activity
20242026
collaborators

5 papers

cs.LG2026

One Model for All Tasks: Leveraging Efficient World Models in Multi-Task Planning

Yuan Pu, Yazhe Niu, Jia Tang +3

In heterogeneous multi-task decision-making, tasks not only exhibit diverse observation and action spaces but also vary substantially in their underlying complexities. While conven…

cs.IR2026

FlashEvaluator: Expanding Search Space with Parallel Sequence-Level Evaluation

Chao Feng, Yuanhao Pu, Chenghao Zhang +9

The Generator-Evaluator (G-E) framework generates K candidate sequences and uses an evaluator to select the highest-scoring one, which is widely used in recommender systems (RecSys…

cs.IR2026

SOLAR: SVD-Optimized Lifelong Attention for Recommendation

Chenghao Zhang, Chao Feng, Yuanhao Pu +8

Attention mechanism remains the defining operator in Transformers since it provides expressive global credit assignment, yet its time and memory cost in sequence length…

cs.LG2025

UniZero: Generalized and Efficient Planning with Scalable Latent World Models

Yuan Pu, Yazhe Niu, Zhenjie Yang +3

Learning predictive world models is crucial for enhancing the planning capabilities of reinforcement learning (RL) agents. Recently, MuZero-style algorithms, leveraging the value e…

cs.AI2024

ReZero: Boosting MCTS-based Algorithms by Backward-view and Entire-buffer Reanalyze

Chunyu Xuan, Yazhe Niu, Yuan Pu +3

Monte Carlo Tree Search (MCTS)-based algorithms, such as MuZero and its derivatives, have achieved widespread success in various decision-making domains. These algorithms employ th…