This project seeks to develop new computational models of adaptive exploratory mechanisms that can accurately capture how these iterative functional responses to unanticipated circumstances originate and exhibit feedback effects on both developmental and evolutionary time scales. In particular, it will generate the first models of the inherent trade-offs between being biased to produce particular responses efficiently and the cost of being able to explore novel responses when needed. A key output of this work will be computational models based on insights from machine learning theory regarding multiple adaptive timescales in learning systems that demonstrate the conditions under which ‘explorability’ evolves and thereby suggest novel experimental tests for confirming its significance for evolutionary theory.
Kevin Lala
Agential Behavior and Plasticity in Evolution
Exploratory mechanisms, agency, and evolution
Subaward Principal Investigator
University of St Andrews
Kevin Lala is Professor of Behavioural and Evolutionary Biology at the University of St Andrews, and prior to that held positions at UCL, UC Berkeley and Cambridge Universities. His principle academic interests are in the general area of animal behaviour and evolution, with a specific focus on: (i) animal social learning, innovation and intelligence, (ii) niche construction, inclusive inheritance and the extended evolutionary synthesis, and (iii) human evolution, particularly the evolution of cognition. He has published c. 300 scientific articles on these topics, been the recipient of more than £17m in grant income, and authored 12 books, including Darwin’s Unfinished Symphony. How Culture Made the Human Mind (Princeton UP 2017), and Niche Construction: The Neglected Process in Evolution (with John Odling-Smee and Marc Feldman, Princeton UP, 2003). Laland is an Elected Fellow of the Royal Society of Edinburgh and a Fellow of the Royal Society of Biology.
Richard Watson
Agential Behavior and Plasticity in Evolution
Evolution and organismal goal-directedness, Exploratory mechanisms, agency, and evolution
Subaward Principal Investigator
University of Southampton
Dr Richard Watson studies evolution, learning, cognition and society and their unifying algorithmic principles. He studied Artificial Intelligence and Adaptive Systems at Sussex University, then PhD Computer Science at Brandeis in Boston. His current work deepens the unification of evolution and learning - specifically, with connectionist models of learning and cognition, familiar in neural network research – to address topics such as evolvability, ecological memory, evolutionary transitions in individuality (ETIs), phenotypic plasticity, the extended evolutionary synthesis, collective intelligence and 'design'. He has also developed new computational methods for combinatorial optimisation (deep optimisation), exploiting a unification of deep learning and ‘deep evolution’ (i.e. ETIs). He is author of "Compositional evolution" (MIT Press), was featured as "one to watch in AI” in Intelligent Systems magazine, and his paper “How Can Evolution Learn” in TREE, attracted the ISAL award 2016. He is now Associate Professor at the University of Southampton.