Relational reinforcement learning

By Kurt Driessens. 2001

360 citations recorded by the source corpus.

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

Citations in the corpus

Interpretable Machine Learning: Fundamental Principles and 10 Grand Challenges (2021, background): “Typically, in relational reinforcement learning, policies are represented as relational regression trees [90, 91, 143, 96, 76] that are more interpretable than neural networks.”
Towards Deep Symbolic Reinforcement Learning (2016, methodology): “First, the incorporation of inductive logic programming [34] would enable a more powerful form of gener-alisation to be applied to the Q function [35].”
Relational inductive biases, deep learning, and graph networks (2018, background): “Entire sub-ﬁelds have focused on explicit entity-and relation-centric learning, such as relational reinforcement learning (Dˇzeroski et al., 2001) and statistical relational learning (Getoor and Taskar, 2007).”
Fitted Q-Learning for Relational Domains (2020, methodology): “Tree-based RRL: Our work is inspired by Dzeroski et al. (2001) on learning single regression tree for RRL.”
Relational Deep Reinforcement Learning (2018, methodology): “Here we improve on deep RL architectures by leveraging insights introduced in the RL literature over 20 years ago under the Relational RL umbrella (RRL, [8, 9]).”
Emergent Tool Use From Multi-Agent Autocurricula (2019, methodology): “[70] Sašo Džeroski, Luc De Raedt, and Kurt Driessens.”
Automatically Composing Representation Transformations as a Means for Generalization (2018, background): “sumes pre-specified transformations and learns the structure (from dense supervision on execution traces to sparserewards) [13, 14, 18, 22, 25, 28, 65, 87, 90].”
Learning Multi-Agent Communication through Structured Attentive Reasoning (2020, background): “The core idea of relational reinforcement learning (RRL) combines inductive logic programming [21, 22] and reinforcement learning to perform reasoning steps iterated over entities in the environment.”
Autonomous transfer for reinforcement learning (2008, neutral): “If the problem is formulated in the relational reinforcement learning [3] (RRL) framework, source task policies can be directly reused if the number of objects change in subsequent tasks.”
Relational inductive bias for physical construction in humans and machines (2018, background): “…& Loftus, 1975; Shepard, 1980; Grifﬁths et al., 2007; Kemp & Tenenbaum, 2008) as well as classic approaches like relational reinforcement learning (Dˇzeroski et al., 2001), and represents a step forward by showing how relational knowledge can be disentangled from physical knowledge through…”
Autonomous Learning of State Representations for Control: An Emerging Field Aims to Autonomously Learn State Representations for Reinforcement Learning Agents from Their Real-World Sensor Observations (2015, background): “relational RL [10, 24]) to select the subset Ŝ.”
Solving Relational and First-Order Logical Markov Decision Processes: A Survey (2012, neutral): “P-learning can be used to learn policies from the current Q-function and a training set (Džeroski et al, 2001).”
Learning Relational Navigation Policies (2006, positive): “Our approach alleviates this situation by adopting techniques from relational reinforcement learning [9], [10], i.”
Graph kernels and Gaussian processes for relational reinforcement learning (2006, neutral): “Relational reinforcement learning (RRL) [9] is a Q-learning technique that allows structural representations for states and actions.”
Relational Instance Based Regression for Relational Reinforcement Learning (2003, neutral): “Relational reinforcement learning or RRL (Džeroski et al., 1998) is a learning technique that combines Qlearning with relational representations for the states, actions and the resulting Q-function.”
Rule abstraction and transfer in reinforcement learning by decision tree (2012, background): “Relational reinforcement learning (RRL)[4] will be a sound solution in this problem.”
Value Functions for RL-Based Behavior Transfer: A Comparative Study (2005, neutral): “For tasks which can be framed in a relational framework ( Dzeroski, Raedt, & Driessens 2001 ), there is research (Morales 2003) which suggests ways of speeding up learning between two relational reinforcement learning tasks.”