Alert button
Picture for Tian Lan

Tian Lan

Alert button

Every Parameter Matters: Ensuring the Convergence of Federated Learning with Dynamic Heterogeneous Models Reduction

Oct 26, 2023
Hanhan Zhou, Tian Lan, Guru Venkataramani, Wenbo Ding

Viaarxiv icon

Repetition In Repetition Out: Towards Understanding Neural Text Degeneration from the Data Perspective

Oct 16, 2023
Huayang Li, Tian Lan, Zihao Fu, Deng Cai, Lemao Liu, Nigel Collier, Taro Watanabe, Yixuan Su

Viaarxiv icon

Advantage Actor-Critic with Reasoner: Explaining the Agent's Behavior from an Exploratory Perspective

Sep 09, 2023
Muzhe Guo, Feixu Yu, Tian Lan, Fang Jin

Figure 1 for Advantage Actor-Critic with Reasoner: Explaining the Agent's Behavior from an Exploratory Perspective
Figure 2 for Advantage Actor-Critic with Reasoner: Explaining the Agent's Behavior from an Exploratory Perspective
Figure 3 for Advantage Actor-Critic with Reasoner: Explaining the Agent's Behavior from an Exploratory Perspective
Figure 4 for Advantage Actor-Critic with Reasoner: Explaining the Agent's Behavior from an Exploratory Perspective
Viaarxiv icon

Statistically Efficient Variance Reduction with Double Policy Estimation for Off-Policy Evaluation in Sequence-Modeled Reinforcement Learning

Aug 28, 2023
Hanhan Zhou, Tian Lan, Vaneet Aggarwal

Figure 1 for Statistically Efficient Variance Reduction with Double Policy Estimation for Off-Policy Evaluation in Sequence-Modeled Reinforcement Learning
Figure 2 for Statistically Efficient Variance Reduction with Double Policy Estimation for Off-Policy Evaluation in Sequence-Modeled Reinforcement Learning
Figure 3 for Statistically Efficient Variance Reduction with Double Policy Estimation for Off-Policy Evaluation in Sequence-Modeled Reinforcement Learning
Figure 4 for Statistically Efficient Variance Reduction with Double Policy Estimation for Off-Policy Evaluation in Sequence-Modeled Reinforcement Learning
Viaarxiv icon

Discrete Message via Online Clustering Labels in Decentralized POMDP

Aug 14, 2023
Jingdi Chen, Tian Lan

Figure 1 for Discrete Message via Online Clustering Labels in Decentralized POMDP
Figure 2 for Discrete Message via Online Clustering Labels in Decentralized POMDP
Figure 3 for Discrete Message via Online Clustering Labels in Decentralized POMDP
Figure 4 for Discrete Message via Online Clustering Labels in Decentralized POMDP
Viaarxiv icon

Minimizing Return Gaps with Discrete Communications in Decentralized POMDP

Aug 07, 2023
Jingdi Chen, Tian Lan

Figure 1 for Minimizing Return Gaps with Discrete Communications in Decentralized POMDP
Figure 2 for Minimizing Return Gaps with Discrete Communications in Decentralized POMDP
Figure 3 for Minimizing Return Gaps with Discrete Communications in Decentralized POMDP
Figure 4 for Minimizing Return Gaps with Discrete Communications in Decentralized POMDP
Viaarxiv icon

AQUILA: Communication Efficient Federated Learning with Adaptive Quantization of Lazily-Aggregated Gradients

Aug 01, 2023
Zihao Zhao, Yuzhu Mao, Zhenpeng Shi, Yang Liu, Tian Lan, Wenbo Ding, Xiao-Ping Zhang

Figure 1 for AQUILA: Communication Efficient Federated Learning with Adaptive Quantization of Lazily-Aggregated Gradients
Figure 2 for AQUILA: Communication Efficient Federated Learning with Adaptive Quantization of Lazily-Aggregated Gradients
Figure 3 for AQUILA: Communication Efficient Federated Learning with Adaptive Quantization of Lazily-Aggregated Gradients
Figure 4 for AQUILA: Communication Efficient Federated Learning with Adaptive Quantization of Lazily-Aggregated Gradients
Viaarxiv icon

Scalable Multi-agent Skill Discovery based on Kronecker Graphs

Jul 21, 2023
Jiayu Chen, Jingdi Chen, Tian Lan, Vaneet Aggarwal

Figure 1 for Scalable Multi-agent Skill Discovery based on Kronecker Graphs
Figure 2 for Scalable Multi-agent Skill Discovery based on Kronecker Graphs
Figure 3 for Scalable Multi-agent Skill Discovery based on Kronecker Graphs
Figure 4 for Scalable Multi-agent Skill Discovery based on Kronecker Graphs
Viaarxiv icon

Copy Is All You Need

Jul 13, 2023
Tian Lan, Deng Cai, Yan Wang, Heyan Huang, Xian-Ling Mao

Figure 1 for Copy Is All You Need
Figure 2 for Copy Is All You Need
Figure 3 for Copy Is All You Need
Figure 4 for Copy Is All You Need
Viaarxiv icon