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。
Haoquan Fang, Jiafei Duan, Donovan Clay +26 authors
MolmoAct2 presents an open-action reasoning model for robotics that improves upon previous systems through specialized vision-language-model backbones, new datasets, open-weight action tokenizers, architectural redesign for continuous-action prediction, and adaptive reasoning for reduced latency.
50 篇论文 · 按点赞排序
Shuzheng Si, Haozhe Zhao, Yu Lei +10 authors
A self-evolving framework autonomously discovers and refines context-specific skills for language models through a multi-agent self-play loop with Challenger, Reasoner, and Judge components, improving context learning performance without human supervision.
Ruofeng Yang, Yongcan Li, Shuai Li
ARIS is an open-source research harness that uses cross-model adversarial collaboration to ensure reliable long-term research outcomes through coordinated execution, orchestration, and assurance layers.
Bin Wu, Mengqi Huang, Shaojin Wu +4 authors
Stream-R1 improves video diffusion model distillation by adaptively weighting supervision based on reliability and perplexity, enhancing visual quality, motion quality, and text alignment without additional computational overhead.
Dongyoung Kim, Huiwon Jang, Myungkyu Koo +65 authors
RLDX-1 is a general-purpose robotic policy for dexterous manipulation that integrates heterogeneous modalities through a Multi-Stream Action Transformer architecture, demonstrating superior performance in complex real-world tasks compared to existing vision-language-action models.
Zhuofeng Li, Haoxiang Zhang, Cong Wei +16 authors
Direct corpus interaction enables more effective agentic search by allowing agents to query raw text directly, outperforming traditional retrieval methods in complex tasks.
Yaorui Shi, Yuxin Chen, Zhengxi Lu +6 authors
Skill1 is a unified framework that trains a single policy to simultaneously evolve skill selection, utilization, and distillation capabilities using a shared task-outcome objective, demonstrating superior performance over existing baselines in complex task environments.
Yijing Tu, Shaojin Wu, Mengqi Huang +4 authors
Stream-T1 enables efficient test-time video generation through streaming synthesis with temporal guidance, achieving improved consistency and quality over existing diffusion-based methods.
Shuang Chen, Kaituo Feng, Hangting Chen +7 authors
OpenSearch-VL presents an open-source framework for training advanced multimodal search agents using reinforcement learning, featuring specialized data curation, diverse tool environments, and a novel training algorithm that improves performance across multiple benchmarks.
Hongcan Guo, Qinyu Zhao, Yian Zhao +8 authors
Cola DLM presents a hierarchical latent diffusion language model that uses text-to-latent mapping, global semantic prior modeling, and conditional decoding to achieve efficient text generation with flexible non-autoregressive inductive bias.
Houyuan Chen, Hong Li, Xianghao Kong +8 authors
UniVidX is a unified multimodal framework that uses video diffusion model priors for versatile video generation through stochastic condition masking, decoupled gated LoRA, and cross-modal self-attention mechanisms.
Junbo Cui, Bokai Xu, Chongyi Wang +33 authors
MiniCPM-o 4.5 enables real-time full-duplex multimodal interaction through Omni-Flow, a unified streaming framework that aligns inputs and outputs temporally for simultaneous perception and response.
Xin Zhou, Dingkang Liang, Xiwu Chen +4 authors
HERMES++ combines 3D scene understanding and future geometry prediction through BEV representation, LLM-enhanced queries, temporal linking, and joint geometric optimization for autonomous driving applications.
Yuwen Du, Rui Ye, Shuo Tang +4 authors
A simple supervised fine-tuning approach achieves state-of-the-art performance in deep search capabilities using minimal data, outperforming complex industrial pipelines and demonstrating the effectiveness of academic-led development in large language model agents.
Yuqing Li, Jiangnan Li, Mo Yu +3 authors
Researchers propose a compressed representation method for global activation patterns in large language models that approximates full activation states while maintaining computational efficiency and improving performance in long-context tasks.
Sudong Wang, Weiquan Huang, Xiaomin Yu +9 authors
PRISM addresses distributional drift in multimodal models by inserting a distribution-alignment stage between supervised fine-tuning and reinforcement learning with verifiable rewards, using a black-box adversarial game between policy and MoE discriminator for disentangled corrective signals.
Siru Ouyang, Jun Yan, Yanfei Chen +13 authors
SkillOS enables self-evolving LLM agents to learn complex long-term skill curation policies through reinforcement learning, improving performance across diverse tasks while generalizing across different executor architectures.
Ivan Bondarenko, Roman Derunets, Oleg Sedukhin +3 authors
A heterogeneous ensemble of seven large language models with dual prompting strategies achieved top performance in the SemEval-2026 MTRAGEval task through judge selection and demonstrated the importance of model diversity.
Rui Wang, Yue Zhang, Jiehong Lin +4 authors
World Action Models (WAMs) have recently emerged as a promising paradigm for robotic manipulation by jointly predicting future visual observations and future actions. However, current WAMs typically execute a fixed number of predicted actions after each model inference, leaving the robot blind to whether the imagined future remains consistent with the actual physical rollout. In this work, we formulate adaptive WAM execution as a future-reality verification problem: the robot should execute longer when the WAM-predicted future remains reliable, and replan earlier when reality deviates from imagination. To this end, we propose Future Forward Dynamics Causal Attention (FFDC), a lightweight verifier that jointly reasons over predicted future actions, predicted visual dynamics, real observations, and language instructions to estimate whether the remaining action rollout can still be trusted. FFDC enables adaptive action chunk sizes as an emergent consequence of prediction-observation consistency, preserving the efficiency of long-horizon execution while restoring responsiveness in contact-rich or difficult phases. We further introduce Mixture-of-Horizon Training to improve long-horizon trajectory coverage for adaptive execution. Experiments on the RoboTwin benchmark and in the real world demonstrate that our method achieves a strong robustness-efficiency trade-off: on RoboTwin, it reduces WAM forward passes by 69.10% and execution time by 34.02%, while improving success rate by 2.54% over the short-chunk baseline; in real-world experiments, it improves success rate by 35%.
Yilun Zhao, Jinbiao Wei, Tingyu Song +3 authors
Researchers introduce BRIGHT-Pro, an expanded expert-annotated benchmark for reasoning-intensive retrieval, and RTriever-Synth, an aspect-decomposed synthetic corpus, to improve retriever performance through agentic search evaluation and LoRA fine-tuning.
Yuxuan Huang, Yihang Chen, Zhiyuan He +6 authors
Web2BigTable is a multi-agent framework that addresses both broad and deep web search challenges through a bi-level architecture with coordinated agents and iterative improvement mechanisms.
Canyu Zhao, Hao Chen, Yunze Tong +3 authors
A novel gradient-space optimization framework called MARBLE addresses limitations in multi-reward reinforcement learning fine-tuning of diffusion models by maintaining independent advantage estimators and harmonizing policy gradients through quadratic programming without manual reward weighting.
Langlin Huang, Chengsong Huang, Jinyuan Li +3 authors
LoPE addresses the zero-advantage problem in reinforcement learning with verifiable rewards by usingLorem Ipsum perturbations to enhance exploration in large language model training.
Yunhan Yang, Chunshi Wang, Junliang Ye +7 authors
PhysForge generates interactive 3D assets by combining visual-language modeling for physical planning with a physics-grounded diffusion model that synthesizes detailed geometry and kinematic parameters through a novel injection mechanism.
You Qin, Kai Liu, Shengqiong Wu +12 authors
Audio-Visual Intelligence represents a multidisciplinary field integrating auditory and visual modalities through large foundation models, encompassing tasks from understanding and generation to interaction, with unified taxonomies and methodological foundations.
Yan Fang, Mengcheng Lan, Zilong Huang +7 authors
GenLIP is a minimalist generative pretraining framework for Vision Transformers that directly predicts language tokens from visual tokens using language modeling, offering simplicity, scalability, and competitive performance in multimodal tasks.
Xiangyuan Xue, Yifan Zhou, Zidong Wang +5 authors
Strategic Trajectory Abstraction framework enhances long-horizon decision making in large language models by introducing trajectory-level strategies that improve sample efficiency and performance across interactive environments.
Dengyang Jiang, Xin Jin, Dongyang Liu +9 authors
A new training approach called D-OPSD enables efficient supervised fine-tuning for diffusion models by leveraging on-policy self-distillation with text and multimodal features while preserving few-step inference capabilities.
Tao Liu, Hao Yan, Mengting Chen +8 authors
Continuous-Time Distribution Matching migrates diffusion model distillation from discrete to continuous optimization, enabling arbitrary points along sampling trajectories and preserving fine visual details through dynamic scheduling and velocity field extrapolation.
Jaeyoung Chung, Suyoung Lee, Jianfeng Xiang +2 authors
Map2World enables 3D world generation from user-defined segment maps with improved scale consistency and detail enhancement through a pipeline leveraging asset generator priors.
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