Multiple Didi papers accepted by ICML 2026, worth reading!
DiDi has had five academic papers accepted into the prestigious International Conference on Machine Learning (ICML) 2026. These papers, developed in collaboration with universities like Sun Yat-sen University and The Hong Kong University of Science and Technology (Guangzhou), span areas such as large language model agents, reinforcement learning, and causal inference. The research aims to push the boundaries of AI capabilities in complex, real-world scenarios, with a focus on improving efficiency and performance in areas like long-horizon tasks and GUI automation. AI
IMPACT Highlights advancements in LLM agents and reinforcement learning, potentially influencing future AI development in complex task execution.
- Peking University
- DiDi
- Sun Yat-sen University
- International Conference on Machine Learning
- Shanghai University of Finance and Economics
- The Hong Kong University of Science and Technology (Guangzhou)
- Agent-Omit: Adaptive Context Omission for Efficient LLM Agents
- UltraHorizon: Benchmarking LLM-Agent Capabilities in Ultra Long-Horizon Scenarios
- Darwinian Memory: A Training-Free Self-Regulating Memory System for GUI Agent Evolution
- HTAC: Hierarchical Task-Aware Composition for Continual Offline Reinforcement Learning
- Feasible Fusion: Constrained Joint Estimation under Structural Non-Overlap