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。
50 篇论文 · 按点赞排序
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.
Eunbi Choi, Kibong Choi, Seokhee Hong +62 authors
K-EXAONE is a multilingual language model with a Mixture-of-Experts architecture that achieves competitive performance on various benchmarks while supporting multiple languages and long-context windows.
Yuhang Wang, Yuling Shi, Mo Yang +7 authors
SWE-Pruner is a self-adaptive context pruning framework for coding agents that uses task-aware pruning to reduce token usage while maintaining performance.
Taofeng Xue, Chong Peng, Mianqiu Huang +13 authors
EvoCUA introduces an evolutionary approach to computer-use agents that combines autonomous task generation with policy optimization to achieve superior performance in complex, long-horizon tasks.
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.
Haochen Shi, Xingdi Yuan, Bang Liu
Programmatic Skill Network enables continual skill acquisition through executable symbolic programs that evolve via reflection, progressive optimization, and structural refactoring mechanisms.
Daixuan Cheng, Shaohan Huang, Yuxian Gu +6 authors
LLM-in-Sandbox enables large language models to perform general intelligence tasks across diverse domains by allowing them to explore a code sandbox environment, achieving robust generalization without additional training.
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.
Amirhosein Ghasemabadi, Di Niu
Large language models (LLMs) generate fluent and complex outputs but often fail to recognize their own mistakes and hallucinations. Existing approaches typically rely on external judges, multi-sample consistency, or text-based self-critique, which incur additional compute or correlate weakly with true correctness. We ask: can LLMs predict their own failures by inspecting internal states during inference? We introduce Gnosis, a lightweight self-awareness mechanism that enables frozen LLMs to perform intrinsic self-verification by decoding signals from hidden states and attention patterns. Gnosis passively observes internal traces, compresses them into fixed-budget descriptors, and predicts correctness with negligible inference cost, adding only ~5M parameters and operating independently of sequence length. Across math reasoning, open-domain question answering, and academic knowledge benchmarks, and over frozen backbones ranging from 1.7B to 20B parameters, Gnosis consistently outperforms strong internal baselines and large external judges in both accuracy and calibration. Moreover, it generalizes zero-shot to partial generations, enabling early detection of failing trajectories and compute-aware control. These results show that reliable correctness cues are intrinsic to generation process and can be extracted efficiently without external supervision.
Zichen Wen, Boxue Yang, Shuang Chen +31 authors
Innovator-VL demonstrates that principled training design and transparent methodology can achieve strong scientific intelligence with reduced data requirements while maintaining general vision performance.
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.
Hangrui Hu, Xinfa Zhu, Ting He +13 authors
The Qwen3-TTS series presents advanced multilingual text-to-speech models with voice cloning and controllable speech generation capabilities, utilizing dual-track LM architecture and specialized speech tokenizers for efficient streaming synthesis.
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.
Hao Luo, Ye Wang, Wanpeng Zhang +9 authors
Being-H0.5 is a Vision-Language-Action model that enables robust cross-embodiment generalization through human-centric learning and a Mixture-of-Transformers architecture with specialized embodiment handling.
Haozhe Xie, Beichen Wen, Jiarui Zheng +4 authors
DynamicVLA addresses dynamic object manipulation challenges through a compact vision-language-action model with temporal reasoning and closed-loop adaptation, supported by a new benchmark for dynamic manipulation tasks.
Zanlin Ni, Shenzhi Wang, Yang Yue +8 authors
Arbitrary order generation in diffusion large language models limits reasoning capability by causing premature solution space collapse, making standard policy optimization more effective.
Haowei Zhang, Shudong Yang, Jinlan Fu +2 authors
HERMES is a training-free architecture that enables real-time video stream understanding by utilizing a hierarchical memory framework based on KV cache reuse, achieving faster response times and maintained accuracy even with reduced video token input.
Haoran Wei, Yaofeng Sun, Yukun Li
DeepSeek-OCR 2 introduces DeepEncoder V2 that dynamically reorders visual tokens based on semantic content, enabling more human-like causal reasoning in 2D image understanding through cascaded 1D causal structures.
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.
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