Graph kernels and Gaussian processes for relational reinforcement learning

By Thomas Gärtner, Kurt Driessens, J. Ramon. 2006

109 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computational Theory and Mathematics

How researchers see this paper

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

In Artificial Intelligence

1. Sriraam Natarajan
2. Kristian Kersting
3. Martijn van Otterlo
4. Ronald E. Parr
5. Kaushik Roy
6. Srijita Das
7. Robert Givan
8. Prasad Tadepalli

In Computational Theory and Mathematics

1. Martijn van Otterlo
2. Kristian Kersting
3. Sriraam Natarajan
4. Robert Givan
5. Craig Boutilier
6. Scott Sanner
7. Prasad Tadepalli
8. Jia-Hong Wu

Citations in the corpus

• Fitted Q-Learning for Relational Domains (2020, methodology): “This work was later extended by others (Driessens and Ramon 2003; Driessens and Dˇzeroski 2004; Driessens, Ramon, and Blockeel 2001; G¨artner, Driessens, and Ramon 2003) w.r.t building incremental RRTs, providing guidance to the learner and regression algorithms using kernels and gaus-sian…”
• Cyclic pattern kernels for predictive graph mining (2004, unknown): “We note, however, that successful applications using such graph kernels have been reported in [19] and [ 14 ] on smaller datasets.”
• Relational reinforcement learning (2001, unknown): “The third regression algorithm KBR is based on Gaussian processes for regression and uses graph kernels as a covariance function between state-action pairs [8].”
• Solving Relational and First-Order Logical Markov Decision Processes: A Survey (2012, methodology): “Compared to RIB, Gärtner et al (2003) take a more principled approach in the KBR algorithm to distances between relational states and use graph kernels and Gaussian processes for value function approximation in relational RL.”
• Relational Reinforcement Learning An Overview (2004, methodology): “Because Gaussian processes are a Bayesian technique, the KBR algorithm ooers more than just a basic prediction of the Q-value of a new unseen example.”
• Practical solution techniques for first-order MDPs (2009, methodology): “This can be done with pure reinforcement learning using relational decision or regression trees to learn a value or Q-function (Dzeroski et al., 2001), combining this with supervised guidance (Driessens and Dzeroski, 2002), or using Gaussian processes and graph kernels over relational structures to ”
• Non-parametric policy gradients: a unified treatment of propositional and relational domains (2008, methodology): “Basically, a num­ber of relational regression algorithms have been de­veloped for use in this RL system that employ rela­tional regression trees (Driessens et al., 2001), rela­tional instance based regression (Driessens & Ramon, 2003), graph kernels and Gaussian processes (G¨artner et al., 20”
• Discovering Relational Domain Features for Probabilistic Planning (2007, methodology): “We set the discount factor γ to 0.9.”