Explore · 每周精选
发现最受关注的研究论文,追踪研究趋势,订阅感兴趣的期刊与关键词。
606 篇论文 · 按点赞排序
Elisei Rykov, Kseniia Petrushina, Maksim Savkin +6 authors
PsiloQA, a multilingual dataset with span-level hallucinations, enhances hallucination detection in large language models across 14 languages using an automated pipeline and encoder-based models.
David Bensaïd, Noam Rotstein, Roy Velich +2 authors
SingLoRA, a reformulated low-rank adaptation method, enhances parameter-efficient fine-tuning by learning a single low-rank matrix and its transpose, ensuring stable optimization and reducing parameter count.
Luoxin Chen, Jinming Gu, Liankai Huang +33 authors
Seed-Prover, a lemma-style reasoning model using Lean, achieves high performance in formal theorem proving and automated mathematical reasoning through iterative refinement and specialized geometry support.
Shiyin Lu, Yang Li, Yu Xia +39 authors
Ovis2.5, a native-resolution vision transformer with multimodal reasoning, achieves state-of-the-art performance on various benchmarks through advanced training techniques and efficient scaling methods.
Jizhan Fang, Xinle Deng, Haoming Xu +9 authors
LightMem, a memory system inspired by human memory, enhances LLMs by efficiently managing historical interaction information, improving accuracy and reducing computational costs.
Yuqing Li, Jiangnan Li, Zheng Lin +5 authors
MiA-RAG enhances long-context reasoning in LLM-based RAG systems by incorporating hierarchical summarization to create global semantic representations that guide both retrieval and generation processes.
Jiachen Lei, Keli Liu, Julius Berner +4 authors
Pixel-space generative models are often more difficult to train and generally underperform compared to their latent-space counterparts, leaving a persistent performance and efficiency gap. In this paper, we introduce a novel two-stage training framework that closes this gap for pixel-space diffusion and consistency models. In the first stage, we pre-train encoders to capture meaningful semantics from clean images while aligning them with points along the same deterministic sampling trajectory, which evolves points from the prior to the data distribution. In the second stage, we integrate the encoder with a randomly initialized decoder and fine-tune the complete model end-to-end for both diffusion and consistency models. Our training framework demonstrates strong empirical performance on ImageNet dataset. Specifically, our diffusion model reaches an FID of 2.04 on ImageNet-256 and 2.35 on ImageNet-512 with 75 number of function evaluations (NFE), surpassing prior pixel-space methods by a large margin in both generation quality and efficiency while rivaling leading VAE-based models at comparable training cost. Furthermore, on ImageNet-256, our consistency model achieves an impressive FID of 8.82 in a single sampling step, significantly surpassing its latent-space counterpart. To the best of our knowledge, this marks the first successful training of a consistency model directly on high-resolution images without relying on pre-trained VAEs or diffusion models.
Kuang-Huei Lee, Ian Fischer, Yueh-Hua Wu +4 authors
Mind Evolution, an evolutionary search strategy using a language model, outperforms other inference methods in natural language planning tasks by generating, recombining, and refining candidate responses.
LASA Team, Weiwen Xu, Hou Pong Chan +16 authors
A medical-specialized multimodal large language model, Lingshu, is introduced with enhanced data curation and reinforcement learning to address limitations in medical applications.
Chengpeng Li, Mingfeng Xue, Zhenru Zhang +7 authors
START integrates external tools into large reasoning models to enhance capabilities, using techniques like Hint-infer and Hint Rejection Sampling Fine-Tuning, achieving high performance across various benchmarks.
Alan Arazi, Eilam Shapira, Roi Reichart
TabSTAR, a tabular foundation model with semantically target-aware representations, achieves state-of-the-art performance in classification tasks with text features through transfer learning without dataset-specific parameters.
Ryan Wong, Jiawei Wang, Junjie Zhao +10 authors
WideSearch is a new benchmark evaluating the reliability of automated search agents in large-scale information collection tasks, revealing significant deficiencies in current systems.
Shaolei Zhang, Ju Fan, Meihao Fan +2 authors
DeepAnalyze-8B, an agentic LLM, autonomously completes the data science pipeline from raw data to research reports using curriculum-based training and data-grounded trajectory synthesis.
Weizhou Shen, Ziyi Yang, Chenliang Li +11 authors
QwenLong-L1.5 enhances long-context reasoning through data synthesis, stabilized reinforcement learning, and memory-augmented architecture, achieving superior performance on benchmarks and general domains.
Yifan Jiang, Yueying Wang, Rui Zhao +4 authors
VideoP2R, a process-aware reinforcement fine-tuning framework, improves video reasoning and understanding by modeling perception and reasoning separately, achieving state-of-the-art results on multiple benchmarks.
Alexey Gorbatovski, Boris Shaposhnikov, Viacheslav Sinii +2 authors
Direct Alignment Algorithms improve language model alignment by introducing a supervised fine-tuning phase and adjusting preference optimization strength, showing that ranking objectives are crucial for performance.
Zhuofeng Li, Haoxiang Zhang, Seungju Han +6 authors
AgentFlow, a trainable agentic framework with in-the-flow optimization, enhances reasoning in large language models by coordinating specialized modules and outperforms top baselines across various tasks.
Chulun Zhou, Chunkang Zhang, Guoxin Yu +4 authors
Multi-step retrieval-augmented generation (RAG) has become a widely adopted strategy for enhancing large language models (LLMs) on tasks that demand global comprehension and intensive reasoning. Many RAG systems incorporate a working memory module to consolidate retrieved information. However, existing memory designs function primarily as passive storage that accumulates isolated facts for the purpose of condensing the lengthy inputs and generating new sub-queries through deduction. This static nature overlooks the crucial high-order correlations among primitive facts, the compositions of which can often provide stronger guidance for subsequent steps. Therefore, their representational strength and impact on multi-step reasoning and knowledge evolution are limited, resulting in fragmented reasoning and weak global sense-making capacity in extended contexts. We introduce HGMem, a hypergraph-based memory mechanism that extends the concept of memory beyond simple storage into a dynamic, expressive structure for complex reasoning and global understanding. In our approach, memory is represented as a hypergraph whose hyperedges correspond to distinct memory units, enabling the progressive formation of higher-order interactions within memory. This mechanism connects facts and thoughts around the focal problem, evolving into an integrated and situated knowledge structure that provides strong propositions for deeper reasoning in subsequent steps. We evaluate HGMem on several challenging datasets designed for global sense-making. Extensive experiments and in-depth analyses show that our method consistently improves multi-step RAG and substantially outperforms strong baseline systems across diverse tasks.
Pengcheng Jiang, Jiacheng Lin, Zhiyi Shi +31 authors
This paper presents a framework for agent and tool adaptation in agentic AI systems, clarifying design strategies and identifying open challenges for improving AI capabilities.
Jie Jiang, Qi Yang, Bolin Ni +3 authors
R-4B, an auto-thinking multimodal large language model, uses bi-mode annealing and Bi-mode Policy Optimization to adaptively decide on problem-solving strategies, achieving state-of-the-art performance with lower computational cost.
Zhaoyang Liu, JingJing Xie, Zichen Ding +18 authors
ScaleCUA, a large-scale dataset and model for computer use agents, achieves state-of-the-art performance across multiple platforms and tasks by leveraging data-driven scaling.
Xiangyu Zhao, Junming Lin, Tianhao Liang +11 authors
Existing Multimodal Large Language Models show performance deficits in long-chain reflective reasoning, which is addressed by developing MM-HELIX-100K and Adaptive Hybrid Policy Optimization, leading to improved accuracy and generalization.
Peng Xia, Kaide Zeng, Jiaqi Liu +5 authors
Agent0, a self-evolving framework utilizing multi-step co-evolution and tool integration, enhances LLM reasoning capabilities without human-curated data.
Wenqi Zhang, Hang Zhang, Xin Li +6 authors
A new high-quality instructional video-based corpus improves VLM pretraining through more coherent context and richer knowledge, enhancing performance in knowledge-intensive tasks.
Tianyu Fu, Yichen You, Zekai Chen +3 authors
Think-at-Hard (TaH) dynamically refines only hard tokens in LLMs using a neural decider and LoRA, improving reasoning performance with minimal additional parameters or iterations.
Karan Dalal, Daniel Koceja, Gashon Hussein +12 authors
Test-Time Training (TTT) layers enable pre-trained Transformers to generate coherent one-minute videos from text storyboards, outperforming alternatives like Mamba~2 and Gated DeltaNet.
Gleb Rodionov, Roman Garipov, Alina Shutova +5 authors
A parallel inference engine for LLMs allows concurrent task collaboration through a shared attention cache, improving efficiency without additional fine-tuning.
Wei Pang, Kevin Qinghong Lin, Xiangru Jian +2 authors
A benchmark and metric suite for poster generation evaluates visual quality, coherence, and content accuracy, leading to a multi-agent pipeline that outperforms existing models with reduced computational cost.
M-A-P Team, Xinrun Du, Yifan Yao +92 authors
SuperGPQA, a benchmark evaluating LLMs across 285 disciplines, reveals performance gaps and offers insights into the collaborative filtering process with human experts.
Guanting Dong, Licheng Bao, Zhongyuan Wang +11 authors
AEPO, an agentic RL algorithm, addresses entropy-related challenges in web agent training, enhancing performance and stability across various datasets.
北京市昌平区探索星信息技术及软件开发工作室
京ICP备2026059466号