Evolving plastic neural networks with novelty search

By Sebastian Risi, C. Hughes, Kenneth O. Stanley. 2010

99 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

Born to Learn: the Inspiration, Progress, and Future of Evolved Plastic Artificial Neural Networks (2017, methodology): “Using a T-Maze domain as learning task, Risi and Stanley (2010) showed that HyperNEAT, which usually implements a compact encoding of weight patterns for large-scale ANNs (Fig.”
Discovering Gated Recurrent Neural Network Architectures (2018, background): “Artificial curiosity Schmidhuber (2010) is one such concept, where the agent is explicitly rewarded for exploring areas in the environment that provide more information.”
Using a map-based encoding to evolve plastic neural networks (2011, methodology): “It is therefore different from a switch between only two alternatives, as done by most previous authors [8], [27].”
Evolving Plastic Neural Networks for Online Learning: Review and Future Directions (2012, methodology): “It has been noted by several authors that most tasks requiring online learning also require some fixed behaviours, and that often the fixed behaviours can be evolved much more easily than the learning behaviours [20,21].”
Quality Diversity: A New Frontier for Evolutionary Computation (2016, methodology): “The counterintuitive result that NS can sometimes find the best solutions without explicitly searching for them has since sparked considerable research interest in applying NS and methods like it to solving problems that were previously considered to be too difficult (Lehman and Stanley, 2008, 2010,”
Developmental Plasticity in Cartesian Genetic Programming based Neural Networks (2016, background): “learning a specific task is easy but learning the strategy to learn and adapting its architecture to the changing task takes time (Risi et al., 2010).”
Overcoming deception in evolution of cognitive behaviors (2014, methodology): “Importantly, some such studies support the idea that evolving cognitive behaviors may be deceptive [13, 17] and further suggest that methods that are based on novelty rather than fitness may be beneficial [17].”
Evolving Deep LSTM-based Memory Networks using an Information Maximization Objective (2016, background): “Risi et al. [19] used one such T-maze extension (double T-maze) to test plastic neural networks.”
Automated Curriculum Learning by Rewarding Temporally Rare Events (2018, background): “with deceptive reward structures, just as novelty search outperformed traditional evolutionary algorithms in mazes with dead ends [24] or deceptive meta-learning tasks [37].”
Evolving a diversity of virtual creatures through novelty search and local competition (2011, background): “Such disconnect between morphology and functionality differs from prior experiments with novelty search in which the space it explores also distinguishes behaviors of interest [14, 17].”
Real-time Hebbian Learning from Autoencoder Features for Control Tasks (2014, background): “…artificial life and related fields (Baxter, 1992; Floreano and Urzelai, 2000; Niv et al., 2002; Soltoggio et al., 2008, 2007; Soltoggio and Jones, 2009; Soltoggio and Stanley, 2012; Risi and Stanley, 2010; Risi et al., 2011; Risi and Stanley, 2012; Stanley et al., 2003; Coleman and Blair, 2012).”
On the Relationships between Generative Encodings, Regularity, and Learning Abilities when Evolving Plastic Artificial Neural Networks (2013, methodology): “Following [21–23], we distinguish two types of neurons: “stan-dard neurons” and “modulatory neurons”.”
Evolving Neural Turing Machines for Reward-based Learning (2016, methodology): “Especially tasks involving cognitive behaviors and adaptation have shown to be very deceptive for objective-based search methods [7, 9, 12] and benefit from more explorative search methods such as novelty search [8].”
Evolution of Cartesian Genetic Programs for Development of Learning Neural Architecture (2011, background): “Risi and colleagues extended the HyperNEAT encoding to evolve both synaptic weights and parameters of learning rules (Risi et al., 2010).”