Efficient reinforcement learning through symbiotic evolution

By David E. Moriarty, Risto Miikkulainen. 2004

333 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 Artificial Intelligence

Citations in the corpus

Deep learning in neural networks: An overview (2014, neutral): “One particularly effective family of methods coevolves neurons, combining them into networks, and selecting those neurons for reproduction that participated in the best-performing networks (Moriarty and Miikkulainen, 1996; Gomez, 2003; Gomez and Miikkulainen, 2003).”
Real-time neuroevolution in the NERO video game (2005, background): “While some evolve only the connection weight values of fixed-topology networks [37], [47]–[49], others evolve both weights and network topology simultaneously [38], [50]–[62].”
Training Recurrent Networks by Evolino (2007, positive): “In particular, recent progress has been made with cooperatively coevolving recurrent neurons, each with its own rather small, local search space of possible weight vectors [6, 23, 30].”
Modeling systems with internal state using evolino (2005, methodology): “This cooperative coevolutionary approach is an extension to Symbiotic, Adaptive Neuroevolution (SANE; [15]) which also evolves neurons, but in a single population.”
Driven by Compression Progress: A Simple Principle Explains Essential Aspects of Subjective Beauty, Novelty, Surprise, Interestingness, Attention, Curiosity, Creativity, Art, Science, Music, Jokes (2008, background): “The components are coevolvedby combining them into networks, and selecting those for reproduction that participated in the best performing networks [ 36 , 16, 17, 15, 18, 19, 14].”
Cartesian genetic programming encoded artificial neural networks: a comparison using three benchmarks (2013, methodology): “19000 - 100 SAIN [7] 12600 - 50 Q-MPL [7] 10583 - 50 ESP [7] 3800 - 50 NEAT [41] 3578 2704 120 NEvA [43] 2177 - 50 CGPANN 1111 1476 100 CoSyNE [6] 954 - 50 CMA-ES [12] 895 - 50 DirE [8] 410 - >30 Table 2: Comparison of Results for the Ball Throwing Benchmark Method Evaluations Std. Dev.”
Stochastic search using the natural gradient (2009, background): “Method SANE ESP NEAT CMA-ES CoSyNE FEM NES Evaluations 262, 700 7, 374 6, 929 3, 521 1, 249 2, 099 1, 753 9.”
Efficient Reinforcement Learning Through Evolving Neural Network Topologies (2002, methodology): “SANE and ESP results were reported by Gomez (1999).”
Evolving Soccer Keepaway Players Through Task Decomposition (2005, methodology): “Neuroevolution has repeatedly been shown to be an effective reinforcement learning method for non-linear control tasks (Gruau et al., 1996; Moriarty and Miikkulainen, 1996; Stanley and Miikkulainen, 2004; Gomez, 2003).”
Reinforcement Learning in Continuous State and Action Spaces (2012, methodology): “Both approaches have often been used in reinforcement learning with considerable success (Sutton, 1988; Werbos, 1989b,a, 1990; Whitley et al, 1993; Tesauro, 1994, 1995; Moriarty and Miikkulainen, 1996; Moriarty et al, 1999; Whiteson and Stone, 2006; Wierstra et al, 2008; Rückstieß et al, 2010).”
The road to everywhere : Evolution, complexity and progress in natural and artificial systems (2008, neutral): “[66], [131].”
Neuroevolution: from architectures to learning (2008, background): “Redrawn from Stanley and Miikkulainen [79] genotype is a collection of rules governing the process of cell divisions (a single cell is replaced by two ''daughter'' cells) and transformations (new connections can be added and the weights of the connections departing from a cell can be modified).”
Evolving Neural Networks through Augmenting Topologies (2002, methodology): “Past studies have shown NE to be faster and more efficient than reinforcement learning methods such as Adaptive Heuristic Critic and Q-Learning on single pole balancing and robot arm control (Moriarty 1997; Moriarty and Miikkulainen 1996).”
Neuroevolution for reinforcement learning using evolution strategies (2003, methodology): “method evaluations population size NE (Wieland, 1991) 307200 2048 EP (Saravanan and Fogel, 1995) 80000 100 SANE ( Moriarty and Miikulainen, 1996 ) 12600 200 ESP (Gomez and Miikulainen, 1999) 3800 200 NEAT (Stanley and Miikkulainen, 2002) 3600 150”