Bayesian Multi-Task Reinforcement Learning

By Alessandro Lazaric, Mohammad Ghavamzadeh. 2010

120 citations recorded by the source corpus.

Subfields: Reinforcement Learning, Probabilistic & Bayesian Methods

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Reinforcement Learning

In Probabilistic & Bayesian Methods

Citations in the corpus

An Overview of Multi-task Learning (2018, methodology): “In [71], the reinforcement learning model for each task is a Gaussian process temporal-difference value function model and a hierarchical Bayesian model relates value functions of different tasks.”
A Survey on Multi-Task Learning (2017, neutral): “In [135], the Gaussian process temporal-”
Stochastic Neural Networks for Hierarchical Reinforcement Learning (2016, background): “…studied in the past in sequential settings, where the agent is exposed to a sequence of tasks, and should learn to make use of experience gathered from solving earlier tasks to help solve later tasks (Taylor & Stone, 2009; Wilson et al., 2007; Lazaric & Ghavamzadeh, 2010; Devin et al., 2017).”
Reward Shaping via Meta-Learning (2019, background): “…induce a distribution over tasks , which in this case is a distribution over maze conﬁgurations (Wilson et al., 2007) and could elsewhere be a distribution over system parameters (Lazaric & Ghavamzadeh, 2010) for different robot-hand sizes or over game maps for RTS games (Jaderberg et al., 2018).”
Transfer Learning in Deep Reinforcement Learning: A Survey (2020, background): “Multi-Task RL learns an agent with generalized skills across various tasks, hence it can solve MDPs randomly sampled from a ﬁxed yet unknown distribution [48].”
Bayesian Reinforcement Learning (2010, methodology): “While some methods explicitly assume that the shared components are drawn from a common generative model (Wilson et al, 2007; Mehta et al, 2008; Lazaric and Ghavamzadeh, 2010), this assumption is more implicit in others (Taylor et al, 2007; Lazaric et al, 2008).”
Learning Transferable Knowledge Through Embedding Spaces (2019, methodology): “Bayesian approaches take a variety of forms, making use of common priors [173; 174], using regularization terms that couple task parameters [175; 176], and finding mixtures of experts that can be shared across tasks [177].”
Using Task Descriptions in Lifelong Machine Learning for Improved Performance and Zero-Shot Transfer (2017, methodology): “Bayesian approaches take a variety of forms, making use of common priors (Wilson, Fern, Ray, & Tadepalli, 2007; Lazaric & Ghavamzadeh, 2010), using regularization terms that couple task parameters (Evgeniou & Pontil, 2004; Zhong & Kwok, 2012), and ﬁnding mixtures of experts that can be shared…”
RT-1: Robotics Transformer for Real-World Control at Scale (2022, background): “[7, 70, 33]) that consider different tasks to be MDPs with different dynamics or reward functions.”
Online Multi-Task Learning for Policy Gradient Methods (2014, methodology): “Unlike our problem setting, their method focuses on off-policy batch MTL. Finally, Lazaric & Ghavamzadeh (2010) exploit shared structure in the value functions between related MDPs.”
Sharing Knowledge in Multi-Task Deep Reinforcement Learning (2020, background): “…this, interesting analyses consider Bayesian approaches; for instance Wilson et al. (2007) assumes that the tasks are generated from a hierarchical Bayesian model, and likewise Lazaric & Ghavamzadeh (2010) considers the case when the value functions are generated from a common prior distribution.”
Probabilistic Movement Primitives (2013, methodology): “The parameters θ = {μw,Σw} can be learned from multiple demonstrations by maximum likelihood estimation, for example, by using the expectation maximization algorithm for HBMs with Gaussian distributions [18].”
A Model-based Approach for Sample-efficient Multi-task Reinforcement Learning (2019, unknown): “A number of previous works have considered multi-task reinforcement learning [Oh et al., 2017, Ammar et al., 2014, Wilson et al., 2007, Lazaric and Ghavamzadeh, 2010], and Taylor and Stone [2009] survey the area.”
Autonomous Agents and Multi-agent Systems Manuscript No. Learning Potential Functions and Their Representations for Multi-task Reinforcement Learning (unknown): “Other forms of initialization are to use a population of evolved source task policies as initial population for an evolutionary algorithm in the target task [78], or to use source task information to set the prior for Bayesian RL [41,89].”