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。
902 篇论文 · 按点赞排序
Shenghai Yuan, Yuanyang Yin, Zongjian Li +3 authors
Helios is a 14 billion parameter autoregressive diffusion model for video generation that achieves real-time performance and high-quality long-video synthesis without conventional optimization techniques.
Wei Zhou, Jun Zhou, Haoyu Wang +16 authors
LLM-enhanced data preparation methods are transforming data-centric workflows from rule-based pipelines to prompt-driven, context-aware approaches, organized into data cleaning, integration, and enrichment tasks.
Junxiang Xu, Ruisi Wang, Fanyi Pu +49 authors
VBBR-Pro introduces a closed-loop testbed that enables scalable, verifiable, and controllable native visual reasoning through generation across diverse visual substrates.
Kimi Team, Guangyu Chen, Yu Zhang +34 authors
Residual connections with PreNorm are standard in modern LLMs, yet they accumulate all layer outputs with fixed unit weights. This uniform aggregation causes uncontrolled hidden-state growth with depth, progressively diluting each layer's contribution. We propose Attention Residuals (AttnRes), which replaces this fixed accumulation with softmax attention over preceding layer outputs, allowing each layer to selectively aggregate earlier representations with learned, input-dependent weights. To address the memory and communication overhead of attending over all preceding layer outputs for large-scale model training, we introduce Block AttnRes, which partitions layers into blocks and attends over block-level representations, reducing the memory footprint while preserving most of the gains of full AttnRes. Combined with cache-based pipeline communication and a two-phase computation strategy, Block AttnRes becomes a practical drop-in replacement for standard residual connections with minimal overhead. Scaling law experiments confirm that the improvement is consistent across model sizes, and ablations validate the benefit of content-dependent depth-wise selection. We further integrate AttnRes into the Kimi Linear architecture (48B total / 3B activated parameters) and pre-train on 1.4T tokens, where AttnRes mitigates PreNorm dilution, yielding more uniform output magnitudes and gradient distribution across depth, and improves downstream performance across all evaluated tasks.
Haoxiang Zhang, Qinglin Chen, Hiroaki Hayashi +9 authors
Prospective learning via Self-Retrospection Distillation uses hindsight experience to train foresight predictions, improving reinforcement learning when outcome rewards are uniform.
Xinhao Ji, Hui En Pang, Hao Li +12 authors
EgoXperience introduces a unified framework for joint reconstruction of camera motion, scene geometry, hands, and objects from egocentric video using automatically annotated large-scale data and a single-pass model.
MiroMind Team, S. Bai, L. Bing +41 authors
MiroThinker-1.7 and MiroThinker-H1 are research agents that enhance complex reasoning through structured planning, contextual reasoning, and tool interaction, with MiroThinker-H1 incorporating verification at local and global levels for more reliable multi-step problem solving.
Yujia Zhang, Xiaoyang Wu, Yunhan Yang +6 authors
Utonia enables cross-domain point cloud representation learning through a unified self-supervised transformer encoder, enhancing perception and supporting embodied and multimodal reasoning tasks.
Zihao Liu, Xiaolong Shen, Zhenglin Zhou +2 authors
Latent-to-4D enables reusable direct 4D generation from video diffusion latents via alignment with a pretrained decoder and spatiotemporal attention, transferring across generators without retraining.
Hao Li, MeiJia Chen, Weijie Ren +4 authors
RIDE improves on-policy distillation by extrapolating reinforcement-learning-induced representation shifts directly in hidden states, yielding stable gains over RL-trained teachers.
Junlin Yang, Che Jiang, Yu Fu +21 authors
OpenMLE enables recursive self-improvement in machine learning engineering via a full-stack system combining execution-grounded operator learning and long-horizon evolutionary search, significantly boosting benchmark performance with a 35B model.
Yoav HaCohen, Benny Brazowski, Nisan Chiprut +26 authors
LTX-2 is an open-source audiovisual diffusion model that generates synchronized video and audio content using a dual-stream transformer architecture with cross-modal attention and classifier-free guidance.
Meituan LongCat Team, Anchun Gui, Bei Li +159 authors
A 560-billion-parameter Mixture-of-Experts reasoning model achieves state-of-the-art performance on agentic benchmarks through a unified training framework combining domain-parallel expert training with fusion, along with enhancements for real-world robustness and complex reasoning.
Tengyue Xu, Zhuoyang Qian, Gaoge Liu +16 authors
Offline knowledge construction through structured methodological graphs enables more reliable and scalable autonomous scientific discovery by reducing reliance on real-time literature processing.
Tencent Robotics X, HY Vision Team, Xumin Yu +19 authors
HY-Embodied-0.5 is a foundation model family for embodied agents featuring Mixture-of-Transformers architecture and iterative post-training for enhanced visual perception and reasoning capabilities.
Xin Li, Hao Jiang, Xin Gao +6 authors
Multi-teacher on-policy distillation merges specialist language models but suffers from unbalanced feedback across domains, which domain-normalized rescaling fixes to recover lost skills.
Chuyan Chen, Haoxing Chen, Kun Chen +27 authors
LLaDA-Image unifies a 6B diffusion transformer with a frozen vision-language module, using image-only pre-training and a Muon optimizer to generate photorealistic images with precise editing, and is distilled into a fast 2-4 step variant that achieves state-of-the-art open-source results.
Yijiang Li, Bingyang Wang, Yijun Liang +3 authors
Unsupervised on-policy self-distillation improves large language models by using internal consistency and majority-vote pseudo-solutions to correct confident errors without external supervision.
Hanyu Guo, Jiedong Yang, Chao Chen +3 authors
TransitLM dataset enables end-to-end transit route planning using large language models trained on structured transit data, eliminating the need for traditional map-based approaches.
Chenxu Yang, Chuanyu Qin, Qingyi Si +7 authors
RLSD combines reinforcement learning with verifiable rewards and self-distillation to achieve stable training with fine-grained updates and reliable policy direction from environmental feedback.
Mind Lab, Vin Bo, Asher Cai +73 authors
Macaron-V1 is an open agent-model family that uses a Mixture-of-LoRA architecture and recursive self-improvement to enable continual learning and collaboration across specialized tasks.
Ye Jin, Yangyang Xu, Jun Zhu +1 authors
MemSlides presents a hierarchical memory framework for personalized presentation agents that separates long-term user profiles, working memory for session constraints, and tool memory for reusable execution experiences to enable stable personalization and reliable local edits across multi-turn revisions.
Chengsong Huang, Zifeng Wang, Rujun Han +14 authors
EnvHarness and EnvRigger dynamically reshape static environments via programmable plugins to target agent weaknesses and improve reinforcement learning co-evolution.
Team Kandinsky, Julia Agafonova, Bulat Akhmatov +85 authors
Kandinsky 6.0 Video is a family of diffusion models that generate synchronized audio-video clips with lip-sync via a dual-stream CrossDiT architecture and multi-stage training.
Haiwen Diao, Jiahao Wang, Chenjing Ding +62 authors
SenseNova-U1.5 is an 8B native unified multimodal model that performs visual understanding, reasoning, and generation without encoders or VAEs, achieving high fidelity and instruction following through patch reconstruction, curated data, expert optimization, and on-policy distillation.
Xuan He, Cong Wei, Yuhao Cheng +20 authors
VGI-bench evaluates visual reasoning in video generation models through 27 tasks, revealing limited reliability and self-correction capabilities.
Yifu Yuan, Yaoting Huang, Xianze Yao +20 authors
Embodied-R1.5 is a unified embodied foundation model that integrates embodied reasoning capabilities and achieves state-of-the-art performance on embodied vision-language benchmarks through a multi-task balanced reinforcement learning approach.
Xinhao Li, Yuhan Zhu, Xiangyu Zeng +24 authors
VideoChat3 is an open, efficient generalist video-language model that uses an inflated 3D vision transformer and adaptive frame resolution for efficient spatiotemporal modeling, alongside curated synthetic datasets, to achieve broad video understanding with only 4B parameters.
Hao Li, Xiyuan Zheng, Yuanyuan Gao +7 authors
HOLI-WORLD is an integrated ecosystem that automates the construction of large-scale learning resources from egocentric recordings to improve machine understanding of dynamic scenes and interactions.
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.
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