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。
Jian Yang, Xianglong Liu, Weifeng Lv +68 authors
A comprehensive guide to code LLMs, covering their lifecycle from data curation to deployment, including techniques, trade-offs, and research-practice gaps.
50 篇论文 · 按点赞排序
DeepSeek-AI, Aixin Liu, Aoxue Mei +261 authors
DeepSeek-V3.2 introduces DeepSeek Sparse Attention and a scalable reinforcement learning framework, achieving superior reasoning and performance compared to GPT-5 and Gemini-3.0-Pro in complex reasoning tasks.
Z-Image Team, Huanqia Cai, Sihan Cao +18 authors
Z-Image, a 6B-parameter Scalable Single-Stream Diffusion Transformer (S3-DiT) model, achieves high-performance image generation with reduced computational cost, offering sub-second inference and compatibility with consumer hardware.
Zuhao Yang, Sudong Wang, Kaichen Zhang +7 authors
LongVT, an end-to-end framework, enhances long video reasoning by interleaving global and local analysis using multimodal tools, outperforming existing methods on challenging benchmarks.
Yubo Huang, Hailong Guo, Fangtai Wu +8 authors
Live Avatar uses a 14-billion-parameter diffusion model with Timestep-forcing Pipeline Parallelism and Rolling Sink Frame Mechanism to achieve real-time, high-fidelity avatar generation.
Shuai Bai, Yuxuan Cai, Ruizhe Chen +61 authors
Qwen3-VL, a vision-language model, excels in text and multimodal understanding through advanced architectures and larger contexts, achieving superior performance across benchmarks.
Fangyu Lei, Jinxiang Meng, Yiming Huang +14 authors
DAComp is a benchmark of 210 tasks that evaluates the capabilities of agents in real-world data engineering and data analysis workflows, revealing significant deficiencies in both areas.
Hongjin Su, Shizhe Diao, Ximing Lu +13 authors
A small orchestrator using ToolOrchestra method coordinates various intelligent tools with reinforcement learning, achieving higher accuracy and efficiency in solving complex tasks like Humanity's Last Exam compared to larger models.
Chujie Zheng, Kai Dang, Bowen Yu +7 authors
The paper provides a theoretical foundation for optimizing sequence-level rewards in reinforcement learning using token-level objectives, highlighting the importance of techniques like importance sampling correction, clipping, and Routing Replay for stabilizing training, especially with large language models.
Zhihong Shao, Yuxiang Luo, Chengda Lu +6 authors
A self-verifying large language model for theorem proving improves mathematical reasoning by incentivizing rigorous step-by-step derivations and achieves high scores in international competitions.
Juanxi Tian, Siyuan Li, Conghui He +2 authors
A benchmark for chained text-to-multi-image generation assesses models' ability to model dynamic causal processes and world knowledge, revealing that unified multimodal models outperform specialized ones but still struggle with spatiotemporal consistency.
Nex-AGI Team, Yuxuan Cai, Lu Chen +62 authors
The introduction of NexAU, NexA4A, and NexGAP enables the scaling of complexity, diversity, and fidelity in interactive environments for training large language models as autonomous agents, resulting in superior performance.
Zhiheng Liu, Weiming Ren, Haozhe Liu +22 authors
TUNA, a unified multimodal model, uses a cascaded VAE and representation encoder for end-to-end multimodal understanding and generation, outperforming decoupled models and achieving state-of-the-art results across various benchmarks.
Junyi Hou, Andre Lin Huikai, Nuo Chen +2 authors
PaperDebugger is an in-editor academic writing assistant that integrates large language models, enabling direct interaction within LaTeX editors for document state management, revision, and literature search.
Alexander Amini, Anna Banaszak, Harold Benoit +30 authors
LFM2, a family of compact foundation models, achieves high efficiency and performance on-device through hardware-in-the-loop architecture search and advanced training techniques, supporting various tasks including multimodal applications.
Zhengliang Shi, Yiqun Chen, Haitao Li +23 authors
Deep Research systems integrate LLMs with external tools to enhance problem-solving capabilities, involving query planning, information acquisition, memory management, and answer generation.
Qinghe Wang, Xiaoyu Shi, Baolu Li +7 authors
MultiShotMaster extends a single-shot model with novel RoPE variants for flexible and controllable multi-shot video generation, addressing data scarcity with an automated annotation pipeline.
Dingling Zhang, He Zhu, Jincheng Ren +15 authors
FINDER is a benchmark for deep research agents with standardized human-curated tasks and DEFT is a failure taxonomy revealing that DRAs struggle with evidence integration, verification, and reasoning-resilient planning.
Wenhao Yu, Zhenwen Liang, Chengsong Huang +4 authors
R-Few, a guided Self-Play Challenger-Solver framework, enables stable and controllable model self-evolution with minimal human supervision, achieving performance improvements on math and reasoning benchmarks.
Minh-Quan Le, Yuanzhi Zhu, Vicky Kalogeiton +1 authors
A physics-grounded post-training framework using verifiable rewards improves physical realism and motion quality in video diffusion models.
Yifan Zhang, Liang Hu, Haofeng Sun +12 authors
Skywork-R1V4, a 30B parameter multimodal agentic model, achieves top performance in perception and multimodal search benchmarks through supervised fine-tuning and interleaved reasoning without reinforcement learning.
Alla Chepurova, Aydar Bulatov, Yuri Kuratov +1 authors
Wikontic is a multi-stage pipeline that constructs high-quality, ontology-consistent knowledge graphs from open-domain text, achieving state-of-the-art performance in information retention and efficiency.
Hidir Yesiltepe, Tuna Han Salih Meral, Adil Kaan Akan +2 authors
A unified inference-time framework addresses core limitations of autoregressive video diffusion models, enabling infinite-horizon, controllable, and cinematic video generation.
Xingrun Xing, Zhiyuan Fan, Jie Lou +3 authors
PretrainZero is a reinforcement active learning framework that enhances general reasoning capabilities by pretraining large models on a corpus without verifiable labels, improving performance on benchmarks compared to domain-specific training.
Shengyuan Ding, Xinyu Fang, Ziyu Liu +10 authors
ARM-Thinker, an agentic reward model, uses external tools for verification, improving accuracy and interpretability in complex multimodal reasoning tasks.
Bo Wang, Jiehong Lin, Chenzhi Liu +5 authors
MG-Nav, a dual-scale framework for zero-shot visual navigation, combines global memory-guided planning with local geometry-enhanced control using a Sparse Spatial Memory Graph and a VGGT-adapter for robust navigation in unseen environments.
Fukun Yin, Shiyu Liu, Yucheng Han +12 authors
Integrating reasoning mechanisms into image editing models enhances performance by improving instruction understanding and result correction.
Zhenxiong Tan, Zeqing Wang, Xingyi Yang +2 authors
Bridge Models, instantiated as Vision Bridge Transformer (ViBT), efficiently translate data through direct modeling of input-to-output trajectories, achieving robust performance in image and video editing tasks at large scales.
Zhizhou Zhong, Yicheng Ji, Zhe Kong +12 authors
The proposed AnyTalker framework generates high-quality multi-person talking videos by extending Diffusion Transformer with identity-aware attention, leveraging single-person videos for training, and using a specialized dataset for evaluation.
Hongfei Zhang, Kanghao Chen, Zixin Zhang +6 authors
DualCamCtrl is a diffusion model for camera-controlled video generation that uses a dual-branch framework and Semantic Guided Mutual Alignment to improve consistency and disentangle appearance and geometry modeling.
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