Exploiting Open-Endedness to Solve Problems Through the Search for Novelty

By Joel Lehman, Kenneth O. Stanley. 2008

590 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.

Citations in the corpus

Weight Agnostic Neural Networks (2019, background): “…single WANN capable of encoding many different useful tasks in its environment, one might consider developing a WANN with a strong intrinsic bias for intrinsic motivation [82, 84, 94], and continuously optimize its architecture to perform well at pursuing novelty in an open-ended environment [60].”
Born to Learn: the Inspiration, Progress, and Future of Evolved Plastic Artificial Neural Networks (2017, background): “…(Wagner and Altenberg, 1996; Pigliucci, 2008; Lehman and Stanley, 2013), the encoding of genetic information through the genotype-phenotype mapping processes (Hornby et al., 2002), and novelty search and the deception of ﬁtness objectives (Lehman and Stanley, 2008; Stanley and Lehman, 2015).”
Large-Scale Study of Curiosity-Driven Learning (2018, background): “Learning without extrinsic rewards or ﬁtness functions has also been studied extensively in the evolutionary computing where it is referred to as ‘novelty search’ [17, 18, 43].”
Paired Open-Ended Trailblazer (POET): Endlessly Generating Increasingly Complex and Diverse Learning Environments and Their Solutions (2019, background): “While a small community within the field of artificial life [18, 19, 21–28] has studied the prospects of open-ended computation for many years, in machine learning the prevailing assumption that algorithms should be directed (e.”
A survey and critique of multiagent deep reinforcement learning (2018, methodology): “, evolutionary methods [20,85,185,271] or sampling-based methods, have shown good results [24,156,187].”
Open-ended Learning in Symmetric Zero-sum Games (2019, background): “There is a large literature on novelty search, open-ended evolution, and curiosity, which aim to continually expand the frontiers of game knowledge within an agent (Lehman & Stanley, 2008; Taylor et al., 2016; Banzhaf et al., 2016; Brant & Stanley, 2017; Pathak et al., 2017; Wang et al., 2019).”
Formal Theory of Creativity, Fun, and Intrinsic Motivation (1990–2010) (2010, background): “Recent related work in the field of evolutionary computation aims at increasing diversity within populations of individ uals [40], [21], [23].”
Novelty-Based Multiobjectivization (2011, unknown): “Tothat aim, we prolong the maze experiments described in the original novelty search paper [14] by comparing them to our multi-objective method.”
Open-Ended Evolution: Perspectives from the OEE Workshop in York (2016, unknown): “Mentioned in Dolsonʼs talk, Lehman and Stanleyʼs novelty search technique has attracted considerable interest in recent years [25, 26].”
Critical factors in the performance of hyperNEAT (2013, background): “There is some evidence that novelty search can improve the performance of NEAT [18] and HyperNEAT [20].”
How novelty search escapes the deceptive trap of learning to learn (2009, methodology): “Although it means that the search is wider, as Lehman and Stanley [12] write, “It is better to search far and wide and eventually reach a summit than to search narrowly and single-mindedly yet never come close.””
No One Representation to Rule Them All: Overlapping Features of Training Methods (2021, background): “…for novelty, we have stumbled upon the decade-old ﬁnding that some problems are best solved by methods that ignore the objective, since “the objective function does not necessarily reward the stepping stones in the search space that ultimately lead to the objective” (Lehman and Stanley, 2008).”
Effective Diversity in Population-Based Reinforcement Learning (2020, methodology): “In the context of population-based Reinforcement Learning, one prominent approach is the class of novelty search methods for RL (Lehman and Stanley, 2008; 2011; Conti et al., 2018).”
Evolutionary and Computational Advantages of Neuromodulated Plasticity (2008, background): “In landscapes with many local optima, a search based on novelty of behaviour for a navigation problem showed better results than a fitnessbased evolution (Lehman and Stanley, 2008).”
Indirectly Encoding Neural Plasticity as a Pattern of Local Rules (2010, unknown): “[19] showed that novelty search [20], which abandons objective-based fitness and instead simply searches only for novel behavior, significantly outperforms fitness-based search in the traditional T-Maze domain, a standard fitness function is employed in this paper to keep the experiment focused on t”
Quality Diversity: A New Frontier for Evolutionary Computation (2016, methodology): “9 Pugh et al. Quality Diversity: A New Frontier Frontiers in Robotics and AI | www.frontiersin.org July 2016 | Volume 3 | Article 40 solutions to deceptive mazes (Lehman and Stanley, 2008).”
Data-efficient neuroevolution with kernel-based surrogate models (2018, methodology): “networks with the HyperNEAT algorithm [29]. The ability of NEAT to produce networks of increasing complexity has also made it an ideal tool for exploring open-ended evolution and novelty-based search [13, 14]. 2.2 Gaussian Process Models Surrogate models can be constructed with a variety of machine ”