Evolving Stable Strategies
By David R Ha. 2017
33 citations recorded by the source corpus.
Subfields: Genetics, Management Science and Operations Research, Sociology and Political Science, Game & Decision Theory, Evolutionary Computation
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Genetics
- Kenneth O. Stanley
- Jeff Clune
- Joel Lehman
- J. Urgen Schmidhuber
- Adam Gaier
- Sebastian Risi
- Rui Wang
- Yingtao Tian
In Management Science and Operations Research
- J. Urgen Schmidhuber
- Jeff Clune
- Kenneth O. Stanley
- Joel Lehman
- Rui Wang
- Adam Gaier
- Yingtao Tian
- Luke Metz
In Sociology and Political Science
- Kenneth O. Stanley
- Jeff Clune
- Joel Lehman
- Sebastian Risi
- Rui Wang
- J. Urgen Schmidhuber
- Adam Gaier
- Julian Togelius
In Game & Decision Theory
- J. Urgen Schmidhuber
- Kenneth O. Stanley
- Rui Wang
- Adam Gaier
- Jeff Clune
- Joel Lehman
- Yingtao Tian
- Luke Metz
In Evolutionary Computation
- Jeff Clune
- Sebastian Risi
- Kenneth O. Stanley
- Rui Wang
- Joel Lehman
- Julian Togelius
- J. Urgen Schmidhuber
- Adam Gaier
Citations in the corpus
- Recurrent World Models Facilitate Policy Evolution (2018, methodology): “Recent works [14, 26, 73, 94] have demonstrated that ES is a viable alternative to traditional Deep RL methods on many strong baselines.”
- Weight Agnostic Neural Networks (2019, methodology): “Our specific approach is based on the open source estool [36] implementation of population-based REINFORCE.”
- World Models (2018, methodology): “Recent works (Fernando et al., 2017; Salimans et al., 2017; Ha, 2017b; Stanley & Clune, 2017) have confirmed that ES is a viable alternative to traditional Deep RL methods on many strong baselines.”
- Paired Open-Ended Trailblazer (POET): Endlessly Generating Increasingly Complex and Diverse Learning Environments and Their Solutions (2019, methodology): “Note that ES had previously been shown to be a competent approach [58] on the original Bipedal Walker Hardcore environment in OpenAI Gym, which, as illustrated later in Table 2, consists of similar, but much less difficult obstacles.”
- Increasing generality in machine learning through procedural content generation (2019, background): “e nal reward is the average reward over multiple rollouts. A similar approach for encouraging the discovery of general policies worked well for evolving stable policies for a 2D bipedal walker domain [54]. In addition to helping in supervised learning settings (see Section 3), forms of data augmenta”
- Learning to Predict Without Looking Ahead: World Models Without Forward Prediction (2019, methodology): “While it’s possible to optimize this objective with any reinforcement learning method [39, 40, 54, 55], we choose to use population based REINFORCE [69] due to its simplicity and effectiveness at achieving high scores on various tasks [20, 21, 51].”
- Modern Evolution Strategies for Creativity: Fitting Concrete Images and Abstract Concepts (2021, methodology): “With the recent resurgence of interest in evolution strategies (ES) in the machine learning community [17, 44], in this work, we revisit the use of ES for creativity applications as an alternative to gradient-based methods.”
- POET: open-ended coevolution of environments and their optimized solutions (2019, methodology): “ES had previously been shown to be a competent approach [20] on the original Bipedal Walker Hardcore environment in OpenAI Gym, which, as illustrated later in Table 1, consists of similar, but”
- Neuroevolution of self-interpretable agents (2020, background): “ Neuroevolution-based methods for tackling challenging RL tasks have recently gained popularity due to their simplicity and competitiveness to Deep RL methods, even on vision-based RL benchmark tasks [34, 36, 62, 81, 94]. More recent work [77, 78] demonstrated that evolution can even train RL agents”