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
Yeonjun In, Wonjoong Kim, Sangwu Park +2 authors
Large language model-based memory systems can benefit from personalized policies that adapt to individual user contexts, though accurate implementation remains challenging.
Ziwen Xu, Haiwen Hong, Linsong Yu +4 authors
Research investigates the quantitative limits of parametric memory in large language models using LoRA as a probe, establishing a power law relationship and developing a threshold-guided optimization method for improved memory performance.
Siyu An, Junru Lu, Junnan Dong +18 authors
Native multimodal modeling advances beyond traditional fusion approaches by integrating modalities inherently within a unified transformer framework, enabling seamless understanding and generation across diverse input-output configurations.
Junyan Ye, Jun He, Zilong Huang +4 authors
GenClaw presents a code-driven agentic image generation framework that enables precise visual construction through conceptualization, sketching, and coloring stages, integrating programmatic logic with generative models.
MiniMax, Aili Chen, Aonian Li +203 authors
The MiniMax-M2 series introduces Mixture-of-Experts language models with minimal activated parameters that achieve high performance in agentic tasks through specialized training and deployment systems.
Xinle Deng, Ruobin Zhong, Hujin Peng +15 authors
Memory systems in large language models suffer from reliability issues that can be addressed through a novel tracing framework and automated fault attribution for improved performance.
Jin Hyeon Kim, Jaeeun Lee, Claire Kim +8 authors
A novel diffusion-based framework for multi-view 3D reconstruction that restores both scene geometry and high-quality imagery from degraded inputs by operating in the feature space of a 3D reconstructor.
Suji Kim, Kangsan Kim, Sung Ju Hwang
LearnWeak is an annotation-free framework that enhances small computer-use agents by identifying weaknesses through a stronger reference agent and generating targeted training data for improved domain specialization.
Rui Meng, Bhavana Dalvi Mishra, Jiefeng Chen +10 authors
Autonomous research agents exhibit verifiability issues like fabricated citations and unreproducible results, which are addressed through a framework ensuring evidence traceability and an end-to-end system maintaining integrity throughout research processes.
Tengfei Liu, Yang Shi, Xuanyu Zhu +17 authors
LongAV-Compass is a comprehensive benchmark for evaluating minute-long audio-visual generation across multiple modalities, assessing quality, consistency, and alignment over extended temporal sequences.
Longbin Ji, Guan Wang, Xuan Wei +6 authors
NAVA enables joint audio-video generation with improved synchronization and controllability through native audio-visual alignment and context-conditioned denoising.
Bowen Wang, Dunjie Lu, Junli Wang +11 authors
RLVR framework for computer-use agents addresses data scarcity through scalable generation pipeline and synthetic environments, achieving superior performance on verification and transfer benchmarks.
Minju Gwak, Minseo Kwak, Dongseok Lee +3 authors
LaRA is a layer-wise representation analysis framework that detects data contamination in reinforcement learning-post-trained large language models by analyzing geometric deviations across model layers.
Jizhan Fang, Buqiang Xu, Zhixian Wang +12 authors
FluxMem is a memory framework that dynamically evolves memory topology through three stages to improve performance in complex agentic environments.
Yiren Song, Huilin Zhong, Kevin Qinghong Lin +2 authors
A multi-agent framework called Soap2Soap is presented for long-horizon video-to-video generation that maintains narrative structure and character identity across extended sequences through consistent semantic backbone and visual reference anchors.
Zuhao Yang, Kaichen Zhang, Sudong Wang +7 authors
ParaVT enables parallel video tool calling through multi-agent reinforcement learning, addressing limitations of sequential approaches and improving long-video understanding performance.
Ruowen Zhao, Bangguo Li, Zuyan Liu +9 authors
GEM is a vision-language model that integrates depth map generation during pre-training to improve embodied intelligence and physical operation capabilities in robotics.
Jiapeng Zhu, Jianxiang Yu, Yibo Zhao +5 authors
Skill0.5 is a novel agentic reinforcement learning framework that combines general skill internalization with task-specific skill utilization through a dynamic, difficulty-aware router to improve performance in complex task environments.
Funing Fu, Tenghui Wang, Junyong Cen +2 authors
Latent diffusion models using clean-data prediction outperform velocity prediction in compressed representations, demonstrating that prediction targets are geometrically dependent rather than algebraically interchangeable.
Jianing Zhu, Yeonju Ro, John Robertson +5 authors
Long-lived AI agents require lifespan evaluation and mechanism-level diagnosis beyond initial performance testing to ensure reliability over time.
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