Blog
告别冗长 PDF，一文读懂最新顶会与核心期刊的创新价值。
Seedance 2.0: Advancing Video Generation for World Complexity
Agent-World: Scaling Real-World Environment Synthesis for Evolving General Agent Intelligence
Training-Free Group Relative Policy Optimization
Trusting Your Evidence: Hallucinate Less with Context-aware Decoding
HII-DPO: Eliminate Hallucination via Accurate Hallucination-Inducing Counterfactual Images
Topology optimization method to devise liquid-cooling plate for managing thermal field of a large-sized lithium-ion battery
ParamMute: Suppressing Knowledge-Critical FFNs for Faithful Retrieval-Augmented Generation
Dual-View Training for Instruction-Following Information Retrieval
Visually-Guided Policy Optimization for Multimodal Reasoning
Dynamic Attention-Guided Context Decoding for Mitigating Context Faithfulness Hallucinations in Large Language Models
DeTeCtive: Detecting AI-generated Text via Multi-Level Contrastive Learning
Stable-RAG: Mitigating Retrieval-Permutation-Induced Hallucinations in Retrieval-Augmented Generation
Stream-CQSA: Avoiding Out-of-Memory in Attention Computation via Flexible Workload Scheduling
Near-Future Policy Optimization
Semantic Recall for Vector Search
GRPO-VPS: Enhancing Group Relative Policy Optimization with Verifiable Process Supervision for Effective Reasoning
Where and What: Reasoning Dynamic and Implicit Preferences in Situated Conversational Recommendation
OMIBench: Benchmarking Olympiad-Level Multi-Image Reasoning in Large Vision-Language Model
Can "AI" Be a Doctor? A Study of Empathy, Readability, and Alignment in Clinical LLMs
Parallel-SFT: Improving Zero-Shot Cross-Programming-Language Transfer for Code RL
Generative AI and English language teaching: A global Englishes perspective
SeekerGym: A Benchmark for Reliable Information Seeking
OneVL: One-Step Latent Reasoning and Planning with Vision-Language Explanation
DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
From Research Question to Scientific Workflow: Leveraging Agentic AI for Science Automation
LLM Reasoning Is Latent, Not the Chain of Thought
SkillFlow:Benchmarking Lifelong Skill Discovery and Evolution for Autonomous Agents
LiveResearchBench: A Live Benchmark for User-Centric Deep Research in the Wild
The Impact of Off-Policy Training Data on Probe Generalisation
Parallel Test-Time Scaling for Latent Reasoning Models