Adam Stanton, Convener and Permanent Member (ORCiD)
Welcome to the homepage of the Adaptive Informatics Research Collective.
Efficient Adaptive Sampling Lexicase Selection in Evolutionary Robotics (2025) DOI | PDF
Fitness Agnostic Adaptive Sampling Lexicase Selection (2023) DOI | PDF
Lexicase Selection for Multi-task Evolutionary Robotics (2022) DOI | PDF
Where do successful populations originate from? (2021) DOI | PDF
Objective Sampling Strategies for Generalized Locomotion Behavior with Lexicase Selection (2021) DOI | PDF
When Specialists Transition to Generalists: Evolutionary Pressure in Lexicase Selection (2020) DOI | PDF
Viral RNA-dependent RNA polymerase mutants display an altered mutation spectrum resulting in attenuation in both mosquito and vertebrate hosts (2019) DOI | PDF
The Limits of Lexicase Selection in an Evolutionary Robotics Task (2019) DOI | PDF
Stochastic Ontogenesis in Evolutionary Robotics (2018) DOI | PDF
Tiebreaks and Diversity: Isolating Effects in Lexicase Selection (2018) DOI | PDF
RecQ helicases in the malaria parasite Plasmodium falciparum affect genome stability, gene expression patterns and DNA replication dynamics (2018) DOI | PDF
Lexicase Selection Outperforms Previous Strategies for Incremental Evolution of Virtual Creature Controllers (2017) DOI | PDF
Neuroevolution of Feedback Control for Object Manipulation by 3D Agents (2016) DOI | PDF
Open-Ended Evolution: Perspectives from the OEE Workshop in York (2016) DOI | PDF
Recombination events among virulence genes in malaria parasites are associated with G-quadruplex-forming DNA motifs (2016) DOI | PDF
Incremental Neuroevolution of Reactive and Deliberative 3D Virtual Creatures (2015) DOI | PDF
Heterogeneous complexification strategies robustly outperform homogeneous strategies for incremental evolution (2013) DOI | PDF
We apply modern machine learning techniques to real-world problems, and try to do research in bio-inspired, generative approaches to autonomous systems. We experiment with algorithms based on principles of evolution to build machines that have intelligence that shows up through animal-like behaviours. This approach can discover parameters for simulated robots that enable basic patterns like walking, and is a promising approach to finding naturalistic planning and problem-solving behaviours.
We are also interested in finding convivial alternatives to modern-day "big tech" products, and in resisting the intrusion of these technologies into our lives. We explicitly reject the use of language-generating models for any purpose and advocate low-tech, low-impact, sustainable approaches wherever possible. We support the ideas of permacomputing, localism, and community in technology.
Adam Stanton, Convener and Permanent Member (ORCiD)
Whilst "artificial intelligence" is prominent at the moment, we choose Adaptive Informatics, or AI, as the umbrella term for our activity. Adaptation to data is at the heart of machine learning, and radical adaptation to adaptation, as well as the rest of the physical universe, is a strong description of evolution in nature.
We're nowhere near to social media like insta, twitter, linkdin and so on but we found a home in the fediverse.