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
Harold Haodong Chen, Rui Peng, Rongjin Guo +4 authors
World-in-Context unifies view, action, and camera controls via explicit tokens and interaction-aware positional encoding to improve interactive video generation and navigation.
Ammar Ali, Stamatis Lefkimmiatis
TWIG represents neural network weights as shared discrete decision paths, unifying compression methods and enabling efficient fixed-weight circuits with optimized discrete choices.
Yuxuan Chen, Feng Xie, Jiagao Hu +4 authors
LociArt enables precise local video stylization with region-style control and temporal consistency via a generative refinement network, automated dataset construction, and a dedicated benchmark.
Jiayue Pu, Ang Gao, Jinyi Bai +9 authors
This survey defines recursive self-improvement and organizes trustworthy advancement requirements into improvement validity, constraint preservation, and oversight across recursive update stages.
Chaehyun Kim, Chi-Pin Huang, Jisoo Kim +4 authors
ThinkWorld improves robotic generalization by representing actions as dynamics chains and using world imagination to learn from unlabeled human videos.
Aaron Zorel, Guy Van den Broeck, Aditya Grover +1 authors
This paper explores trading inference speed for output quality in speculative decoding by jointly adjusting diffusion effort and acceptance leniency to extend the speed-quality Pareto frontier.
Bowen Ding, Mingda Wan, Yuhan Chen +7 authors
ProAS improves credit assignment in critic-free reinforcement learning for LLM agents by using a dynamic reference graph to mask unreasonable incentives and redistribute penalties asymmetrically, boosting success rates and convergence speed.
Yegor Anichkov, Valery Vinogradov, Andrey Furov +2 authors
TenetBench evaluates whether language models adhere to professional ethical codes by pairing conflicting requests with benign equivalents, revealing frequent unflagged violations across most tested models.
Siyang He, Qiqi Wang, Fan Liu +5 authors
DraMaze improves long-form microdrama script generation by using evolutionary tree search guided by trope memory and backward narrative dependency trees to enhance plot diversity and coherence.
Zhen Li, Mingfa Feng, Yiming Zhang +6 authors
MiniTE enables stable low-bit reinforcement learning for large language models by aligning cross-engine numerical execution and accelerating native 4-bit computation.
Zhihao Yang, Haojin Wang, Juncheng Wu +7 authors
Harness self-evolution improves LLM agents via skill adaptation guided by a meta-skill that reformulates guidance to match executor capabilities, boosting weak models without parameter updates.
Shuo Shen, Guangran Cheng, Chengqi Lyu +3 authors
S3AC introduces a stabilized actor-critic training method with difficulty-aware prompt sampling to improve long reasoning in large language models without relying on grouped rollouts.
Arman Zarei, Sayan Nag, Bryan Wang +5 authors
SlideBench benchmarks continuous image editing and SliderEdit++ improves slider reliability via trajectory regularization and disentanglement for smoother, faster control.
Juekai Lin, Honglin Lin, Yuqian Yuan +7 authors
MMVistaReason improves open multimodal reasoning through structured data construction, difficulty-aware distillation, and specialized reinforcement learning integration.
Liam Dugan, Callum McDougall, Matthew Denton +3 authors
A pipeline using sparse autoencoders extracts interpretable linguistic features from neural detectors, yielding more accurate and explainable AI-generated text classification.
Shuyang Cui, Zhi Zhong, Keisuke Toyama +10 authors
Flow matching is investigated for four-source full-band stereo music separation, evaluating prediction targets, flow paths, modeling strategies, and consistency constraints with both signal-level and perceptual metrics.
Yufeng Liu, Bohao Fan, Juntu Zhao +11 authors
Echo integrates stop and failure signals into a vision-language-action model to guide robot execution and recovery without fine-grained progress labels.
Hongcheng Gao, Zelin Zheng, Alex Zhou +18 authors
EmbodiedLab evaluates embodied agents that must physically interact with simulated laboratory apparatus to discover scientific principles through self-directed experimentation.
Changlian Ma, Limin Wang
AutoTavern demonstrates that self-improving LLM research agents can autonomously build a faithful game simulator and train a reinforcement learning agent that surpasses top human players in Hearthstone Battlegrounds.
Jiacheng Liu, Yaxin Luo, Xiaohan Zhao +5 authors
OPTIC achieves extreme compression of per-layer embeddings by using token-shared indexed codes and layer-specific affine decoders fitted to the output path, preserving most model performance at up to 1000x compression with minimal inference overhead.
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