Task Specific Policy

Task-specific policy learning aims to create robust and efficient robotic agents capable of mastering diverse tasks, addressing limitations of generalist policies that struggle with modularity and interpretability. Current research focuses on developing efficient multi-task architectures, such as transformers and GFlowNets, and algorithms like pessimistic value iteration and behavior distillation, to improve data efficiency and generalization across tasks. These advancements are significant for robotics, enabling faster training, better performance on complex tasks, and improved interpretability of learned behaviors, ultimately leading to more adaptable and reliable autonomous systems.

Papers

October 24, 2024

Data Scaling Laws in Imitation Learning for Robotic Manipulation
Fanqi Lin, Yingdong Hu, Pingyue Sheng, Chuan Wen, Jiacheng You, Yang Gao
Imitation Learning Generalization Performance Robotic Manipulation Scaling Law Real Robot Task Specific Policy Data Scaling

July 10, 2024

Towards Interpretable Foundation Models of Robot Behavior: A Task Specific Policy Generation Approach
Isaac Sheidlower, Reuben Aronson, Elaine Schaertl Short
Foundation Model Robot Behavior Reinforcement Learning Policy Task Specific Policy Robot Experience

June 11, 2024

BAKU: An Efficient Transformer for Multi-Task Policy Learning
Siddhant Haldar, Zhuoran Peng, Lerrel Pinto
Efficient Transformer Offline Imitation Learning Real World Manipulation Task Task Specific Policy Multi Task Policy Simple Transformer

April 30, 2024

Pessimistic Value Iteration for Multi-Task Data Sharing in Offline Reinforcement Learning
Chenjia Bai, Lingxiao Wang, Jianye Hao, Zhuoran Yang, Bin Zhao, Zhen Wang, Xuelong Li
Offline Reinforcement Learning Value Iteration Task Specific Policy Offline Dataset

March 7, 2024

Symmetry Considerations for Learning Task Symmetric Robot Policies
Mayank Mittal, Nikita Rudin, Victor Klemm, Arthur Allshire, Marco Hutter
Cyclic Symmetry Symmetry Detection Task Specific Policy Transformation Invariance

January 22, 2024

P2DT: Mitigating Forgetting in task-incremental Learning with progressive prompt Decision Transformer
Zhiyuan Wang, Xiaoyang Qu, Jing Xiao, Bokui Chen, Jianzong Wang
Catastrophic Forgetting Faulty Negative Mitigation Decision Transformer Traditional Reinforcement Learning Task Specific Policy Task Incremental Learning

October 12, 2023

PolyTask: Learning Unified Policies through Behavior Distillation
Siddhant Haldar, Lerrel Pinto
Unified Model Diverse Policy Task Specific Policy Real Robotic System

October 11, 2023

Contextualized Policy Recovery: Modeling and Interpreting Medical Decisions with Adaptive Imitation Learning
Jannik Deuschel, Caleb N. Ellington, Yingtao Luo, Benjamin J. Lengerich, Pascal Friederich, Eric P. Xing
Decision Making Clinical Decision Task Specific Policy

September 18, 2023

Q-Transformer: Scalable Offline Reinforcement Learning via Autoregressive Q-Functions
Yevgen Chebotar, Quan Vuong, Alex Irpan, Karol Hausman, Fei Xia, Yao Lu, Aviral Kumar, Tianhe Yu, Alexander Herzog, Karl Pertsch, Keerthana Gopalakrishnan, Julian Ibarz, Ofir Nachum, Sumedh Sontakke, Grecia Salazar, Huong T Tran, Jodilyn Peralta, Clayton Tan, Deeksha Manjunath, Jaspiar Singht, Brianna Zitkovich, Tomas Jackson, Kanishka Rao, Chelsea Finn, Sergey Levine
Reinforcement Learning Offline Reinforcement Learning Q Learning Q Function Task Specific Policy Q Transformer

August 29, 2023

Policy composition in reinforcement learning via multi-objective policy optimization
Shruti Mishra, Ankit Anand, Jordan Hoffmann, Nicolas Heess, Martin Riedmiller, Abbas Abdolmaleki, Doina Precup
Reinforcement Learning Action Space Policy Optimization Policy Design Task Specific Policy Behavior Policy Multi Objective Policy

June 16, 2023

Meta Generative Flow Networks with Personalization for Task-Specific Adaptation
Xinyuan Ji, Xu Zhang, Wei Xi, Haozhi Wang, Olga Gadyatskaya, Yinchuan Li
Meta Reinforcement Learning Generative Flow Network Task Adaptation Multi Task Reinforcement Learning Task Specific Policy

May 26, 2023

DynaShare: Task and Instance Conditioned Parameter Sharing for Multi-Task Learning
Elahe Rahimian, Golara Javadi, Frederick Tung, Gabriel Oliveira
Multi Task Learning Related Task Parameter Sharing Task Specific Policy Multi Task Network

May 22, 2023

Multi-task Hierarchical Adversarial Inverse Reinforcement Learning
Jiayu Chen, Dipesh Tamboli, Tian Lan, Vaneet Aggarwal
Task Specific Policy Adversarial Inverse Reinforcement Learning Hierarchical Policy Learning

March 14, 2023

Merging Decision Transformers: Weight Averaging for Forming Multi-Task Policies
Daniel Lawson, Ahmed H. Qureshi
Decision Transformer Sequential Decision Making Problem Multi Task Model WEight AVERaging Task Specific Policy

February 6, 2023

Robust Subtask Learning for Compositional Generalization
Kishor Jothimurugan, Steve Hsu, Osbert Bastani, Rajeev Alur
Compositional Generalization Zero Sum Game Task Specific Policy Compositional Reinforcement Learning

October 10, 2022

Using Both Demonstrations and Language Instructions to Efficiently Learn Robotic Tasks
Albert Yu, Raymond J. Mooney
Noisy Demonstration Robotic Task Language Conditioned Novel Task Task Specific Policy Task Embeddings Language Instruction

March 29, 2022

Demonstration-Bootstrapped Autonomous Practicing via Multi-Task Reinforcement Learning
Abhishek Gupta, Corey Lynch, Brandon Kinman, Garrett Peake, Sergey Levine, Karol Hausman
Reinforcement Learning Multi Task Reinforcement Learning Reinforcement Learning System Task Specific Policy

Task Specific Policy

Papers

Data Scaling Laws in Imitation Learning for Robotic Manipulation

Towards Interpretable Foundation Models of Robot Behavior: A Task Specific Policy Generation Approach

BAKU: An Efficient Transformer for Multi-Task Policy Learning

Pessimistic Value Iteration for Multi-Task Data Sharing in Offline Reinforcement Learning

Symmetry Considerations for Learning Task Symmetric Robot Policies

P2DT: Mitigating Forgetting in task-incremental Learning with progressive prompt Decision Transformer

PolyTask: Learning Unified Policies through Behavior Distillation

Contextualized Policy Recovery: Modeling and Interpreting Medical Decisions with Adaptive Imitation Learning

Q-Transformer: Scalable Offline Reinforcement Learning via Autoregressive Q-Functions

Policy composition in reinforcement learning via multi-objective policy optimization

Meta Generative Flow Networks with Personalization for Task-Specific Adaptation

DynaShare: Task and Instance Conditioned Parameter Sharing for Multi-Task Learning

Multi-task Hierarchical Adversarial Inverse Reinforcement Learning

Merging Decision Transformers: Weight Averaging for Forming Multi-Task Policies

Robust Subtask Learning for Compositional Generalization

Using Both Demonstrations and Language Instructions to Efficiently Learn Robotic Tasks

Demonstration-Bootstrapped Autonomous Practicing via Multi-Task Reinforcement Learning