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。
Chengwen Liu, Xiaomin Yu, Zhuoyue Chang +13 authors
VideoDR benchmark enables video question answering by combining cross-frame visual extraction, web retrieval, and multi-hop reasoning in open-domain settings.
50 篇论文 · 按点赞排序
Liang Chen, Weichu Xie, Yiyan Liang +26 authors
Current multimodal large language models exhibit significant gaps in fundamental visual understanding compared to human children, as demonstrated by the BabyVision benchmark.
Ailin Huang, Chengyuan Yao, Chunrui Han +90 authors
STEP3-VL-10B achieves superior multimodal performance through unified pre-training with a language-aligned Perception Encoder and Qwen3-8B decoder, combined with scaled post-training and Parallel Coordinated Reasoning for efficient large-scale visual reasoning.
Yuxiang Ji, Yong Wang, Ziyu Ma +6 authors
Large vision-language models are enhanced for image geolocalization by incorporating map-based reasoning and agent-in-the-map loop optimization, achieving superior accuracy compared to existing models.
Yu Wang, Yi Wang, Rui Dai +4 authors
Urban socio-semantic segmentation is achieved through a vision-language model framework that combines cross-modal recognition and multi-stage reasoning with reinforcement learning optimization.
Zhiyuan Hu, Yucheng Wang, Yufei He +7 authors
Reinforcement learning for large language models is enhanced by a rollout-level objective that rewards rare high-level reasoning strategies, improving diverse solution discovery without sacrificing initial performance.
Yibo Wang, Lei Wang, Yue Deng +7 authors
DeepResearchEval presents an automated framework for creating complex research tasks and evaluating them through agent-based methods that adapt to task specifics and verify facts without relying on citations.
Tu Hu, Ronghao Chen, Shuo Zhang +9 authors
Controlled Self-Evolution method improves code generation through diversified initialization, feedback-guided genetic evolution, and hierarchical memory to enhance exploration efficiency and solution quality.
Jing Xiong, Qi Han, Yunta Hsieh +11 authors
MMFormalizer enables multimodal autoformalization by integrating visual perception with formal mathematical reasoning, supporting complex physical domains from classical mechanics to quantum mechanics.
Jian Zhang, Zhiyuan Wang, Zhangqi Wang +7 authors
MAXS is a meta-adaptive reasoning framework for LLM agents that improves multi-tool reasoning through lookahead strategies and trajectory convergence mechanisms, balancing global effectiveness and computational efficiency.
Zhiyuan Hu, Yunhai Hu, Juncheng Liu +9 authors
Multi-Agent Test-Time Reinforcement Learning (MATTRL) enhances multi-agent reasoning through structured textual experience injection and consensus-based decision making at inference time.
Jingcheng Hu, Yinmin Zhang, Shijie Shang +17 authors
Parallel Coordinated Reasoning enables large-scale test-time compute scaling beyond sequential reasoning limitations through parallel exploration and message-passing architecture.
Jian Zhang, Yu He, Zhiyuan Wang +5 authors
Scientific reasoning relies not only on logical inference but also on activating prior knowledge and experiential structures. Memory can efficiently reuse knowledge and enhance reasoning consistency and stability. However, existing benchmarks mainly evaluate final answers or step-by-step coherence, overlooking the memory-driven mechanisms that underlie human reasoning, which involves activating anchors and attractors, then integrating them into multi-step inference. To address this gap, we propose A^3-Bench~ https://a3-bench.github.io, a benchmark designed to evaluate scientific reasoning through dual-scale memory-driven activation, grounded in Anchor and Attractor Activation. First, we annotate 2,198 science reasoning problems across domains using the SAPM process(subject, anchor & attractor, problem, and memory developing). Second, we introduce a dual-scale memory evaluation framework utilizing anchors and attractors, along with the AAUI(Anchor--Attractor Utilization Index) metric to measure memory activation rates. Finally, through experiments with various base models and paradigms, we validate A^3-Bench and analyze how memory activation impacts reasoning performance, providing insights into memory-driven scientific reasoning.
Qihao Wang, Ziming Cheng, Shuo Zhang +12 authors
MemGovern framework transforms unstructured GitHub data into structured experiential memory for autonomous software engineering agents, improving bug resolution rates through enhanced experience retrieval.
Lizhen Wang, Yongming Zhu, Zhipeng Ge +15 authors
FlowAct-R1 enables real-time interactive humanoid video generation with high-fidelity synthesis and low-latency responsiveness through MMDiT architecture and chunkwise diffusion forcing strategies.
Xindi Wu, Despoina Paschalidou, Jun Gao +5 authors
Motive is a gradient-based data attribution framework that identifies influential video clips for motion improvement in text-to-video models through motion-weighted loss masking.
Sungrae Park, Sanghoon Kim, Jungho Cho +34 authors
Solar Open presents a 102B-parameter bilingual Mixture-of-Experts language model that addresses data scarcity in underserved languages through synthetic data generation, progressive curriculum coordination, and scalable reinforcement learning optimization.
Alexander H. Liu, Kartik Khandelwal, Sandeep Subramanian +117 authors
The Ministral 3 series consists of parameter-efficient dense language models with three sizes (3B, 8B, 14B) and three variants per size, trained using cascade distillation for compute-constrained applications.
Shaotian Yan, Kaiyuan Liu, Chen Shen +6 authors
A lightweight open-source reasoning model achieves state-of-the-art performance through enhanced sequence-level distillation that addresses limitations in current teacher-student knowledge transfer methods.
Grigorii Alekseenko, Aleksandr Gordeev, Irina Tolstykh +7 authors
A compact image editing system uses a 2B-parameter model for guidance and a 1.6B-parameter diffusion model to achieve high-quality edits with low computational requirements and strict source consistency.
Qiguang Chen, Yantao Du, Ziniu Li +10 authors
Large language models struggle with long chain-of-thought reasoning due to unstable structural patterns, but a molecular-inspired approach using effective semantic isomers and distribution-transfer-graph methods improves training stability and performance.
Tingyu Wu, Zhisheng Chen, Ziyan Weng +8 authors
Long-horizon memory benchmarks based on autobiographical narratives evaluate models' ability to infer stable motivations and decision principles through evidence-linked questions spanning factual recall, subjective state attribution, and principle-level reasoning.
Mingxin Li, Yanzhao Zhang, Dingkun Long +9 authors
The Qwen3-VL-Embedding and Qwen3-VL-Reranker models form an end-to-end multimodal search pipeline, leveraging multi-stage training and cross-attention mechanisms to achieve high-precision retrieval across diverse modalities.
Chi-Pin Huang, Yunze Man, Zhiding Yu +4 authors
Fast-ThinkAct is an efficient vision-language-action framework that reduces inference latency by 89.3% through compact latent reasoning while maintaining long-horizon planning and few-shot adaptation capabilities.
Qiang Zhang, Boli Chen, Fanrui Zhang +14 authors
Reinforcement learning for large language model agents suffers from discrimination collapse in open-ended tasks due to pointwise scalar scoring, which ArenaRL addresses through relative ranking and pairwise evaluation mechanisms.
Eldad Matmon, Amit Bracha, Noam Rotstein +1 authors
A photorealistic 3D caricaturization framework combines Gaussian curvature-based surface exaggeration with 3D Gaussian Splatting to create controllable, realistic avatars with improved fidelity and real-time deformation capabilities.
Jungho Cho, Minbyul Jeong, Sungrae Park
Large reasoning models enable scalable multi-turn dialogue generation through automated task-oriented simulation and user-oriented behavioral modeling for enhanced human-agent interaction datasets.
Kewei Zhang, Ye Huang, Yufan Deng +5 authors
Multi-Head Linear Attention addresses the performance degradation in linear attention by preserving representational diversity through head-wise token dimension computation, maintaining linear complexity while recovering softmax attention's expressive power across multiple domains.
Dongchao Yang, Yuxin Xie, Yuguo Yin +25 authors
A suite of open-source music foundation models is introduced, featuring components for audio-text alignment, lyric recognition, music coding, and large language model-based song generation with controllable attributes and scalable parameterization.
Jiaxuan Lu, Ziyu Kong, Yemin Wang +10 authors
Test-Time Tool Evolution enables AI agents to dynamically create and refine computational tools during inference, overcoming limitations of static tool libraries in scientific applications.
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