K.4.1']
Despite recent efforts in Large Language Models (LLMs) safety and alignment, current adversarial attacks on frontier LLMs are still able to force harmful generations consistently. Although adversarial training has been widely studied and shown to significantly improve the robustness of traditional machine learning models, its strengths and weaknesses in the context of LLMs are less understood. Specifically, while existing discrete adversarial attacks are effective at producing harmful content, training LLMs with concrete adversarial prompts is often computationally expensive, leading to reliance on continuous relaxations. As these relaxations do not correspond to discrete input tokens, such latent training methods often leave models vulnerable to a diverse set of discrete attacks. In this work, we aim to bridge this gap by introducing MixAT, a novel method that combines stronger discrete and faster continuous attacks during training. We rigorously evaluate MixAT across a wide spectrum of state-of-the-art attacks, proposing the At Least One Attack Success Rate (ALO-ASR) metric to capture the worst-case vulnerability of models. We show MixAT achieves substantially better robustness (ALO-ASR < 20%) compared to prior defenses (ALO-ASR > 50%), while maintaining a runtime comparable to methods based on continuous relaxations. We further analyze MixAT in realistic deployment settings, exploring how chat templates, quantization, low-rank adapters, and temperature affect both adversarial training and evaluation, revealing additional blind spots in current methodologies. Our results demonstrate that MixAT's discrete-continuous defense offers a principled and superior robustness-accuracy tradeoff with minimal computational overhead, highlighting its promise for building safer LLMs. We provide our code and models at https://github.com/insait-institute/MixAT.
2025-05-22T17:32:50Z
null
null
null
MixAT: Combining Continuous and Discrete Adversarial Training for LLMs
["Csaba D'ek'any", 'Stefan Balauca', 'Robin Staab', 'Dimitar I. Dimitrov', 'Martin T. Vechev']
2,025
arXiv.org
0
41
['Computer Science']
2,505.16968
CASS: Nvidia to AMD Transpilation with Data, Models, and Benchmark
['Ahmed Heakl', 'Sarim Hashmi', 'Gustavo Bertolo Stahl', 'Seung Hun Eddie Han', 'Salman Khan', 'Abdulrahman Mahmoud']
['cs.AR', 'cs.AI', 'cs.CL', 'cs.LG', 'cs.PL']
We introduce CASS, the first large-scale dataset and model suite for cross-architecture GPU code transpilation, targeting both source-level (CUDA <--> HIP) and assembly-level (Nvidia SASS <--> AMD RDNA3) translation. The dataset comprises 70k verified code pairs across host and device, addressing a critical gap in low-level GPU code portability. Leveraging this resource, we train the CASS family of domain-specific language models, achieving 95% source translation accuracy and 37.5% assembly translation accuracy, substantially outperforming commercial baselines such as GPT-4o, Claude, and Hipify. Our generated code matches native performance in over 85% of test cases, preserving runtime and memory behavior. To support rigorous evaluation, we introduce CASS-Bench, a curated benchmark spanning 16 GPU domains with ground-truth execution. All data, models, and evaluation tools are released as open source to foster progress in GPU compiler tooling, binary compatibility, and LLM-guided hardware translation.
2025-05-22T17:48:53Z
20 pages, 11 figures, 5 tables
null
null
null
null
null
null
null
null
null
2,505.16973
VeriFastScore: Speeding up long-form factuality evaluation
['Rishanth Rajendhran', 'Amir Zadeh', 'Matthew Sarte', 'Chuan Li', 'Mohit Iyyer']
['cs.CL']
Metrics like FactScore and VeriScore that evaluate long-form factuality operate by decomposing an input response into atomic claims and then individually verifying each claim. While effective and interpretable, these methods incur numerous LLM calls and can take upwards of 100 seconds to evaluate a single response, limiting their practicality in large-scale evaluation and training scenarios. To address this, we propose VeriFastScore, which leverages synthetic data to fine-tune Llama3.1 8B for simultaneously extracting and verifying all verifiable claims within a given text based on evidence from Google Search. We show that this task cannot be solved via few-shot prompting with closed LLMs due to its complexity: the model receives ~4K tokens of evidence on average and needs to concurrently decompose claims, judge their verifiability, and verify them against noisy evidence. However, our fine-tuned VeriFastScore model demonstrates strong correlation with the original VeriScore pipeline at both the example level (r=0.80) and system level (r=0.94) while achieving an overall speedup of 6.6x (9.9x excluding evidence retrieval) over VeriScore. To facilitate future factuality research, we publicly release our VeriFastScore model and synthetic datasets.
2025-05-22T17:51:25Z
null
null
null
VeriFastScore: Speeding up long-form factuality evaluation
['Rishanth Rajendhran', 'Amir Zadeh', 'Matthew Sarte', 'Chuan Li', 'Mohit Iyyer']
2,025
arXiv.org
0
29
['Computer Science']
2,505.16983
LLM as Effective Streaming Processor: Bridging Streaming-Batch Mismatches with Group Position Encoding
['Junlong Tong', 'Jinlan Fu', 'Zixuan Lin', 'Yingqi Fan', 'Anhao Zhao', 'Hui Su', 'Xiaoyu Shen']
['cs.CL']
Large Language Models (LLMs) are primarily designed for batch processing. Existing methods for adapting LLMs to streaming rely either on expensive re-encoding or specialized architectures with limited scalability. This work identifies three key mismatches in adapting batch-oriented LLMs to streaming: (1) input-attention, (2) output-attention, and (3) position-ID mismatches. While it is commonly assumed that the latter two mismatches require frequent re-encoding, our analysis reveals that only the input-attention mismatch significantly impacts performance, indicating re-encoding outputs is largely unnecessary. To better understand this discrepancy with the common assumption, we provide the first comprehensive analysis of the impact of position encoding on LLMs in streaming, showing that preserving relative positions within source and target contexts is more critical than maintaining absolute order. Motivated by the above analysis, we introduce a group position encoding paradigm built on batch architectures to enhance consistency between streaming and batch modes. Extensive experiments on cross-lingual and cross-modal tasks demonstrate that our method outperforms existing approaches. Our method requires no architectural modifications, exhibits strong generalization in both streaming and batch modes. The code is available at repository https://github.com/EIT-NLP/StreamingLLM.
2025-05-22T17:53:28Z
ACL 2025 Findings
null
null
LLM as Effective Streaming Processor: Bridging Streaming-Batch Mismatches with Group Position Encoding
['Junlong Tong', 'Jinlan Fu', 'Zixuan Lin', 'Yingqi Fan', 'Anhao Zhao', 'Hui Su', 'Xiaoyu Shen']
2,025
arXiv.org
0
49
['Computer Science']
2,505.16984
UFT: Unifying Supervised and Reinforcement Fine-Tuning
['Mingyang Liu', 'Gabriele Farina', 'Asuman Ozdaglar']
['cs.LG', 'cs.CL']
Post-training has demonstrated its importance in enhancing the reasoning capabilities of large language models (LLMs). The primary post-training methods can be categorized into supervised fine-tuning (SFT) and reinforcement fine-tuning (RFT). SFT is efficient and well-suited for small language models, but it may lead to overfitting and limit the reasoning abilities of larger models. In contrast, RFT generally yields better generalization but depends heavily on the strength of the base model. To address the limitations of SFT and RFT, we propose Unified Fine-Tuning (UFT), a novel post-training paradigm that unifies SFT and RFT into a single, integrated process. UFT enables the model to effectively explore solutions while incorporating informative supervision signals, bridging the gap between memorizing and thinking underlying existing methods. Notably, UFT outperforms both SFT and RFT in general, regardless of model sizes. Furthermore, we theoretically prove that UFT breaks RFT's inherent exponential sample complexity bottleneck, showing for the first time that unified training can exponentially accelerate convergence on long-horizon reasoning tasks.
2025-05-22T17:53:57Z
null
null
null
null
null
null
null
null
null
null
2,505.1699
Dimple: Discrete Diffusion Multimodal Large Language Model with Parallel Decoding
['Runpeng Yu', 'Xinyin Ma', 'Xinchao Wang']
['cs.CV']
In this work, we propose Dimple, the first Discrete Diffusion Multimodal Large Language Model (DMLLM). We observe that training with a purely discrete diffusion approach leads to significant training instability, suboptimal performance, and severe length bias issues. To address these challenges, we design a novel training paradigm that combines an initial autoregressive phase with a subsequent diffusion phase. This approach yields the Dimple-7B model, trained on the same dataset and using a similar training pipeline as LLaVA-NEXT. Dimple-7B ultimately surpasses LLaVA-NEXT in performance by 3.9%, demonstrating that DMLLM can achieve performance comparable to that of autoregressive models. To improve inference efficiency, we propose a decoding strategy termed confident decoding, which dynamically adjusts the number of tokens generated at each step, significantly reducing the number of generation iterations. In autoregressive models, the number of forward iterations during generation equals the response length. With confident decoding, however, the number of iterations needed by Dimple is even only $\frac{\text{response length}}{3}$. We also re-implement the prefilling technique in autoregressive models and demonstrate that it does not significantly impact performance on most benchmark evaluations, while offering a speedup of 1.5x to 7x. Additionally, we explore Dimple's capability to precisely control its response using structure priors. These priors enable structured responses in a manner distinct from instruction-based or chain-of-thought prompting, and allow fine-grained control over response format and length, which is difficult to achieve in autoregressive models. Overall, this work validates the feasibility and advantages of DMLLM and enhances its inference efficiency and controllability. Code and models are available at https://github.com/yu-rp/Dimple.
2025-05-22T17:55:04Z
null
null
null
null
null
null
null
null
null
null
2,505.16994
$\text{R}^2\text{ec}$: Towards Large Recommender Models with Reasoning
['Runyang You', 'Yongqi Li', 'Xinyu Lin', 'Xin Zhang', 'Wenjie Wang', 'Wenjie Li', 'Liqiang Nie']
['cs.IR', 'cs.AI', 'cs.CL']
Large recommender models have extended LLMs as powerful recommenders via encoding or item generation, and recent breakthroughs in LLM reasoning synchronously motivate the exploration of reasoning in recommendation. Current studies usually position LLMs as external reasoning modules to yield auxiliary thought for augmenting conventional recommendation pipelines. However, such decoupled designs are limited in significant resource cost and suboptimal joint optimization. To address these issues, we propose \name, a unified large recommender model with intrinsic reasoning capabilities. Initially, we reconceptualize the model architecture to facilitate interleaved reasoning and recommendation in the autoregressive process. Subsequently, we propose RecPO, a corresponding reinforcement learning framework that optimizes \name\ both the reasoning and recommendation capabilities simultaneously in a single policy update; RecPO introduces a fused reward scheme that solely leverages recommendation labels to simulate the reasoning capability, eliminating dependency on specialized reasoning annotations. Experiments on three datasets with various baselines verify the effectiveness of \name, showing relative improvements of 68.67\% in Hit@5 and 45.21\% in NDCG@20. Code available at https://github.com/YRYangang/RRec.
2025-05-22T17:55:43Z
null
null
null
null
null
null
null
null
null
null
2,505.17012
SpatialScore: Towards Unified Evaluation for Multimodal Spatial Understanding
['Haoning Wu', 'Xiao Huang', 'Yaohui Chen', 'Ya Zhang', 'Yanfeng Wang', 'Weidi Xie']
['cs.CV', 'cs.AI']
Multimodal large language models (MLLMs) have achieved impressive success in question-answering tasks, yet their capabilities for spatial understanding are less explored. This work investigates a critical question: do existing MLLMs possess 3D spatial perception and understanding abilities? Concretely, we make the following contributions in this paper: (i) we introduce VGBench, a benchmark specifically designed to assess MLLMs for visual geometry perception, e.g., camera pose and motion estimation; (ii) we propose SpatialScore, the most comprehensive and diverse multimodal spatial understanding benchmark to date, integrating VGBench with relevant data from the other 11 existing datasets. This benchmark comprises 28K samples across various spatial understanding tasks, modalities, and QA formats, along with a carefully curated challenging subset, SpatialScore-Hard; (iii) we develop SpatialAgent, a novel multi-agent system incorporating 9 specialized tools for spatial understanding, supporting both Plan-Execute and ReAct reasoning paradigms; (iv) we conduct extensive evaluations to reveal persistent challenges in spatial reasoning while demonstrating the effectiveness of SpatialAgent. We believe SpatialScore will offer valuable insights and serve as a rigorous benchmark for the next evolution of MLLMs.
2025-05-22T17:59:03Z
Technical Report; Project Page: https://haoningwu3639.github.io/SpatialScore
null
null
null
null
null
null
null
null
null
2,505.17016
Interactive Post-Training for Vision-Language-Action Models
['Shuhan Tan', 'Kairan Dou', 'Yue Zhao', 'Philipp Krähenbühl']
['cs.LG', 'cs.AI', 'cs.CV', 'cs.RO']
We introduce RIPT-VLA, a simple and scalable reinforcement-learning-based interactive post-training paradigm that fine-tunes pretrained Vision-Language-Action (VLA) models using only sparse binary success rewards. Existing VLA training pipelines rely heavily on offline expert demonstration data and supervised imitation, limiting their ability to adapt to new tasks and environments under low-data regimes. RIPT-VLA addresses this by enabling interactive post-training with a stable policy optimization algorithm based on dynamic rollout sampling and leave-one-out advantage estimation. RIPT-VLA has the following characteristics. First, it applies to various VLA models, resulting in an improvement on the lightweight QueST model by 21.2%, and the 7B OpenVLA-OFT model to an unprecedented 97.5% success rate. Second, it is computationally efficient and data-efficient: with only one demonstration, RIPT-VLA enables an unworkable SFT model (4%) to succeed with a 97% success rate within 15 iterations. Furthermore, we demonstrate that the policy learned by RIPT-VLA generalizes across different tasks and scenarios and is robust to the initial state context. These results highlight RIPT-VLA as a practical and effective paradigm for post-training VLA models through minimal supervision.
2025-05-22T17:59:45Z
Project page: https://ariostgx.github.io/ript_vla/
null
null
Interactive Post-Training for Vision-Language-Action Models
['Shuhan Tan', 'Kairan Dou', 'Yue Zhao', 'Philipp Krähenbühl']
2,025
arXiv.org
1
39
['Computer Science']
2,505.17018
SophiaVL-R1: Reinforcing MLLMs Reasoning with Thinking Reward
['Kaixuan Fan', 'Kaituo Feng', 'Haoming Lyu', 'Dongzhan Zhou', 'Xiangyu Yue']
['cs.CV']
Recent advances have shown success in eliciting strong reasoning abilities in multimodal large language models (MLLMs) through rule-based reinforcement learning (RL) with outcome rewards. However, this paradigm typically lacks supervision over the thinking process leading to the final outcome.As a result, the model may learn sub-optimal reasoning strategies, which can hinder its generalization ability. In light of this, we propose SophiaVL-R1, as an attempt to add reward signals for the thinking process in this paradigm. To achieve this, we first train a thinking reward model that evaluates the quality of the entire thinking process. Given that the thinking reward may be unreliable for certain samples due to reward hacking, we propose the Trust-GRPO method, which assigns a trustworthiness weight to the thinking reward during training. This weight is computed based on the thinking reward comparison of responses leading to correct answers versus incorrect answers, helping to mitigate the impact of potentially unreliable thinking rewards. Moreover, we design an annealing training strategy that gradually reduces the thinking reward over time, allowing the model to rely more on the accurate rule-based outcome reward in later training stages. Experiments show that our SophiaVL-R1 surpasses a series of reasoning MLLMs on various benchmarks (e.g., MathVisita, MMMU), demonstrating strong reasoning and generalization capabilities. Notably, our SophiaVL-R1-7B even outperforms LLaVA-OneVision-72B on most benchmarks, despite the latter having 10 times more parameters. All code, models, and datasets are made publicly available at https://github.com/kxfan2002/SophiaVL-R1.
2025-05-22T17:59:53Z
Project page:https://github.com/kxfan2002/SophiaVL-R1
null
null
null
null
null
null
null
null
null
2,505.17082
GemMaroc: Unlocking Darija Proficiency in LLMs with Minimal Data
['Abderrahman Skiredj', 'Ferdaous Azhari', 'Houdaifa Atou', 'Nouamane Tazi', 'Ismail Berrada']
['cs.CL', 'cs.AI']
Open-source large language models (LLMs) still marginalise Moroccan Arabic (Darija), forcing practitioners either to bolt on heavyweight Arabic adapters or to sacrifice the very reasoning skills that make LLMs useful. We show that a rigorously quality-over-quantity alignment strategy can surface fluent Darija while safeguarding the backbone s cross-lingual reasoning at a sliver of the usual compute. We translate three compact instruction suites LIMA 1 K, DEITA 6 K and TULU 50 K into Darija, preserve 20 of the English originals, and add mathematics, coding and scientific prompts. A LoRA-tuned Gemma 3-4B trained on 5 K mixed instructions lifts DarijaMMLU from 32.8 to 42.7 ; adding the reasoning-dense TULU portion pushes it to 47.5 with no English regression. Scaling the identical recipe to Gemma 3-27B produces GemMaroc-27B, which matches Atlas-Chat on DarijaMMLU (61.6 ) and leaps ahead on Darija commonsense, scoring 60.5 on HellaSwag versus Atlas-Chat s 48.4 . Crucially, GemMaroc retains Gemma-27B s strong maths and general-reasoning ability, showing only minimal movement on GSM8K and English benchmarks. The entire model is trained in just 48 GPU.h, underscoring a Green AI pathway to inclusive, sustainable language technology. We release code, data and checkpoints to spur Darija-centric applications in education, public services and everyday digital interaction.
2025-05-20T12:38:42Z
null
null
null
null
null
null
null
null
null
null
2,505.17102
BanglaByT5: Byte-Level Modelling for Bangla
['Pramit Bhattacharyya', 'Arnab Bhattacharya']
['cs.CL']
Large language models (LLMs) have achieved remarkable success across various natural language processing tasks. However, most LLM models use traditional tokenizers like BPE and SentencePiece, which fail to capture the finer nuances of a morphologically rich language like Bangla (Bengali). In this work, we introduce BanglaByT5, the first byte-level encoder-decoder model explicitly tailored for Bangla. Built upon a small variant of Googles ByT5 architecture, BanglaByT5 is pre-trained on a 14GB curated corpus combining high-quality literary and newspaper articles. Through zeroshot and supervised evaluations across generative and classification tasks, BanglaByT5 demonstrates competitive performance, surpassing several multilingual and larger models. Our findings highlight the efficacy of byte-level modelling for morphologically rich languages and highlight BanglaByT5 potential as a lightweight yet powerful tool for Bangla NLP, particularly in both resource-constrained and scalable environments.
2025-05-21T07:39:07Z
null
null
null
null
null
null
null
null
null
null
2,505.17166
ViDoRe Benchmark V2: Raising the Bar for Visual Retrieval
['Quentin Macé', 'António Loison', 'Manuel Faysse']
['cs.IR']
The ViDoRe Benchmark V1 was approaching saturation with top models exceeding 90% nDCG@5, limiting its ability to discern improvements. ViDoRe Benchmark V2 introduces realistic, challenging retrieval scenarios via blind contextual querying, long and cross-document queries, and a hybrid synthetic and human-in-the-loop query generation process. It comprises four diverse, multilingual datasets and provides clear evaluation instructions. Initial results demonstrate substantial room for advancement and highlight insights on model generalization and multilingual capability. This benchmark is designed as a living resource, inviting community contributions to maintain relevance through future evaluations.
2025-05-22T16:13:02Z
Published as a HuggingFace Blog
null
null
null
null
null
null
null
null
null
2,505.17266
Select2Reason: Efficient Instruction-Tuning Data Selection for Long-CoT Reasoning
['Cehao Yang', 'Xueyuan Lin', 'Chengjin Xu', 'Xuhui Jiang', 'Xiaojun Wu', 'Honghao Liu', 'Hui Xiong', 'Jian Guo']
['cs.CL', 'cs.AI']
A practical approach to activate long chain-of-thoughts reasoning ability in pre-trained large language models is to perform supervised fine-tuning on instruction datasets synthesized by strong Large Reasoning Models such as DeepSeek-R1, offering a cost-effective alternative to reinforcement learning. However, large-scale instruction sets with more than 100k samples incur significant training overhead, while effective strategies for automatic long-CoT instruction selection still remain unexplored. In this work, we propose Select2Reason, a novel and efficient instruction-tuning data selection framework for long-CoT reasoning. From the perspective of emergence of rethinking behaviors like self-correction and backtracking, we investigate common metrics that may determine the quality of long-CoT reasoning instructions. Select2Reason leverages a quantifier to estimate difficulty of question and jointly incorporates a reasoning trace length-based heuristic through a weighted scheme for ranking to prioritize high-utility examples. Empirical results on OpenR1-Math-220k demonstrate that fine-tuning LLM on only 10% of the data selected by Select2Reason achieves performance competitive with or superior to full-data tuning and open-source baseline OpenR1-Qwen-7B across three competition-level and six comprehensive mathematical benchmarks. Further experiments highlight the scalability in varying data size, efficiency during inference, and its adaptability to other instruction pools with minimal cost.
2025-05-22T20:24:08Z
null
null
null
Select2Reason: Efficient Instruction-Tuning Data Selection for Long-CoT Reasoning
['Cehao Yang', 'Xueyuan Lin', 'Chengjin Xu', 'Xuhui Jiang', 'Xiaojun Wu', 'Honghao Liu', 'Hui Xiong', 'Jian Guo']
2,025
arXiv.org
0
67
['Computer Science']
2,505.17373
Value-Guided Search for Efficient Chain-of-Thought Reasoning
['Kaiwen Wang', 'Jin Peng Zhou', 'Jonathan Chang', 'Zhaolin Gao', 'Nathan Kallus', 'Kianté Brantley', 'Wen Sun']
['cs.LG', 'cs.AI', 'cs.CL']
In this paper, we propose a simple and efficient method for value model training on long-context reasoning traces. Compared to existing process reward models (PRMs), our method does not require a fine-grained notion of "step," which is difficult to define for long-context reasoning models. By collecting a dataset of 2.5 million reasoning traces, we train a 1.5B token-level value model and apply it to DeepSeek models for improved performance with test-time compute scaling. We find that block-wise value-guided search (VGS) with a final weighted majority vote achieves better test-time scaling than standard methods such as majority voting or best-of-n. With an inference budget of 64 generations, VGS with DeepSeek-R1-Distill-1.5B achieves an average accuracy of 45.7% across four competition math benchmarks (AIME 2024 & 2025, HMMT Feb 2024 & 2025), reaching parity with o3-mini-medium. Moreover, VGS significantly reduces the inference FLOPs required to achieve the same performance of majority voting. Our dataset, model and codebase are open-sourced.
2025-05-23T01:05:07Z
null
null
null
Value-Guided Search for Efficient Chain-of-Thought Reasoning
['Kaiwen Wang', 'Jin Peng Zhou', 'Jonathan D. Chang', 'Zhaolin Gao', 'Nathan Kallus', 'Kianté Brantley', 'Wen Sun']
2,025
arXiv.org
1
54
['Computer Science']
2,505.17412
Direct3D-S2: Gigascale 3D Generation Made Easy with Spatial Sparse Attention
['Shuang Wu', 'Youtian Lin', 'Feihu Zhang', 'Yifei Zeng', 'Yikang Yang', 'Yajie Bao', 'Jiachen Qian', 'Siyu Zhu', 'Xun Cao', 'Philip Torr', 'Yao Yao']
['cs.CV']
Generating high-resolution 3D shapes using volumetric representations such as Signed Distance Functions (SDFs) presents substantial computational and memory challenges. We introduce Direct3D-S2, a scalable 3D generation framework based on sparse volumes that achieves superior output quality with dramatically reduced training costs. Our key innovation is the Spatial Sparse Attention (SSA) mechanism, which greatly enhances the efficiency of Diffusion Transformer (DiT) computations on sparse volumetric data. SSA allows the model to effectively process large token sets within sparse volumes, substantially reducing computational overhead and achieving a 3.9x speedup in the forward pass and a 9.6x speedup in the backward pass. Our framework also includes a variational autoencoder (VAE) that maintains a consistent sparse volumetric format across input, latent, and output stages. Compared to previous methods with heterogeneous representations in 3D VAE, this unified design significantly improves training efficiency and stability. Our model is trained on public available datasets, and experiments demonstrate that Direct3D-S2 not only surpasses state-of-the-art methods in generation quality and efficiency, but also enables training at 1024 resolution using only 8 GPUs, a task typically requiring at least 32 GPUs for volumetric representations at 256 resolution, thus making gigascale 3D generation both practical and accessible. Project page: https://www.neural4d.com/research/direct3d-s2.
2025-05-23T02:58:01Z
Project page: https://www.neural4d.com/research/direct3d-s2
null
null
Direct3D-S2: Gigascale 3D Generation Made Easy with Spatial Sparse Attention
['Shuang Wu', 'Youtian Lin', 'Feihu Zhang', 'Yifei Zeng', 'Yikang Yang', 'Yajie Bao', 'Jiachen Qian', 'Siyu Zhu', 'Xun Cao', 'Philip Torr', 'Yao Yao']
2,025
arXiv.org
1
49
['Computer Science']
2,505.17426
UniTTS: An end-to-end TTS system without decoupling of acoustic and semantic information
['Rui Wang', 'Qianguo Sun', 'Tianrong Chen', 'Zhiyun Zeng', 'Junlong Wu', 'Jiaxing Zhang']
['cs.SD', 'cs.AI', 'eess.AS']
The emergence of multi-codebook neutral audio codecs such as Residual Vector Quantization (RVQ) and Group Vector Quantization (GVQ) has significantly advanced Large-Language-Model (LLM) based Text-to-Speech (TTS) systems. These codecs are crucial in separating semantic and acoustic information while efficiently harnessing semantic priors. However, since semantic and acoustic information cannot be fully aligned, a significant drawback of these methods when applied to LLM-based TTS is that large language models may have limited access to comprehensive audio information. To address this limitation, we propose DistilCodec and UniTTS, which collectively offer the following advantages: 1) This method can distill a multi-codebook audio codec into a single-codebook audio codec with 32,768 codes while achieving a near 100\% utilization. 2) As DistilCodec does not employ a semantic alignment scheme, a large amount of high-quality unlabeled audio (such as audiobooks with sound effects, songs, etc.) can be incorporated during training, further expanding data diversity and broadening its applicability. 3) Leveraging the comprehensive audio information modeling of DistilCodec, we integrated three key tasks into UniTTS's pre-training framework: audio modality autoregression, text modality autoregression, and speech-text cross-modal autoregression. This allows UniTTS to accept interleaved text and speech/audio prompts while substantially preserving LLM's text capabilities. 4) UniTTS employs a three-stage training process: Pre-Training, Supervised Fine-Tuning (SFT), and Alignment. Source code and model checkpoints are publicly available at https://github.com/IDEA-Emdoor-Lab/UniTTS and https://github.com/IDEA-Emdoor-Lab/DistilCodec.
2025-05-23T03:13:46Z
null
null
null
UniTTS: An end-to-end TTS system without decoupling of acoustic and semantic information
['Rui Wang', 'Qianguo Sun', 'Tianrong Chen', 'Zhiyun Zeng', 'Junlong Wu', 'Jiaxing Zhang']
2,025
arXiv.org
0
42
['Computer Science', 'Engineering']
2,505.17496
Analyzing Mitigation Strategies for Catastrophic Forgetting in End-to-End Training of Spoken Language Models
['Chi-Yuan Hsiao', 'Ke-Han Lu', 'Kai-Wei Chang', 'Chih-Kai Yang', 'Wei-Chih Chen', 'Hung-yi Lee']
['cs.CL', 'cs.AI', 'cs.LG', 'cs.SD', 'eess.AS']
End-to-end training of Spoken Language Models (SLMs) commonly involves adapting pre-trained text-based Large Language Models (LLMs) to the speech modality through multi-stage training on diverse tasks such as ASR, TTS and spoken question answering (SQA). Although this multi-stage continual learning equips LLMs with both speech understanding and generation capabilities, the substantial differences in task and data distributions across stages can lead to catastrophic forgetting, where previously acquired knowledge is lost. This paper investigates catastrophic forgetting and evaluates three mitigation strategies-model merging, discounting the LoRA scaling factor, and experience replay to balance knowledge retention with new learning. Results show that experience replay is the most effective, with further gains achieved by combining it with other methods. These findings provide insights for developing more robust and efficient SLM training pipelines.
2025-05-23T05:50:14Z
Accepted to Interspeech 2025
null
null
null
null
null
null
null
null
null
2,505.17538
Swedish Whispers; Leveraging a Massive Speech Corpus for Swedish Speech Recognition
['Leonora Vesterbacka', 'Faton Rekathati', 'Robin Kurtz', 'Justyna Sikora', 'Agnes Toftgård']
['cs.CL', 'cs.SD', 'eess.AS']
This work presents a suite of fine-tuned Whisper models for Swedish, trained on a dataset of unprecedented size and variability for this mid-resourced language. As languages of smaller sizes are often underrepresented in multilingual training datasets, substantial improvements in performance can be achieved by fine-tuning existing multilingual models, as shown in this work. This work reports an overall improvement across model sizes compared to OpenAI's Whisper evaluated on Swedish. Most notably, we report an average 47% reduction in WER comparing our best performing model to OpenAI's whisper-large-v3, in evaluations across FLEURS, Common Voice, and NST.
2025-05-23T06:42:16Z
Submitted to Interspeech 2025
null
null
null
null
null
null
null
null
null
2,505.17592
AstroMLab 4: Benchmark-Topping Performance in Astronomy Q&A with a 70B-Parameter Domain-Specialized Reasoning Model
['Tijmen de Haan', 'Yuan-Sen Ting', 'Tirthankar Ghosal', 'Tuan Dung Nguyen', 'Alberto Accomazzi', 'Emily Herron', 'Vanessa Lama', 'Rui Pan', 'Azton Wells', 'Nesar Ramachandra']
['astro-ph.IM', 'cs.LG']
General-purpose large language models, despite their broad capabilities, often struggle with specialized domain knowledge, a limitation particularly pronounced in more accessible, lower-parameter versions. This gap hinders their deployment as effective agents in demanding fields such as astronomy. Building on our prior work with AstroSage-8B, this study introduces AstroSage-70B, a significantly larger and more advanced domain-specialized natural-language AI assistant. It is designed for research and education across astronomy, astrophysics, space science, astroparticle physics, cosmology, and astronomical instrumentation. Developed from the Llama-3.1-70B foundation, AstroSage-70B underwent extensive continued pre-training on a vast corpus of astronomical literature, followed by supervised fine-tuning and model merging. Beyond its 70-billion parameter scale, this model incorporates refined datasets, judiciously chosen learning hyperparameters, and improved training procedures, achieving state-of-the-art performance on complex astronomical tasks. Notably, we integrated reasoning chains into the SFT dataset, enabling AstroSage-70B to either answer the user query immediately, or first emit a human-readable thought process. Evaluated on the AstroMLab-1 benchmark -- comprising 4,425 questions from literature withheld during training -- AstroSage-70B achieves state-of-the-art performance. It surpasses all other tested open-weight and proprietary models, including leading systems like o3, Gemini-2.5-Pro, Claude-3.7-Sonnet, Deepseek-R1, and Qwen-3-235B, even those with API costs two orders of magnitude higher. This work demonstrates that domain specialization, when applied to large-scale models, can enable them to outperform generalist counterparts in specialized knowledge areas like astronomy, thereby advancing the frontier of AI capabilities in the field.
2025-05-23T07:58:50Z
null
null
null
AstroMLab 4: Benchmark-Topping Performance in Astronomy Q&A with a 70B-Parameter Domain-Specialized Reasoning Model
['Tijmen de Haan', 'Y.-S. Ting', 'Tirthankar Ghosal', 'Tuan Dung Nguyen', 'Alberto Accomazzi', 'Emily Herron', 'Vanessa Lama', 'Rui Pan', 'Azton Wells', 'Nesar Ramachandra']
2,025
arXiv.org
0
10
['Physics', 'Computer Science']
2,505.17612
Distilling LLM Agent into Small Models with Retrieval and Code Tools
['Minki Kang', 'Jongwon Jeong', 'Seanie Lee', 'Jaewoong Cho', 'Sung Ju Hwang']
['cs.CL', 'cs.AI']
Large language models (LLMs) excel at complex reasoning tasks but remain computationally expensive, limiting their practical deployment. To address this, recent works have focused on distilling reasoning capabilities into smaller language models (sLMs) using chain-of-thought (CoT) traces from teacher LLMs. However, this approach struggles in scenarios requiring rare factual knowledge or precise computation, where sLMs often hallucinate due to limited capability. In this work, we propose Agent Distillation, a framework for transferring not only reasoning capability but full task-solving behavior from LLM-based agents into sLMs with retrieval and code tools. We improve agent distillation along two complementary axes: (1) we introduce a prompting method called first-thought prefix to enhance the quality of teacher-generated trajectories; and (2) we propose a self-consistent action generation for improving test-time robustness of small agents. We evaluate our method on eight reasoning tasks across factual and mathematical domains, covering both in-domain and out-of-domain generalization. Our results show that sLMs as small as 0.5B, 1.5B, 3B parameters can achieve performance competitive with next-tier larger 1.5B, 3B, 7B models fine-tuned using CoT distillation, demonstrating the potential of agent distillation for building practical, tool-using small agents. Our code is available at https://github.com/Nardien/agent-distillation.
2025-05-23T08:20:15Z
preprint, v1
null
null
Distilling LLM Agent into Small Models with Retrieval and Code Tools
['Minki Kang', 'Jongwon Jeong', 'Seanie Lee', 'Jaewoong Cho', 'Sung Ju Hwang']
2,025
arXiv.org
2
74
['Computer Science']
2,505.17625
Enhancing Large Vision-Language Models with Layout Modality for Table Question Answering on Japanese Annual Securities Reports
['Hayato Aida', 'Kosuke Takahashi', 'Takahiro Omi']
['cs.CL', 'cs.CV', '68T50', 'I.2']
With recent advancements in Large Language Models (LLMs) and growing interest in retrieval-augmented generation (RAG), the ability to understand table structures has become increasingly important. This is especially critical in financial domains such as securities reports, where highly accurate question answering (QA) over tables is required. However, tables exist in various formats-including HTML, images, and plain text-making it difficult to preserve and extract structural information. Therefore, multimodal LLMs are essential for robust and general-purpose table understanding. Despite their promise, current Large Vision-Language Models (LVLMs), which are major representatives of multimodal LLMs, still face challenges in accurately understanding characters and their spatial relationships within documents. In this study, we propose a method to enhance LVLM-based table understanding by incorporating in-table textual content and layout features. Experimental results demonstrate that these auxiliary modalities significantly improve performance, enabling robust interpretation of complex document layouts without relying on explicitly structured input formats.
2025-05-23T08:36:22Z
Accepted at IIAI AAI 2025, the 3rd International Conference on Computational and Data Sciences in Economics and Finance
null
null
null
null
null
null
null
null
null
2,505.17667
QwenLong-L1: Towards Long-Context Large Reasoning Models with Reinforcement Learning
['Fanqi Wan', 'Weizhou Shen', 'Shengyi Liao', 'Yingcheng Shi', 'Chenliang Li', 'Ziyi Yang', 'Ji Zhang', 'Fei Huang', 'Jingren Zhou', 'Ming Yan']
['cs.CL']
Recent large reasoning models (LRMs) have demonstrated strong reasoning capabilities through reinforcement learning (RL). These improvements have primarily been observed within the short-context reasoning tasks. In contrast, extending LRMs to effectively process and reason on long-context inputs via RL remains a critical unsolved challenge. To bridge this gap, we first formalize the paradigm of long-context reasoning RL, and identify key challenges in suboptimal training efficiency and unstable optimization process. To address these issues, we propose QwenLong-L1, a framework that adapts short-context LRMs to long-context scenarios via progressive context scaling. Specifically, we utilize a warm-up supervised fine-tuning (SFT) stage to establish a robust initial policy, followed by a curriculum-guided phased RL technique to stabilize the policy evolution, and enhanced with a difficulty-aware retrospective sampling strategy to incentivize the policy exploration. Experiments on seven long-context document question-answering benchmarks demonstrate that QwenLong-L1-32B outperforms flagship LRMs like OpenAI-o3-mini and Qwen3-235B-A22B, achieving performance on par with Claude-3.7-Sonnet-Thinking, demonstrating leading performance among state-of-the-art LRMs. This work advances the development of practical long-context LRMs capable of robust reasoning across information-intensive environments.
2025-05-23T09:31:55Z
Technical Report
null
null
QwenLong-L1: Towards Long-Context Large Reasoning Models with Reinforcement Learning
['Fanqi Wan', 'Weizhou Shen', 'Shengyi Liao', 'Yingcheng Shi', 'Chenliang Li', 'Ziyi Yang', 'Ji Zhang', 'Fei Huang', 'Jingren Zhou', 'Ming Yan']
2,025
arXiv.org
0
64
['Computer Science']
2,505.17778
TextFlux: An OCR-Free DiT Model for High-Fidelity Multilingual Scene Text Synthesis
['Yu Xie', 'Jielei Zhang', 'Pengyu Chen', 'Ziyue Wang', 'Weihang Wang', 'Longwen Gao', 'Peiyi Li', 'Huyang Sun', 'Qiang Zhang', 'Qian Qiao', 'Jiaqing Fan', 'Zhouhui Lian']
['cs.CV']
Diffusion-based scene text synthesis has progressed rapidly, yet existing methods commonly rely on additional visual conditioning modules and require large-scale annotated data to support multilingual generation. In this work, we revisit the necessity of complex auxiliary modules and further explore an approach that simultaneously ensures glyph accuracy and achieves high-fidelity scene integration, by leveraging diffusion models' inherent capabilities for contextual reasoning. To this end, we introduce TextFlux, a DiT-based framework that enables multilingual scene text synthesis. The advantages of TextFlux can be summarized as follows: (1) OCR-free model architecture. TextFlux eliminates the need for OCR encoders (additional visual conditioning modules) that are specifically used to extract visual text-related features. (2) Strong multilingual scalability. TextFlux is effective in low-resource multilingual settings, and achieves strong performance in newly added languages with fewer than 1,000 samples. (3) Streamlined training setup. TextFlux is trained with only 1% of the training data required by competing methods. (4) Controllable multi-line text generation. TextFlux offers flexible multi-line synthesis with precise line-level control, outperforming methods restricted to single-line or rigid layouts. Extensive experiments and visualizations demonstrate that TextFlux outperforms previous methods in both qualitative and quantitative evaluations.
2025-05-23T11:46:46Z
null
null
null
TextFlux: An OCR-Free DiT Model for High-Fidelity Multilingual Scene Text Synthesis
['Yu Xie', 'Jielei Zhang', 'Pengyu Chen', 'Ziyue Wang', 'Weihang Wang', 'Longwen Gao', 'Peiyi Li', 'Huyang Sun', 'Qiang Zhang', 'Qian Qiao', 'Jiaqing Fan', 'Zhouhui Lian']
2,025
arXiv.org
1
55
['Computer Science']
2,505.17941
VeriThinker: Learning to Verify Makes Reasoning Model Efficient
['Zigeng Chen', 'Xinyin Ma', 'Gongfan Fang', 'Ruonan Yu', 'Xinchao Wang']
['cs.LG']
Large Reasoning Models (LRMs) excel at complex tasks using Chain-of-Thought (CoT) reasoning. However, their tendency to overthinking leads to unnecessarily lengthy reasoning chains, dramatically increasing inference costs. To mitigate this issue, we introduce VeriThinker, a novel approach for CoT compression. Unlike conventional methods that fine-tune LRMs directly on the original reasoning task using synthetic concise CoT data, we innovatively fine-tune the model solely through an auxiliary verification task. By training LRMs to accurately verify the correctness of CoT solutions, the LRMs inherently become more discerning about the necessity of subsequent self-reflection steps, thereby effectively suppressing overthinking. Extensive experiments validate that VeriThinker substantially reduces reasoning chain lengths while maintaining or even slightly improving accuracy. When applied to DeepSeek-R1-Distill-Qwen-7B, our approach reduces reasoning tokens on MATH500 from 3790 to 2125 while improving accuracy by 0.8% (94.0% to 94.8%), and on AIME25, tokens decrease from 14321 to 10287 with a 2.1% accuracy gain (38.7% to 40.8%). Additionally, our experiments demonstrate that VeriThinker can also be zero-shot generalized to speculative reasoning. Code is available at https://github.com/czg1225/VeriThinker
2025-05-23T14:17:56Z
Working in progress. Code Repo: https://github.com/czg1225/VeriThinker
null
null
VeriThinker: Learning to Verify Makes Reasoning Model Efficient
['Zigeng Chen', 'Xinyin Ma', 'Gongfan Fang', 'Ruonan Yu', 'Xinchao Wang']
2,025
arXiv.org
1
71
['Computer Science']
2,505.17952
Beyond Distillation: Pushing the Limits of Medical LLM Reasoning with Minimalist Rule-Based RL
['Che Liu', 'Haozhe Wang', 'Jiazhen Pan', 'Zhongwei Wan', 'Yong Dai', 'Fangzhen Lin', 'Wenjia Bai', 'Daniel Rueckert', 'Rossella Arcucci']
['cs.CL', 'cs.AI']
Improving performance on complex tasks and enabling interpretable decision making in large language models (LLMs), especially for clinical applications, requires effective reasoning. Yet this remains challenging without supervised fine-tuning (SFT) on costly chain-of-thought (CoT) data distilled from closed-source models (e.g., GPT-4o). In this work, we present AlphaMed, the first medical LLM to show that reasoning capability can emerge purely through reinforcement learning (RL), using minimalist rule-based rewards on public multiple-choice QA datasets, without relying on SFT or distilled CoT data. AlphaMed achieves state-of-the-art results on six medical QA benchmarks, outperforming models trained with conventional SFT+RL pipelines. On challenging benchmarks (e.g., MedXpert), AlphaMed even surpasses larger or closed-source models such as DeepSeek-V3-671B and Claude-3.5-Sonnet. To understand the factors behind this success, we conduct a comprehensive data-centric analysis guided by three questions: (i) Can minimalist rule-based RL incentivize reasoning without distilled CoT supervision? (ii) How do dataset quantity and diversity impact reasoning? (iii) How does question difficulty shape the emergence and generalization of reasoning? Our findings show that dataset informativeness is a key driver of reasoning performance, and that minimalist RL on informative, multiple-choice QA data is effective at inducing reasoning without CoT supervision. We also observe divergent trends across benchmarks, underscoring limitations in current evaluation and the need for more challenging, reasoning-oriented medical QA benchmarks.
2025-05-23T14:27:37Z
Under Review
null
null
Beyond Distillation: Pushing the Limits of Medical LLM Reasoning with Minimalist Rule-Based RL
['Che Liu', 'Haozhe Wang', 'Jiazhen Pan', 'Zhongwei Wan', 'Yong Dai', 'Fangzhen Lin', 'Wenjia Bai', 'D. Rueckert', 'Rossella Arcucci']
2,025
arXiv.org
1
48
['Computer Science']
2,505.18092
QwenLong-CPRS: Towards $\infty$-LLMs with Dynamic Context Optimization
['Weizhou Shen', 'Chenliang Li', 'Fanqi Wan', 'Shengyi Liao', 'Shaopeng Lai', 'Bo Zhang', 'Yingcheng Shi', 'Yuning Wu', 'Gang Fu', 'Zhansheng Li', 'Bin Yang', 'Ji Zhang', 'Fei Huang', 'Jingren Zhou', 'Ming Yan']
['cs.CL']
This technical report presents QwenLong-CPRS, a context compression framework designed for explicit long-context optimization, addressing prohibitive computation overhead during the prefill stage and the "lost in the middle" performance degradation of large language models (LLMs) during long sequence processing. Implemented through a novel dynamic context optimization mechanism, QwenLong-CPRS enables multi-granularity context compression guided by natural language instructions, achieving both efficiency gains and improved performance. Evolved from the Qwen architecture series, QwenLong-CPRS introduces four key innovations: (1) Natural language-guided dynamic optimization, (2) Bidirectional reasoning layers for enhanced boundary awareness, (3) Token critic mechanisms with language modeling heads, and (4) Window-parallel inference. Comprehensive evaluations across five benchmarks (4K-2M word contexts) demonstrate QwenLong-CPRS's threefold effectiveness: (1) Consistent superiority over other context management methods like RAG and sparse attention in both accuracy and efficiency. (2) Architecture-agnostic integration with all flagship LLMs, including GPT-4o, Gemini2.0-pro, Claude3.7-sonnet, DeepSeek-v3, and Qwen2.5-max, achieves 21.59$\times$ context compression alongside 19.15-point average performance gains; (3) Deployed with Qwen2.5-32B-Instruct, QwenLong-CPRS surpasses leading proprietary LLMs by 4.85 and 10.88 points on Ruler-128K and InfiniteBench, establishing new SOTA performance.
2025-05-23T16:47:00Z
null
null
null
QwenLong-CPRS: Towards ∞-LLMs with Dynamic Context Optimization
['Weizhou Shen', 'Chenliang Li', 'Fanqi Wan', 'Shengyi Liao', 'Shaopeng Lai', 'Bo Zhang', 'Yingcheng Shi', 'Yuning Wu', 'Gang Fu', 'Zhansheng Li', 'Bin Yang', 'Ji Zhang', 'Fei Huang', 'Jingren Zhou', 'Ming Yan']
2,025
arXiv.org
1
40
['Computer Science']
2,505.18125
TabSTAR: A Foundation Tabular Model With Semantically Target-Aware Representations
['Alan Arazi', 'Eilam Shapira', 'Roi Reichart']
['cs.LG', 'cs.CL']
While deep learning has achieved remarkable success across many domains, it has historically underperformed on tabular learning tasks, which remain dominated by gradient boosting decision trees (GBDTs). However, recent advancements are paving the way for Tabular Foundation Models, which can leverage real-world knowledge and generalize across diverse datasets, particularly when the data contains free-text. Although incorporating language model capabilities into tabular tasks has been explored, most existing methods utilize static, target-agnostic textual representations, limiting their effectiveness. We introduce TabSTAR: a Foundation Tabular Model with Semantically Target-Aware Representations. TabSTAR is designed to enable transfer learning on tabular data with textual features, with an architecture free of dataset-specific parameters. It unfreezes a pretrained text encoder and takes as input target tokens, which provide the model with the context needed to learn task-specific embeddings. TabSTAR achieves state-of-the-art performance for both medium- and large-sized datasets across known benchmarks of classification tasks with text features, and its pretraining phase exhibits scaling laws in the number of datasets, offering a pathway for further performance improvements.
2025-05-23T17:34:28Z
null
null
null
null
null
null
null
null
null
null
2,505.18129
One RL to See Them All: Visual Triple Unified Reinforcement Learning
['Yan Ma', 'Linge Du', 'Xuyang Shen', 'Shaoxiang Chen', 'Pengfei Li', 'Qibing Ren', 'Lizhuang Ma', 'Yuchao Dai', 'Pengfei Liu', 'Junjie Yan']
['cs.CV', 'cs.CL']
Reinforcement learning (RL) has significantly advanced the reasoning capabilities of vision-language models (VLMs). However, the use of RL beyond reasoning tasks remains largely unexplored, especially for perceptionintensive tasks like object detection and grounding. We propose V-Triune, a Visual Triple Unified Reinforcement Learning system that enables VLMs to jointly learn visual reasoning and perception tasks within a single training pipeline. V-Triune comprises triple complementary components: Sample-Level Data Formatting (to unify diverse task inputs), Verifier-Level Reward Computation (to deliver custom rewards via specialized verifiers) , and Source-Level Metric Monitoring (to diagnose problems at the data-source level). We further introduce a novel Dynamic IoU reward, which provides adaptive, progressive, and definite feedback for perception tasks handled by V-Triune. Our approach is instantiated within off-the-shelf RL training framework using open-source 7B and 32B backbone models. The resulting model, dubbed Orsta (One RL to See Them All), demonstrates consistent improvements across both reasoning and perception tasks. This broad capability is significantly shaped by its training on a diverse dataset, constructed around four representative visual reasoning tasks (Math, Puzzle, Chart, and Science) and four visual perception tasks (Grounding, Detection, Counting, and OCR). Subsequently, Orsta achieves substantial gains on MEGA-Bench Core, with improvements ranging from +2.1 to an impressive +14.1 across its various 7B and 32B model variants, with performance benefits extending to a wide range of downstream tasks. These results highlight the effectiveness and scalability of our unified RL approach for VLMs. The V-Triune system, along with the Orsta models, is publicly available at https://github.com/MiniMax-AI.
2025-05-23T17:41:14Z
Technical Report
null
null
null
null
null
null
null
null
null
2,505.18179
GAIA: A Foundation Model for Operational Atmospheric Dynamics
['Ata Akbari Asanjan', 'Olivia Alexander', 'Tom Berg', 'Clara Zhang', 'Matt Yang', 'Jad Makki', 'Disha Shidham', 'Srija Chakraborty', 'William Bender', 'Stephen Peng', 'Arun Ravindran', 'Olivier Raiman', 'David Potere', 'David Bell']
['cs.LG', 'cs.AI']
We present the GAIA (Geospatial Artificial Intelligence for Atmospheres) Foundation Model, a novel model that combines masked autoencoders (MAE) and self-DIstillation with NO labels (DINO) for analyzing global atmospheric patterns in satellite imagery. By integrating these complementary self-supervised learning approaches, our model simultaneously captures both local features and global dependencies. We address two critical challenges in satellite data analysis: reconstructing missing regions and estimating precipitation patterns as our first downstream tasks. The model demonstrates superior temporal pattern capture compared to standard MAE approaches, while maintaining robust performance in downstream tasks. Our experimental results show strong gap-filling capabilities across varying mask ratios and accurate precipitation estimation with limited training data, achieving a false alarm ratio of 0.088 and structural similarity of 0.881. This work represents an advancement in self-supervised learning for atmospheric science, providing a foundation for improved weather monitoring and climate analysis. The trained model weights and accompanying code are publicly available as open-source on Hugging Face here: https://huggingface.co/bcg-usra-nasa-gaia/GAIA-v1.
2025-05-15T05:07:09Z
14 pages, 7 figures
null
null
null
null
null
null
null
null
null
2,505.18383
NileChat: Towards Linguistically Diverse and Culturally Aware LLMs for Local Communities
['Abdellah El Mekki', 'Houdaifa Atou', 'Omer Nacar', 'Shady Shehata', 'Muhammad Abdul-Mageed']
['cs.CL']
Enhancing the linguistic capabilities of Large Language Models (LLMs) to include low-resource languages is a critical research area. Current research directions predominantly rely on synthetic data generated by translating English corpora, which, while demonstrating promising linguistic understanding and translation abilities, often results in models aligned with source language culture. These models frequently fail to represent the cultural heritage and values of local communities. This work proposes a methodology to create both synthetic and retrieval-based pre-training data tailored to a specific community, considering its (i) language, (ii) cultural heritage, and (iii) cultural values. We demonstrate our methodology using Egyptian and Moroccan dialects as testbeds, chosen for their linguistic and cultural richness and current underrepresentation in LLMs. As a proof-of-concept, we develop NileChat, a 3B parameter LLM adapted for Egyptian and Moroccan communities, incorporating their language, cultural heritage, and values. Our results on various understanding, translation, and cultural and values alignment benchmarks show that NileChat outperforms existing Arabic-aware LLMs of similar size and performs on par with larger models. We share our methods, data, and models with the community to promote the inclusion and coverage of more diverse communities in LLM development.
2025-05-23T21:18:40Z
null
null
null
null
null
null
null
null
null
null
2,505.18405
RaDeR: Reasoning-aware Dense Retrieval Models
['Debrup Das', "Sam O' Nuallain", 'Razieh Rahimi']
['cs.CL', 'cs.IR']
We propose RaDeR, a set of reasoning-based dense retrieval models trained with data derived from mathematical problem solving using large language models (LLMs). Our method leverages retrieval-augmented reasoning trajectories of an LLM and self-reflective relevance evaluation, enabling the creation of both diverse and hard-negative samples for reasoning-intensive relevance. RaDeR retrievers, trained for mathematical reasoning, effectively generalize to diverse reasoning tasks in the BRIGHT and RAR-b benchmarks, consistently outperforming strong baselines in overall performance. Notably, RaDeR achieves significantly higher performance than baselines on the Math and Coding splits. In addition, RaDeR presents the first dense retriever that outperforms BM25 when queries are Chain-of-Thought reasoning steps, underscoring the critical role of reasoning-based retrieval to augment reasoning language models. Furthermore, RaDeR achieves comparable or superior performance while using only 2.5% of the training data used by the concurrent work REASONIR, highlighting the quality of our synthesized training data.
2025-05-23T22:18:32Z
26 pages
null
null
RaDeR: Reasoning-aware Dense Retrieval Models
['Debrup Das', "Sam O' Nuallain", 'Razieh Rahimi']
2,025
arXiv.org
1
47
['Computer Science']
2,505.18445
OmniConsistency: Learning Style-Agnostic Consistency from Paired Stylization Data
['Yiren Song', 'Cheng Liu', 'Mike Zheng Shou']
['cs.CV']
Diffusion models have advanced image stylization significantly, yet two core challenges persist: (1) maintaining consistent stylization in complex scenes, particularly identity, composition, and fine details, and (2) preventing style degradation in image-to-image pipelines with style LoRAs. GPT-4o's exceptional stylization consistency highlights the performance gap between open-source methods and proprietary models. To bridge this gap, we propose \textbf{OmniConsistency}, a universal consistency plugin leveraging large-scale Diffusion Transformers (DiTs). OmniConsistency contributes: (1) an in-context consistency learning framework trained on aligned image pairs for robust generalization; (2) a two-stage progressive learning strategy decoupling style learning from consistency preservation to mitigate style degradation; and (3) a fully plug-and-play design compatible with arbitrary style LoRAs under the Flux framework. Extensive experiments show that OmniConsistency significantly enhances visual coherence and aesthetic quality, achieving performance comparable to commercial state-of-the-art model GPT-4o.
2025-05-24T01:00:20Z
null
null
null
OmniConsistency: Learning Style-Agnostic Consistency from Paired Stylization Data
['Yiren Song', 'Cheng Liu', 'Mike Zheng Shou']
2,025
arXiv.org
2
46
['Computer Science']
2,505.18495
Beyond Masked and Unmasked: Discrete Diffusion Models via Partial Masking
['Chen-Hao Chao', 'Wei-Fang Sun', 'Hanwen Liang', 'Chun-Yi Lee', 'Rahul G. Krishnan']
['cs.LG']
Masked diffusion models (MDM) are powerful generative models for discrete data that generate samples by progressively unmasking tokens in a sequence. Each token can take one of two states: masked or unmasked. We observe that token sequences often remain unchanged between consecutive sampling steps; consequently, the model repeatedly processes identical inputs, leading to redundant computation. To address this inefficiency, we propose the Partial masking scheme (Prime), which augments MDM by allowing tokens to take intermediate states interpolated between the masked and unmasked states. This design enables the model to make predictions based on partially observed token information, and facilitates a fine-grained denoising process. We derive a variational training objective and introduce a simple architectural design to accommodate intermediate-state inputs. Our method demonstrates superior performance across a diverse set of generative modeling tasks. On text data, it achieves a perplexity of 15.36 on OpenWebText, outperforming previous MDM (21.52), autoregressive models (17.54), and their hybrid variants (17.58), without relying on an autoregressive formulation. On image data, it attains competitive FID scores of 3.26 on CIFAR-10 and 6.98 on ImageNet-32, comparable to leading continuous generative models.
2025-05-24T04:16:40Z
null
null
null
Beyond Masked and Unmasked: Discrete Diffusion Models via Partial Masking
['Chen-Hao Chao', 'Wei-Fang Sun', 'Hanwen Liang', 'Chun-Yi Lee', 'Rahul G. Krishnan']
2,025
arXiv.org
0
65
['Computer Science']
2,505.18499
G1: Teaching LLMs to Reason on Graphs with Reinforcement Learning
['Xiaojun Guo', 'Ang Li', 'Yifei Wang', 'Stefanie Jegelka', 'Yisen Wang']
['cs.LG', 'cs.AI', 'stat.ML']
Although Large Language Models (LLMs) have demonstrated remarkable progress, their proficiency in graph-related tasks remains notably limited, hindering the development of truly general-purpose models. Previous attempts, including pretraining graph foundation models or employing supervised fine-tuning, often face challenges such as the scarcity of large-scale, universally represented graph data. We introduce G1, a simple yet effective approach demonstrating that Reinforcement Learning (RL) on synthetic graph-theoretic tasks can significantly scale LLMs' graph reasoning abilities. To enable RL training, we curate Erd\~os, the largest graph reasoning dataset to date comprising 50 diverse graph-theoretic tasks of varying difficulty levels, 100k training data and 5k test data, all drived from real-world graphs. With RL on Erd\~os, G1 obtains substantial improvements in graph reasoning, where our finetuned 3B model even outperforms Qwen2.5-72B-Instruct (24x size). RL-trained models also show strong zero-shot generalization to unseen tasks, domains, and graph encoding schemes, including other graph-theoretic benchmarks as well as real-world node classification and link prediction tasks, without compromising general reasoning abilities. Our findings offer an efficient, scalable path for building strong graph reasoners by finetuning LLMs with RL on graph-theoretic tasks, which combines the strengths of pretrained LLM capabilities with abundant, automatically generated synthetic data, suggesting that LLMs possess graph understanding abilities that RL can elicit successfully. Our implementation is open-sourced at https://github.com/PKU-ML/G1, with models and datasets hosted on Hugging Face collections https://huggingface.co/collections/PKU-ML/g1-683d659e992794fc99618cf2 for broader accessibility.
2025-05-24T04:33:41Z
null
null
null
null
null
null
null
null
null
null
2,505.18601
Flex-Judge: Think Once, Judge Anywhere
['Jongwoo Ko', 'Sungnyun Kim', 'Sungwoo Cho', 'Se-Young Yun']
['cs.CL', 'cs.AI']
Human-generated reward signals are critical for aligning generative models with human preferences, guiding both training and inference-time evaluations. While large language models (LLMs) employed as proxy evaluators, i.e., LLM-as-a-Judge, significantly reduce the costs associated with manual annotations, they typically require extensive modality-specific training data and fail to generalize well across diverse multimodal tasks. In this paper, we propose Flex-Judge, a reasoning-guided multimodal judge model that leverages minimal textual reasoning data to robustly generalize across multiple modalities and evaluation formats. Our core intuition is that structured textual reasoning explanations inherently encode generalizable decision-making patterns, enabling an effective transfer to multimodal judgments, e.g., with images or videos. Empirical results demonstrate that Flex-Judge, despite being trained on significantly fewer text data, achieves competitive or superior performance compared to state-of-the-art commercial APIs and extensively trained multimodal evaluators. Notably, Flex-Judge presents broad impact in modalities like molecule, where comprehensive evaluation benchmarks are scarce, underscoring its practical value in resource-constrained domains. Our framework highlights reasoning-based text supervision as a powerful, cost-effective alternative to traditional annotation-intensive approaches, substantially advancing scalable multimodal model-as-a-judge.
2025-05-24T08:50:53Z
The code is available at https://github.com/jongwooko/flex-judge
null
null
Flex-Judge: Think Once, Judge Anywhere
['Jongwoo Ko', 'Sungnyun Kim', 'Sungwoo Cho', 'Se-Young Yun']
2,025
arXiv.org
0
87
['Computer Science']
2,505.18842
Don't Look Only Once: Towards Multimodal Interactive Reasoning with Selective Visual Revisitation
['Jiwan Chung', 'Junhyeok Kim', 'Siyeol Kim', 'Jaeyoung Lee', 'Min Soo Kim', 'Youngjae Yu']
['cs.CL', 'cs.CV']
We present v1, a lightweight extension to Multimodal Large Language Models (MLLMs) that enables selective visual revisitation during inference. While current MLLMs typically consume visual input only once and reason purely over internal memory, v1 introduces a simple point-and-copy mechanism that allows the model to dynamically retrieve relevant image regions throughout the reasoning process. This mechanism augments existing architectures with minimal modifications, enabling contextual access to visual tokens based on the model's evolving hypotheses. To train this capability, we construct v1g, a dataset of 300K multimodal reasoning traces with interleaved visual grounding annotations. Experiments on three multimodal mathematical reasoning benchmarks -- MathVista, MathVision, and MathVerse -- demonstrate that v1 consistently improves performance over comparable baselines, particularly on tasks requiring fine-grained visual reference and multi-step reasoning. Our results suggest that dynamic visual access is a promising direction for enhancing grounded multimodal reasoning. Code, models, and data will be released to support future research.
2025-05-24T19:30:47Z
null
null
null
null
null
null
null
null
null
null
2,505.19
VerIPO: Cultivating Long Reasoning in Video-LLMs via Verifier-Gudied Iterative Policy Optimization
['Yunxin Li', 'Xinyu Chen', 'Zitao Li', 'Zhenyu Liu', 'Longyue Wang', 'Wenhan Luo', 'Baotian Hu', 'Min Zhang']
['cs.CL', 'cs.CV']
Applying Reinforcement Learning (RL) to Video Large Language Models (Video-LLMs) shows significant promise for complex video reasoning. However, popular Reinforcement Fine-Tuning (RFT) methods, such as outcome-based Group Relative Policy Optimization (GRPO), are limited by data preparation bottlenecks (e.g., noise or high cost) and exhibit unstable improvements in the quality of long chain-of-thoughts (CoTs) and downstream performance.To address these limitations, we propose VerIPO, a Verifier-guided Iterative Policy Optimization method designed to gradually improve video LLMs' capacity for generating deep, long-term reasoning chains. The core component is Rollout-Aware Verifier, positioned between the GRPO and Direct Preference Optimization (DPO) training phases to form the GRPO-Verifier-DPO training loop. This verifier leverages small LLMs as a judge to assess the reasoning logic of rollouts, enabling the construction of high-quality contrastive data, including reflective and contextually consistent CoTs. These curated preference samples drive the efficient DPO stage (7x faster than GRPO), leading to marked improvements in reasoning chain quality, especially in terms of length and contextual consistency. This training loop benefits from GRPO's expansive search and DPO's targeted optimization. Experimental results demonstrate: 1) Significantly faster and more effective optimization compared to standard GRPO variants, yielding superior performance; 2) Our trained models exceed the direct inference of large-scale instruction-tuned Video-LLMs, producing long and contextually consistent CoTs on diverse video reasoning tasks; and 3) Our model with one iteration outperforms powerful LMMs (e.g., Kimi-VL) and long reasoning models (e.g., Video-R1), highlighting its effectiveness and stability.
2025-05-25T06:41:28Z
19 pages, 9 figures, Project Link: https://github.com/HITsz-TMG/VerIPO
null
null
null
null
null
null
null
null
null
2,505.19084
Jodi: Unification of Visual Generation and Understanding via Joint Modeling
['Yifeng Xu', 'Zhenliang He', 'Meina Kan', 'Shiguang Shan', 'Xilin Chen']
['cs.CV', 'cs.AI', 'cs.LG']
Visual generation and understanding are two deeply interconnected aspects of human intelligence, yet they have been traditionally treated as separate tasks in machine learning. In this paper, we propose Jodi, a diffusion framework that unifies visual generation and understanding by jointly modeling the image domain and multiple label domains. Specifically, Jodi is built upon a linear diffusion transformer along with a role switch mechanism, which enables it to perform three particular types of tasks: (1) joint generation, where the model simultaneously generates images and multiple labels; (2) controllable generation, where images are generated conditioned on any combination of labels; and (3) image perception, where multiple labels can be predicted at once from a given image. Furthermore, we present the Joint-1.6M dataset, which contains 200,000 high-quality images collected from public sources, automatic labels for 7 visual domains, and LLM-generated captions. Extensive experiments demonstrate that Jodi excels in both generation and understanding tasks and exhibits strong extensibility to a wider range of visual domains. Code is available at https://github.com/VIPL-GENUN/Jodi.
2025-05-25T10:40:52Z
Code: https://github.com/VIPL-GENUN/Jodi
null
null
null
null
null
null
null
null
null
2,505.19094
SATORI-R1: Incentivizing Multimodal Reasoning with Spatial Grounding and Verifiable Rewards
['Chuming Shen', 'Wei Wei', 'Xiaoye Qu', 'Yu Cheng']
['cs.CV', 'cs.AI']
DeepSeek-R1 has demonstrated powerful reasoning capabilities in the text domain through stable reinforcement learning (RL). Recently, in the multimodal domain, works have begun to directly apply RL to generate R1-like free-form reasoning for Visual Question Answering (VQA) tasks. However, multimodal tasks share an intrinsically different nature from textual tasks, which heavily rely on the understanding of the input image to solve the problem. Therefore, such free-form reasoning faces two critical limitations in the VQA task: (1) Extended reasoning chains diffuse visual focus away from task-critical regions, degrading answer accuracy. (2) Unverifiable intermediate steps amplify policy-gradient variance and computational costs overhead. To address these issues, in this paper, we introduce SATORI ($\textbf{S}patially$ $\textbf{A}nchored$ $\textbf{T}ask$ $\textbf{O}ptimization$ with $\textbf{R}e\textbf{I}nforcement$ Learning), which decomposes VQA into three verifiable stages, including global image captioning, region localization, and answer prediction, each supplying explicit reward signals. Furthermore, we also introduce VQA-Verify, a 12k dataset annotated with answer-aligned captions and bounding-boxes to facilitate training. Experiments demonstrate consistent performance improvements across seven VQA benchmarks, achieving up to $15.7\%$ improvement in accuracy in accuracy compared to the R1-like baseline. Our analysis of the attention map confirms enhanced focus on critical regions, which brings improvements in accuracy. Our code is available at https://github.com/justairr/SATORI-R1.
2025-05-25T11:11:06Z
Under review
null
null
null
null
null
null
null
null
null
2,505.19095
ScreenExplorer: Training a Vision-Language Model for Diverse Exploration in Open GUI World
['Runliang Niu', 'Jinglong Ji', 'Yi Chang', 'Qi Wang']
['cs.AI']
The rapid progress of large language models (LLMs) has sparked growing interest in building Artificial General Intelligence (AGI) within Graphical User Interface (GUI) environments. However, existing GUI agents based on LLMs or vision-language models (VLMs) often fail to generalize to novel environments and rely heavily on manually curated, diverse datasets. To overcome these limitations, we introduce ScreenExplorer, a VLM trained via Group Relative Policy Optimization(GRPO) in real, dynamic, and open-ended GUI environments. Innovatively, we introduced a world-model-based curiosity reward function to help the agent overcome the cold-start phase of exploration. Additionally, distilling experience streams further enhances the model's exploration capabilities. Our training framework enhances model exploration in open GUI environments, with trained models showing better environmental adaptation and sustained exploration compared to static deployment models. Our findings offer a scalable pathway toward AGI systems with self-improving capabilities in complex interactive settings.
2025-05-25T11:13:03Z
null
null
null
null
null
null
null
null
null
null
2,505.19103
WHISTRESS: Enriching Transcriptions with Sentence Stress Detection
['Iddo Yosha', 'Dorin Shteyman', 'Yossi Adi']
['cs.CL', 'cs.SD', 'eess.AS']
Spoken language conveys meaning not only through words but also through intonation, emotion, and emphasis. Sentence stress, the emphasis placed on specific words within a sentence, is crucial for conveying speaker intent and has been extensively studied in linguistics. In this work, we introduce WHISTRESS, an alignment-free approach for enhancing transcription systems with sentence stress detection. To support this task, we propose TINYSTRESS-15K, a scalable, synthetic training data for the task of sentence stress detection which resulted from a fully automated dataset creation process. We train WHISTRESS on TINYSTRESS-15K and evaluate it against several competitive baselines. Our results show that WHISTRESS outperforms existing methods while requiring no additional input priors during training or inference. Notably, despite being trained on synthetic data, WHISTRESS demonstrates strong zero-shot generalization across diverse benchmarks. Project page: https://pages.cs.huji.ac.il/adiyoss-lab/whistress.
2025-05-25T11:45:08Z
Accepted to Interspeech2025
null
null
null
null
null
null
null
null
null
2,505.19114
CreatiDesign: A Unified Multi-Conditional Diffusion Transformer for Creative Graphic Design
['Hui Zhang', 'Dexiang Hong', 'Maoke Yang', 'Yutao Cheng', 'Zhao Zhang', 'Jie Shao', 'Xinglong Wu', 'Zuxuan Wu', 'Yu-Gang Jiang']
['cs.CV']
Graphic design plays a vital role in visual communication across advertising, marketing, and multimedia entertainment. Prior work has explored automated graphic design generation using diffusion models, aiming to streamline creative workflows and democratize design capabilities. However, complex graphic design scenarios require accurately adhering to design intent specified by multiple heterogeneous user-provided elements (\eg images, layouts, and texts), which pose multi-condition control challenges for existing methods. Specifically, previous single-condition control models demonstrate effectiveness only within their specialized domains but fail to generalize to other conditions, while existing multi-condition methods often lack fine-grained control over each sub-condition and compromise overall compositional harmony. To address these limitations, we introduce CreatiDesign, a systematic solution for automated graphic design covering both model architecture and dataset construction. First, we design a unified multi-condition driven architecture that enables flexible and precise integration of heterogeneous design elements with minimal architectural modifications to the base diffusion model. Furthermore, to ensure that each condition precisely controls its designated image region and to avoid interference between conditions, we propose a multimodal attention mask mechanism. Additionally, we develop a fully automated pipeline for constructing graphic design datasets, and introduce a new dataset with 400K samples featuring multi-condition annotations, along with a comprehensive benchmark. Experimental results show that CreatiDesign outperforms existing models by a clear margin in faithfully adhering to user intent.
2025-05-25T12:14:23Z
null
null
null
null
null
null
null
null
null
null
2,505.19201
DREAM: Drafting with Refined Target Features and Entropy-Adaptive Cross-Attention Fusion for Multimodal Speculative Decoding
['Yunhai Hu', 'Tianhua Xia', 'Zining Liu', 'Rahul Raman', 'Xingyu Liu', 'Bo Bao', 'Eric Sather', 'Vithursan Thangarasa', 'Sai Qian Zhang']
['cs.CL']
Speculative decoding (SD) has emerged as a powerful method for accelerating autoregressive generation in large language models (LLMs), yet its integration into vision-language models (VLMs) remains underexplored. We introduce DREAM, a novel speculative decoding framework tailored for VLMs that combines three key innovations: (1) a cross-attention-based mechanism to inject intermediate features from the target model into the draft model for improved alignment, (2) adaptive intermediate feature selection based on attention entropy to guide efficient draft model training, and (3) visual token compression to reduce draft model latency. DREAM enables efficient, accurate, and parallel multimodal decoding with significant throughput improvement. Experiments across a diverse set of recent popular VLMs, including LLaVA, Pixtral, SmolVLM and Gemma3, demonstrate up to 3.6x speedup over conventional decoding and significantly outperform prior SD baselines in both inference throughput and speculative draft acceptance length across a broad range of multimodal benchmarks. The code is publicly available at: https://github.com/SAI-Lab-NYU/DREAM.git
2025-05-25T15:56:50Z
null
null
null
null
null
null
null
null
null
null
2,505.19225
MedITok: A Unified Tokenizer for Medical Image Synthesis and Interpretation
['Chenglong Ma', 'Yuanfeng Ji', 'Jin Ye', 'Zilong Li', 'Chenhui Wang', 'Junzhi Ning', 'Wei Li', 'Lihao Liu', 'Qiushan Guo', 'Tianbin Li', 'Junjun He', 'Hongming Shan']
['eess.IV', 'cs.CV']
Advanced autoregressive models have reshaped multimodal AI. However, their transformative potential in medical imaging remains largely untapped due to the absence of a unified visual tokenizer -- one capable of capturing fine-grained visual structures for faithful image reconstruction and realistic image synthesis, as well as rich semantics for accurate diagnosis and image interpretation. To this end, we present MedITok, the first unified tokenizer tailored for medical images, encoding both low-level structural details and high-level clinical semantics within a unified latent space. To balance these competing objectives, we introduce a novel two-stage training framework: a visual representation alignment stage that cold-starts the tokenizer reconstruction learning with a visual semantic constraint, followed by a textual semantic representation alignment stage that infuses detailed clinical semantics into the latent space. Trained on the meticulously collected large-scale dataset with over 30 million medical images and 2 million image-caption pairs, MedITok achieves state-of-the-art performance on more than 30 datasets across 9 imaging modalities and 4 different tasks. By providing a unified token space for autoregressive modeling, MedITok supports a wide range of tasks in clinical diagnostics and generative healthcare applications. Model and code will be made publicly available at: https://github.com/Masaaki-75/meditok.
2025-05-25T16:39:35Z
null
null
null
MedITok: A Unified Tokenizer for Medical Image Synthesis and Interpretation
['Chenglong Ma', 'Yuanfeng Ji', 'Jin Ye', 'Zilong Li', 'Chenhui Wang', 'Junzhi Ning', 'Wei Li', 'Lihao Liu', 'Qiushan Guo', 'Tian-Xin Li', 'Junjun He', 'Hongming Shan']
2,025
arXiv.org
0
0
['Computer Science', 'Engineering']
2,505.19274
Conventional Contrastive Learning Often Falls Short: Improving Dense Retrieval with Cross-Encoder Listwise Distillation and Synthetic Data
['Manveer Singh Tamber', 'Suleman Kazi', 'Vivek Sourabh', 'Jimmy Lin']
['cs.IR']
We investigate improving the retrieval effectiveness of embedding models through the lens of corpus-specific fine-tuning. Prior work has shown that fine-tuning with queries generated using a dataset's retrieval corpus can boost retrieval effectiveness for the dataset. However, we find that surprisingly, fine-tuning using the conventional InfoNCE contrastive loss often reduces effectiveness in state-of-the-art models. To overcome this, we revisit cross-encoder listwise distillation and demonstrate that, unlike using contrastive learning alone, listwise distillation can help more consistently improve retrieval effectiveness across multiple datasets. Additionally, we show that synthesizing more training data using diverse query types (such as claims, keywords, and questions) yields greater effectiveness than using any single query type alone, regardless of the query type used in evaluation. Our findings further indicate that synthetic queries offer comparable utility to human-written queries for training. We use our approach to train an embedding model that achieves state-of-the-art effectiveness among BERT embedding models. We release our model and both query generation and training code to facilitate further research.
2025-05-25T19:06:19Z
updated version of arxiv:2502.19712
null
null
Conventional Contrastive Learning Often Falls Short: Improving Dense Retrieval with Cross-Encoder Listwise Distillation and Synthetic Data
['M. Tamber', 'Suleman Kazi', 'Vivek Sourabh', 'Jimmy Lin']
2,025
arXiv.org
0
63
['Computer Science']
2,505.19314
SoloSpeech: Enhancing Intelligibility and Quality in Target Speech Extraction through a Cascaded Generative Pipeline
['Helin Wang', 'Jiarui Hai', 'Dongchao Yang', 'Chen Chen', 'Kai Li', 'Junyi Peng', 'Thomas Thebaud', 'Laureano Moro Velazquez', 'Jesus Villalba', 'Najim Dehak']
['eess.AS', 'cs.AI', 'cs.SD']
Target Speech Extraction (TSE) aims to isolate a target speaker's voice from a mixture of multiple speakers by leveraging speaker-specific cues, typically provided as auxiliary audio (a.k.a. cue audio). Although recent advancements in TSE have primarily employed discriminative models that offer high perceptual quality, these models often introduce unwanted artifacts, reduce naturalness, and are sensitive to discrepancies between training and testing environments. On the other hand, generative models for TSE lag in perceptual quality and intelligibility. To address these challenges, we present SoloSpeech, a novel cascaded generative pipeline that integrates compression, extraction, reconstruction, and correction processes. SoloSpeech features a speaker-embedding-free target extractor that utilizes conditional information from the cue audio's latent space, aligning it with the mixture audio's latent space to prevent mismatches. Evaluated on the widely-used Libri2Mix dataset, SoloSpeech achieves the new state-of-the-art intelligibility and quality in target speech extraction and speech separation tasks while demonstrating exceptional generalization on out-of-domain data and real-world scenarios.
2025-05-25T21:00:48Z
null
null
null
null
null
null
null
null
null
null
2,505.19356
Optimized Text Embedding Models and Benchmarks for Amharic Passage Retrieval
['Kidist Amde Mekonnen', 'Yosef Worku Alemneh', 'Maarten de Rijke']
['cs.IR', 'cs.AI', 'cs.CL', 'cs.LG', '68T50 (Primary), 68T05 (Secondary)', 'H.3.3; H.3.1; I.2.7']
Neural retrieval methods using transformer-based pre-trained language models have advanced multilingual and cross-lingual retrieval. However, their effectiveness for low-resource, morphologically rich languages such as Amharic remains underexplored due to data scarcity and suboptimal tokenization. We address this gap by introducing Amharic-specific dense retrieval models based on pre-trained Amharic BERT and RoBERTa backbones. Our proposed RoBERTa-Base-Amharic-Embed model (110M parameters) achieves a 17.6% relative improvement in MRR@10 and a 9.86% gain in Recall@10 over the strongest multilingual baseline, Arctic Embed 2.0 (568M parameters). More compact variants, such as RoBERTa-Medium-Amharic-Embed (42M), remain competitive while being over 13x smaller. Additionally, we train a ColBERT-based late interaction retrieval model that achieves the highest MRR@10 score (0.843) among all evaluated models. We benchmark our proposed models against both sparse and dense retrieval baselines to systematically assess retrieval effectiveness in Amharic. Our analysis highlights key challenges in low-resource settings and underscores the importance of language-specific adaptation. To foster future research in low-resource IR, we publicly release our dataset, codebase, and trained models at https://github.com/kidist-amde/amharic-ir-benchmarks.
2025-05-25T23:06:20Z
10 pages (excl. refs/appendix), 10 figures. Accepted to ACL 2025 Findings. Kidist and Yosef contributed equally to this work. Public resources: https://github.com/kidist-amde/amharic-ir-benchmarks
null
null
Optimized Text Embedding Models and Benchmarks for Amharic Passage Retrieval
['Kidist Amde Mekonnen', 'Yosef Alemneh', 'M. D. Rijke']
2,025
arXiv.org
0
52
['Computer Science']
2,505.19536
FlowCut: Rethinking Redundancy via Information Flow for Efficient Vision-Language Models
['Jintao Tong', 'Wenwei Jin', 'Pengda Qin', 'Anqi Li', 'Yixiong Zou', 'Yuhong Li', 'Yuhua Li', 'Ruixuan Li']
['cs.CV', 'cs.AI', 'cs.CL']
Large vision-language models (LVLMs) excel at multimodal understanding but suffer from high computational costs due to redundant vision tokens. Existing pruning methods typically rely on single-layer attention scores to rank and prune redundant visual tokens to solve this inefficiency. However, as the interaction between tokens and layers is complicated, this raises a basic question: Is such a simple single-layer criterion sufficient to identify redundancy? To answer this question, we rethink the emergence of redundant visual tokens from a fundamental perspective: information flow, which models the interaction between tokens and layers by capturing how information moves between tokens across layers. We find (1) the CLS token acts as an information relay, which can simplify the complicated flow analysis; (2) the redundancy emerges progressively and dynamically via layer-wise attention concentration; and (3) relying solely on attention scores from single layers can lead to contradictory redundancy identification. Based on this, we propose FlowCut, an information-flow-aware pruning framework, mitigating the insufficiency of the current criterion for identifying redundant tokens and better aligning with the model's inherent behaviors. Extensive experiments show that FlowCut achieves superior results, outperforming SoTA by 1.6% on LLaVA-1.5-7B with 88.9% token reduction, and by 4.3% on LLaVA-NeXT-7B with 94.4% reduction, delivering 3.2x speed-up in the prefilling stage. Our code is available at https://github.com/TungChintao/FlowCut
2025-05-26T05:54:48Z
19 pages, 11 figures
null
null
FlowCut: Rethinking Redundancy via Information Flow for Efficient Vision-Language Models
['Jintao Tong', 'Wenwei Jin', 'Pengda Qin', 'Anqi Li', 'Yixiong Zou', 'Yuhong Li', 'Yuhua Li', 'Ruixuan Li']
2,025
arXiv.org
0
51
['Computer Science']
2,505.1959
Learning to Reason without External Rewards
['Xuandong Zhao', 'Zhewei Kang', 'Aosong Feng', 'Sergey Levine', 'Dawn Song']
['cs.LG', 'cs.CL']
Training large language models (LLMs) for complex reasoning via Reinforcement Learning with Verifiable Rewards (RLVR) is effective but limited by reliance on costly, domain-specific supervision. We explore Reinforcement Learning from Internal Feedback (RLIF), a framework that enables LLMs to learn from intrinsic signals without external rewards or labeled data. We propose Intuitor, an RLIF method that uses a model's own confidence, termed self-certainty, as its sole reward signal. Intuitor replaces external rewards in Group Relative Policy Optimization (GRPO) with self-certainty scores, enabling fully unsupervised learning. Experiments demonstrate that Intuitor matches GRPO's performance on mathematical benchmarks while achieving superior generalization to out-of-domain tasks like code generation, without requiring gold solutions or test cases. Our findings show that intrinsic model signals can drive effective learning across domains, offering a scalable alternative to RLVR for autonomous AI systems where verifiable rewards are unavailable. Code is available at https://github.com/sunblaze-ucb/Intuitor
2025-05-26T07:01:06Z
null
null
null
null
null
null
null
null
null
null
2,505.19641
SynLogic: Synthesizing Verifiable Reasoning Data at Scale for Learning Logical Reasoning and Beyond
['Junteng Liu', 'Yuanxiang Fan', 'Zhuo Jiang', 'Han Ding', 'Yongyi Hu', 'Chi Zhang', 'Yiqi Shi', 'Shitong Weng', 'Aili Chen', 'Shiqi Chen', 'Yunan Huang', 'Mozhi Zhang', 'Pengyu Zhao', 'Junjie Yan', 'Junxian He']
['cs.AI', 'cs.CL']
Recent advances such as OpenAI-o1 and DeepSeek R1 have demonstrated the potential of Reinforcement Learning (RL) to enhance reasoning abilities in Large Language Models (LLMs). While open-source replication efforts have primarily focused on mathematical and coding domains, methods and resources for developing general reasoning capabilities remain underexplored. This gap is partly due to the challenge of collecting diverse and verifiable reasoning data suitable for RL. We hypothesize that logical reasoning is critical for developing general reasoning capabilities, as logic forms a fundamental building block of reasoning. In this work, we present SynLogic, a data synthesis framework and dataset that generates diverse logical reasoning data at scale, encompassing 35 diverse logical reasoning tasks. The SynLogic approach enables controlled synthesis of data with adjustable difficulty and quantity. Importantly, all examples can be verified by simple rules, making them ideally suited for RL with verifiable rewards. In our experiments, we validate the effectiveness of RL training on the SynLogic dataset based on 7B and 32B models. SynLogic leads to state-of-the-art logical reasoning performance among open-source datasets, surpassing DeepSeek-R1-Distill-Qwen-32B by 6 points on BBEH. Furthermore, mixing SynLogic data with mathematical and coding tasks improves the training efficiency of these domains and significantly enhances reasoning generalization. Notably, our mixed training model outperforms DeepSeek-R1-Zero-Qwen-32B across multiple benchmarks. These findings position SynLogic as a valuable resource for advancing the broader reasoning capabilities of LLMs. We open-source both the data synthesis pipeline and the SynLogic dataset at https://github.com/MiniMax-AI/SynLogic.
2025-05-26T07:59:36Z
null
null
null
null
null
null
null
null
null
null
2,505.1965
Modality Curation: Building Universal Embeddings for Advanced Multimodal Information Retrieval
['Fanheng Kong', 'Jingyuan Zhang', 'Yahui Liu', 'Hongzhi Zhang', 'Shi Feng', 'Xiaocui Yang', 'Daling Wang', 'Yu Tian', 'Victoria W.', 'Fuzheng Zhang', 'Guorui Zhou']
['cs.CV', 'cs.IR', 'cs.MM']
Multimodal information retrieval (MIR) faces inherent challenges due to the heterogeneity of data sources and the complexity of cross-modal alignment. While previous studies have identified modal gaps in feature spaces, a systematic approach to address these challenges remains unexplored. In this work, we introduce UNITE, a universal framework that tackles these challenges through two critical yet underexplored aspects: data curation and modality-aware training configurations. Our work provides the first comprehensive analysis of how modality-specific data properties influence downstream task performance across diverse scenarios. Moreover, we propose Modal-Aware Masked Contrastive Learning (MAMCL) to mitigate the competitive relationships among the instances of different modalities. Our framework achieves state-of-the-art results on multiple multimodal retrieval benchmarks, outperforming existing methods by notable margins. Through extensive experiments, we demonstrate that strategic modality curation and tailored training protocols are pivotal for robust cross-modal representation learning. This work not only advances MIR performance but also provides a foundational blueprint for future research in multimodal systems. Our project is available at https://friedrichor.github.io/projects/UNITE.
2025-05-26T08:09:44Z
26 pages, project page: https://friedrichor.github.io/projects/UNITE
null
null
null
null
null
null
null
null
null
2,505.19706
Error Typing for Smarter Rewards: Improving Process Reward Models with Error-Aware Hierarchical Supervision
['Tej Deep Pala', 'Panshul Sharma', 'Amir Zadeh', 'Chuan Li', 'Soujanya Poria']
['cs.CL', 'cs.AI']
Large Language Models (LLMs) are prone to hallucination, especially during multi-hop and reasoning-intensive tasks such as mathematical problem solving. While Outcome Reward Models verify only final answers, Process Reward Models (PRMs) score each intermediate step to steer generation toward coherent solutions. We introduce PathFinder-PRM, a novel hierarchical, error-aware discriminative PRM that first classifies math and consistency errors at each step, then combines these fine-grained signals to estimate step correctness. To train PathFinder-PRM, we construct a 400K-sample dataset by enriching the human-annotated PRM800K corpus and RLHFlow Mistral traces with three-dimensional step-level labels. On PRMBench, PathFinder-PRM achieves a new state-of-the-art PRMScore of 67.7, outperforming the prior best (65.5) while using 3 times less data. When applied to reward guided greedy search, our model yields prm@8 48.3, a +1.5 point gain over the strongest baseline. These results demonstrate that decoupled error detection and reward estimation not only boost fine-grained error detection but also substantially improve end-to-end, reward-guided mathematical reasoning with greater data efficiency.
2025-05-26T08:56:36Z
https://github.com/declare-lab/PathFinder-PRM
null
null
Error Typing for Smarter Rewards: Improving Process Reward Models with Error-Aware Hierarchical Supervision
['Tej Deep Pala', 'Panshul Sharma', 'Amir Zadeh', 'Chuan Li', 'Soujanya Poria']
2,025
arXiv.org
0
20
['Computer Science']
2,505.19743
Token-level Accept or Reject: A Micro Alignment Approach for Large Language Models
['Yang Zhang', 'Yu Yu', 'Bo Tang', 'Yu Zhu', 'Chuxiong Sun', 'Wenqiang Wei', 'Jie Hu', 'Zipeng Xie', 'Zhiyu Li', 'Feiyu Xiong', 'Edward Chung']
['cs.CL', 'cs.LG']
With the rapid development of Large Language Models (LLMs), aligning these models with human preferences and values is critical to ensuring ethical and safe applications. However, existing alignment techniques such as RLHF or DPO often require direct fine-tuning on LLMs with billions of parameters, resulting in substantial computational costs and inefficiencies. To address this, we propose Micro token-level Accept-Reject Aligning (MARA) approach designed to operate independently of the language models. MARA simplifies the alignment process by decomposing sentence-level preference learning into token-level binary classification, where a compact three-layer fully-connected network determines whether candidate tokens are "Accepted" or "Rejected" as part of the response. Extensive experiments across seven different LLMs and three open-source datasets show that MARA achieves significant improvements in alignment performance while reducing computational costs. The source code and implementation details are publicly available at https://github.com/IAAR-Shanghai/MARA, and the trained models are released at https://huggingface.co/IAAR-Shanghai/MARA_AGENTS.
2025-05-26T09:24:36Z
Accepted to 34th International Joint Conference on Artificial Intelligence (IJCAI 2025)
null
null
null
null
null
null
null
null
null
2,505.19789
What Can RL Bring to VLA Generalization? An Empirical Study
['Jijia Liu', 'Feng Gao', 'Bingwen Wei', 'Xinlei Chen', 'Qingmin Liao', 'Yi Wu', 'Chao Yu', 'Yu Wang']
['cs.LG']
Large Vision-Language Action (VLA) models have shown significant potential for embodied AI. However, their predominant training via supervised fine-tuning (SFT) limits generalization due to susceptibility to compounding errors under distribution shifts. Reinforcement learning (RL) offers a path to overcome these limitations by optimizing for task objectives via trial-and-error, yet a systematic understanding of its specific generalization benefits for VLAs compared to SFT is lacking. To address this, our study introduces a comprehensive benchmark for evaluating VLA generalization and systematically investigates the impact of RL fine-tuning across diverse visual, semantic, and execution dimensions. Our extensive experiments reveal that RL fine-tuning, particularly with PPO, significantly enhances generalization in semantic understanding and execution robustness over SFT, while maintaining comparable visual robustness. We identify PPO as a more effective RL algorithm for VLAs than LLM-derived methods like DPO and GRPO. We also develop a simple recipe for efficient PPO training on VLAs, and demonstrate its practical utility for improving VLA generalization. The project page is at https://rlvla.github.io
2025-05-26T10:19:26Z
null
null
null
What Can RL Bring to VLA Generalization? An Empirical Study
['Jijia Liu', 'Feng Gao', 'Bingwen Wei', 'Xinlei Chen', 'Qingmin Liao', 'Yi Wu', 'Chaoyang Yu', 'Yu Wang']
2,025
arXiv.org
0
75
['Computer Science']
2,505.19819
FinLoRA: Benchmarking LoRA Methods for Fine-Tuning LLMs on Financial Datasets
['Dannong Wang', 'Jaisal Patel', 'Daochen Zha', 'Steve Y. Yang', 'Xiao-Yang Liu']
['cs.CE', 'cs.AI']
Low-rank adaptation (LoRA) methods show great potential for scaling pre-trained general-purpose Large Language Models (LLMs) to hundreds or thousands of use scenarios. However, their efficacy in high-stakes domains like finance is rarely explored, e.g., passing CFA exams and analyzing SEC filings. In this paper, we present the open-source FinLoRA project that benchmarks LoRA methods on both general and highly professional financial tasks. First, we curated 19 datasets covering diverse financial applications; in particular, we created four novel XBRL analysis datasets based on 150 SEC filings. Second, we evaluated five LoRA methods and five base LLMs. Finally, we provide extensive experimental results in terms of accuracy, F1, and BERTScore and report computational cost in terms of time and GPU memory during fine-tuning and inference stages. We find that LoRA methods achieved substantial performance gains of 36\% on average over base models. Our FinLoRA project provides an affordable and scalable approach to democratize financial intelligence to the general public. Datasets, LoRA adapters, code, and documentation are available at https://github.com/Open-Finance-Lab/FinLoRA
2025-05-26T10:58:51Z
null
null
null
null
null
null
null
null
null
null
2,505.1984
One Surrogate to Fool Them All: Universal, Transferable, and Targeted Adversarial Attacks with CLIP
['Binyan Xu', 'Xilin Dai', 'Di Tang', 'Kehuan Zhang']
['cs.CR', 'cs.LG', '68T07', 'I.2.6']
Deep Neural Networks (DNNs) have achieved widespread success yet remain prone to adversarial attacks. Typically, such attacks either involve frequent queries to the target model or rely on surrogate models closely mirroring the target model -- often trained with subsets of the target model's training data -- to achieve high attack success rates through transferability. However, in realistic scenarios where training data is inaccessible and excessive queries can raise alarms, crafting adversarial examples becomes more challenging. In this paper, we present UnivIntruder, a novel attack framework that relies solely on a single, publicly available CLIP model and publicly available datasets. By using textual concepts, UnivIntruder generates universal, transferable, and targeted adversarial perturbations that mislead DNNs into misclassifying inputs into adversary-specified classes defined by textual concepts. Our extensive experiments show that our approach achieves an Attack Success Rate (ASR) of up to 85% on ImageNet and over 99% on CIFAR-10, significantly outperforming existing transfer-based methods. Additionally, we reveal real-world vulnerabilities, showing that even without querying target models, UnivIntruder compromises image search engines like Google and Baidu with ASR rates up to 84%, and vision language models like GPT-4 and Claude-3.5 with ASR rates up to 80%. These findings underscore the practicality of our attack in scenarios where traditional avenues are blocked, highlighting the need to reevaluate security paradigms in AI applications.
2025-05-26T11:25:00Z
21 pages, 15 figures, 18 tables. To appear in the Proceedings of The ACM Conference on Computer and Communications Security (CCS), 2025
null
null
One Surrogate to Fool Them All: Universal, Transferable, and Targeted Adversarial Attacks with CLIP
['Binyan Xu', 'Xilin Dai', 'Di Tang', 'Kehuan Zhang']
2,025
arXiv.org
0
68
['Computer Science']
2,505.19897
ScienceBoard: Evaluating Multimodal Autonomous Agents in Realistic Scientific Workflows
['Qiushi Sun', 'Zhoumianze Liu', 'Chang Ma', 'Zichen Ding', 'Fangzhi Xu', 'Zhangyue Yin', 'Haiteng Zhao', 'Zhenyu Wu', 'Kanzhi Cheng', 'Zhaoyang Liu', 'Jianing Wang', 'Qintong Li', 'Xiangru Tang', 'Tianbao Xie', 'Xiachong Feng', 'Xiang Li', 'Ben Kao', 'Wenhai Wang', 'Biqing Qi', 'Lingpeng Kong', 'Zhiyong Wu']
['cs.AI', 'cs.CL', 'cs.CV', 'cs.HC']
Large Language Models (LLMs) have extended their impact beyond Natural Language Processing, substantially fostering the development of interdisciplinary research. Recently, various LLM-based agents have been developed to assist scientific discovery progress across multiple aspects and domains. Among these, computer-using agents, capable of interacting with operating systems as humans do, are paving the way to automated scientific problem-solving and addressing routines in researchers' workflows. Recognizing the transformative potential of these agents, we introduce ScienceBoard, which encompasses two complementary contributions: (i) a realistic, multi-domain environment featuring dynamic and visually rich scientific workflows with integrated professional software, where agents can autonomously interact via different interfaces to accelerate complex research tasks and experiments; and (ii) a challenging benchmark of 169 high-quality, rigorously validated real-world tasks curated by humans, spanning scientific-discovery workflows in domains such as biochemistry, astronomy, and geoinformatics. Extensive evaluations of agents with state-of-the-art backbones (e.g., GPT-4o, Claude 3.7, UI-TARS) show that, despite some promising results, they still fall short of reliably assisting scientists in complex workflows, achieving only a 15% overall success rate. In-depth analysis further provides valuable insights for addressing current agent limitations and more effective design principles, paving the way to build more capable agents for scientific discovery. Our code, environment, and benchmark are at https://qiushisun.github.io/ScienceBoard-Home/.
2025-05-26T12:27:27Z
work in progress
null
null
null
null
null
null
null
null
null
2,505.19954
An Explainable Diagnostic Framework for Neurodegenerative Dementias via Reinforcement-Optimized LLM Reasoning
['Andrew Zamai', 'Nathanael Fijalkow', 'Boris Mansencal', 'Laurent Simon', 'Eloi Navet', 'Pierrick Coupe']
['cs.LG', 'cs.CL']
The differential diagnosis of neurodegenerative dementias is a challenging clinical task, mainly because of the overlap in symptom presentation and the similarity of patterns observed in structural neuroimaging. To improve diagnostic efficiency and accuracy, deep learning-based methods such as Convolutional Neural Networks and Vision Transformers have been proposed for the automatic classification of brain MRIs. However, despite their strong predictive performance, these models find limited clinical utility due to their opaque decision making. In this work, we propose a framework that integrates two core components to enhance diagnostic transparency. First, we introduce a modular pipeline for converting 3D T1-weighted brain MRIs into textual radiology reports. Second, we explore the potential of modern Large Language Models (LLMs) to assist clinicians in the differential diagnosis between Frontotemporal dementia subtypes, Alzheimer's disease, and normal aging based on the generated reports. To bridge the gap between predictive accuracy and explainability, we employ reinforcement learning to incentivize diagnostic reasoning in LLMs. Without requiring supervised reasoning traces or distillation from larger models, our approach enables the emergence of structured diagnostic rationales grounded in neuroimaging findings. Unlike post-hoc explainability methods that retrospectively justify model decisions, our framework generates diagnostic rationales as part of the inference process-producing causally grounded explanations that inform and guide the model's decision-making process. In doing so, our framework matches the diagnostic performance of existing deep learning methods while offering rationales that support its diagnostic conclusions.
2025-05-26T13:18:32Z
null
null
null
null
null
null
null
null
null
null
2,505.20046
REARANK: Reasoning Re-ranking Agent via Reinforcement Learning
['Le Zhang', 'Bo Wang', 'Xipeng Qiu', 'Siva Reddy', 'Aishwarya Agrawal']
['cs.IR', 'cs.CL']
We present REARANK, a large language model (LLM)-based listwise reasoning reranking agent. REARANK explicitly reasons before reranking, significantly improving both performance and interpretability. Leveraging reinforcement learning and data augmentation, REARANK achieves substantial improvements over baseline models across popular information retrieval benchmarks, notably requiring only 179 annotated samples. Built on top of Qwen2.5-7B, our REARANK-7B demonstrates performance comparable to GPT-4 on both in-domain and out-of-domain benchmarks and even surpasses GPT-4 on reasoning-intensive BRIGHT benchmarks. These results underscore the effectiveness of our approach and highlight how reinforcement learning can enhance LLM reasoning capabilities in reranking.
2025-05-26T14:31:48Z
null
null
null
REARANK: Reasoning Re-ranking Agent via Reinforcement Learning
['Le Zhang', 'Bo Wang', 'Xipeng Qiu', 'Siva Reddy', 'Aishwarya Agrawal']
2,025
arXiv.org
0
31
['Computer Science']
2,505.20052
Ankh3: Multi-Task Pretraining with Sequence Denoising and Completion Enhances Protein Representations
['Hazem Alsamkary', 'Mohamed Elshaffei', 'Mohamed Elkerdawy', 'Ahmed Elnaggar']
['cs.LG', 'q-bio.QM']
Protein language models (PLMs) have emerged as powerful tools to detect complex patterns of protein sequences. However, the capability of PLMs to fully capture information on protein sequences might be limited by focusing on single pre-training tasks. Although adding data modalities or supervised objectives can improve the performance of PLMs, pre-training often remains focused on denoising corrupted sequences. To push the boundaries of PLMs, our research investigated a multi-task pre-training strategy. We developed Ankh3, a model jointly optimized on two objectives: masked language modeling with multiple masking probabilities and protein sequence completion relying only on protein sequences as input. This multi-task pre-training demonstrated that PLMs can learn richer and more generalizable representations solely from protein sequences. The results demonstrated improved performance in downstream tasks, such as secondary structure prediction, fluorescence, GB1 fitness, and contact prediction. The integration of multiple tasks gave the model a more comprehensive understanding of protein properties, leading to more robust and accurate predictions.
2025-05-26T14:41:10Z
8 pages, 0 figures
null
null
Ankh3: Multi-Task Pretraining with Sequence Denoising and Completion Enhances Protein Representations
['Hazem Alsamkary', 'Mohamed Elshaffei', 'Mohamed Elkerdawy', 'Ahmed Elnaggar']
2,025
arXiv.org
0
21
['Computer Science', 'Biology']
2,505.20156
HunyuanVideo-Avatar: High-Fidelity Audio-Driven Human Animation for Multiple Characters
['Yi Chen', 'Sen Liang', 'Zixiang Zhou', 'Ziyao Huang', 'Yifeng Ma', 'Junshu Tang', 'Qin Lin', 'Yuan Zhou', 'Qinglin Lu']
['cs.CV']
Recent years have witnessed significant progress in audio-driven human animation. However, critical challenges remain in (i) generating highly dynamic videos while preserving character consistency, (ii) achieving precise emotion alignment between characters and audio, and (iii) enabling multi-character audio-driven animation. To address these challenges, we propose HunyuanVideo-Avatar, a multimodal diffusion transformer (MM-DiT)-based model capable of simultaneously generating dynamic, emotion-controllable, and multi-character dialogue videos. Concretely, HunyuanVideo-Avatar introduces three key innovations: (i) A character image injection module is designed to replace the conventional addition-based character conditioning scheme, eliminating the inherent condition mismatch between training and inference. This ensures the dynamic motion and strong character consistency; (ii) An Audio Emotion Module (AEM) is introduced to extract and transfer the emotional cues from an emotion reference image to the target generated video, enabling fine-grained and accurate emotion style control; (iii) A Face-Aware Audio Adapter (FAA) is proposed to isolate the audio-driven character with latent-level face mask, enabling independent audio injection via cross-attention for multi-character scenarios. These innovations empower HunyuanVideo-Avatar to surpass state-of-the-art methods on benchmark datasets and a newly proposed wild dataset, generating realistic avatars in dynamic, immersive scenarios.
2025-05-26T15:57:27Z
null
null
null
null
null
null
null
null
null
null