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。
873 篇论文 · 按点赞排序
Julianna Piskorz, Antonin Berthon, Mihaela van der Schaar
In controlled distillation experiments, rollout policy has limited impact compared to token-level KL direction and learning rate, with on-policy benefits depending on objective and hyperparameters.
DataFlow Team, Bohan Zeng, Daili Hua +38 authors
OpenWorldLib presents a standardized framework for advanced world models that integrate perception, interaction, and long-term memory capabilities for comprehensive world understanding and prediction.
Jingxuan Wei, Siyuan Li, Yuhang Xu +21 authors
World Models require three consistency principles—modal, spatial, and temporal—for general artificial intelligence, with a proposed benchmark evaluating multimodal learning systems.
Yuhao Zheng, Li'an Zhong, Yi Wang +6 authors
Code2World enables autonomous GUI agents to predict next visual states through renderable code generation, achieving high visual fidelity and structural controllability while improving navigation performance.
Liang Chen, Weichu Xie, Yiyan Liang +26 authors
Current multimodal large language models exhibit significant gaps in fundamental visual understanding compared to human children, as demonstrated by the BabyVision benchmark.
Ailin Huang, Ang Li, Aobo Kong +212 authors
Step 3.5 Flash is a sparse Mixture-of-Experts model that achieves frontier-level agentic intelligence through efficient parameter utilization and optimized attention mechanisms, demonstrating strong performance across multiple benchmarks.
Zhixia Zhang, Zixuan Huang, Xin Xia +7 authors
HACRL enables collaborative reinforcement learning where heterogeneous agents share verified rollouts during training to improve collectively while maintaining independent operation at inference time, with HACPO achieving superior performance through efficient sample utilization and cross-agent knowledge transfer.
Qingchuan Ma, Yuexiao Ma, Xiang Meng +7 authors
Vision-language models struggle to convert 2D visual observations into precise 3D rotational control, showing large gaps versus humans in projection alignment tasks.
Kehan Li, Bohan Hou, Minghao Zhu +27 authors
RynnBrain 1.1 is a family of embodied foundation models that integrates spatio-temporal and physical grounding with contact-point prediction, 3D grounding, and cross-embodiment action spaces to improve robot perception, reasoning, and manipulation.
Guanting Dong, Xinyu Yang, Jiajie Jin +9 authors
Agent-Reflex trains interactive agents to reflect and recover from noisy tool feedback and user corrections through skill discovery, supervised fine-tuning, and on-policy optimization.
Chenxu Wang, Chaozhuo Li, Songyang Liu +10 authors
Multi-agent LLM systems face fundamental limitations in achieving continuous self-improvement while maintaining safety alignment due to inherent statistical blind spots in isolated evolution.
Guobin Shen, Xiang Cheng, Chenxiao Zhao +4 authors
Anti-Self-Distillation reverses the direction of knowledge transfer in self-distillation to improve math reasoning efficiency and accuracy.
GLM-5 Team, Aohan Zeng, Xin Lv +183 authors
GLM-5 advances foundation models with DSA for cost reduction, asynchronous reinforcement learning for improved alignment, and enhanced coding capabilities for real-world software engineering.
Ailin Huang, Chengyuan Yao, Chunrui Han +90 authors
STEP3-VL-10B achieves superior multimodal performance through unified pre-training with a language-aligned Perception Encoder and Qwen3-8B decoder, combined with scaled post-training and Parallel Coordinated Reasoning for efficient large-scale visual reasoning.
Pengfei Zhou, Hexin Wang, Zhengfeiyang Zhang +5 authors
Game engines provide executable verification and long-horizon trajectories for reinforcement learning post-training of spatial world models, motivating a human-engine verification paradigm.
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.
Haiwen Diao, Penghao Wu, Hanming Deng +55 authors
Unified vision-language models treat understanding and generation as integrated processes rather than separate tasks, demonstrating strong performance across multiple multimodal capabilities including image synthesis and action reasoning.
Kwai Keye Team, Bin Wen, Changyi Liu +50 authors
Kwai Keye-VL-2.0-30B-A3B is an open-source Mixture-of-Experts multimodal foundation model that enables long-video understanding and agentic intelligence through DeepSeek Sparse Attention and specialized training infrastructure.
Deyuan Liu, Yihao Hu, Jingxuan Zhang +13 authors
UltraText Bench evaluates dense visual text generation across bilingual scenes using structured references and a vision-language judge, revealing trade-offs between clarity, fidelity, and workload difficulty.
Jianshu Zhang, Keliang Wu, Haoran Lu +8 authors
A survey unifies progress reward modeling for robotic learning by analyzing model interfaces, internal construction methods, and evaluation benchmarks.
Ruixiang Zhao, Hualei Wang, Renhe Sun +24 authors
A proactive full-duplex system with separate audio-visual and audio models enables continuous perception, native speech generation, and asynchronous tool execution for real-time dialogue.
Jun Han, Shuo Zhang, Wei Li +21 authors
Financial markets are noisy and non-stationary, making alpha mining highly sensitive to noise in backtesting results and sudden market regime shifts. While recent agentic frameworks improve alpha mining automation, they often lack controllable multi-round search and reliable reuse of validated experience. To address these challenges, we propose QuantaAlpha, an evolutionary alpha mining framework that treats each end-to-end mining run as a trajectory and improves factors through trajectory-level mutation and crossover operations. QuantaAlpha localizes suboptimal steps in each trajectory for targeted revision and recombines complementary high-reward segments to reuse effective patterns, enabling structured exploration and refinement across mining iterations. During factor generation, QuantaAlpha enforces semantic consistency across the hypothesis, factor expression, and executable code, while constraining the complexity and redundancy of the generated factor to mitigate crowding. Extensive experiments on the China Securities Index 300 (CSI 300) demonstrate consistent gains over strong baseline models and prior agentic systems. When utilizing GPT-5.2, QuantaAlpha achieves an Information Coefficient (IC) of 0.1501, with an Annualized Rate of Return (ARR) of 27.75% and a Maximum Drawdown (MDD) of 7.98%. Moreover, factors mined on CSI 300 transfer effectively to the China Securities Index 500 (CSI 500) and the Standard & Poor's 500 Index (S&P 500), delivering 160% and 137% cumulative excess return over four years, respectively, which indicates strong robustness of QuantaAlpha under market distribution shifts.
Jiaqi Liu, Shi Qiu, Mairui Li +32 authors
AutoResearchClaw is a multi-agent autonomous research system that improves scientific discovery through structured debate, self-healing execution, verifiable reporting, human collaboration, and evolutionary learning, outperforming previous systems on a benchmark while maintaining human oversight.
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.
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.
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.
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.
Chuyang Wei, Haoxiang Guan, Minghang Zhu +5 authors
A framework for self-evolving agent harnesses uses feedback-driven revisions and evolutionary optimization with theoretical generalization guarantees while keeping model weights fixed.
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.
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