Richard Watson received a B.A. in Artificial Intelligence in 1990, MSc in Evolutionary and Adaptive Systems 1996, and PhD. Computer Science 2002, before becoming a postdoctoral research fellow in Organismic and Evolutionary Biology at Harvard. Since 2004 he has been at the University of Southampton where he teaches Artificial Intelligence and Evolution. He now has ~80 publications (h-index 38) in evolutionary theory, complex systems and computational optimisation including topics in coevolutionary algorithms, fitness landscapes, artificial life models, dynamical systems, model-building optimisation, modularity, multi-objective optimisation, optimisation, machine learning, and collective robotics – as well as population genetics, the benefit of sexual recombination, evolution of evolvability, evolutionary transitions in individuality, adaptive plasticity, cellular differentiation, social evolution, biological agency, exploratory mechanisms and developmental bias. Esteem factors include: IEEE international award “Ten to Watch in AI”, celebrating 50 years of AI (2006), Best paper in the field of Artificial Life 2016 (International ISAL award), Featured in cover articles of New Scientist magazine 2016 and 2020.
Watson’s research detailed the close relationship between evolution and learning and, in particular, how the action of natural selection on the connections of a gene-regulation network follows the same principles as learning mechanisms on the connections of a neural network. This means that evolution (when acting on the interactions of a dynamical system) has the same memory, learning and generalisation capabilities as learning neural networks. Recently, with Chris Buckley at Sussex, he showed that the same cognitive capabilities of memory, learning and adaptive behaviour occur spontaneously in physical systems under simple conditions (e.g. a network of masses connected by springs, analogous to many natural dynamical systems). He calls this “Natural Induction” to contrast with natural selection and emphasise the learning principles involved. The next step in this line of work is to recognise that if natural systems can learn and adapt without natural selection, and natural selection canalises solutions that have already been learned, then natural selection becomes a follower not a leader in evolutionary change – hence, “cognition-first evolution”. From this point of view, “survival of the fittest” and “what persists exists” fail to capture the prime mover in living systems. These are replaced by ideas like “the differential easing of frustrations between things”, “deeply vulnerable mutual knowing” and “what relates creates”. Most recently, he has been working on a theory of life based on complex harmonic resonance or ‘songs of life and mind’. This aims to develop a calculus relating development (the organisational transformation of material structures), evolution (adaptation and the accumulation of information) and cognition (memory, learning and agential behaviour) into one unified theory.
