Dr Ashvin Shah
This person has left and cannot be contacted through the department. If they have provided contact details, they will be displayed below.
I can still be reached at: a.shah at sheffield dot ac dot uk.
My website is at this link.
I joined the ABRG in the summer of 2010. Although my contract ended in the spring of 2013, I continue with work with Kevin Gurney and colleagues in finishing up some projects. I recieved my PhD in 2008 from the Neuroscience and Behavior Program at the University of Massachusetts Amherst (in the US) under the supervision of Andrew Barto. I am interested in how computational mechanisms that we think are implemented in brain areas are used to develop behaviour we see in animals, specifically in animals engaged in learning tasks. (And yes, humans are animals.) For my PhD research, I looked at motor skill acquisition by implementing models on a fairly abstract level, similar to those used in reinforcement learning research within computer science. The abstract to my thesis provides a better description of that stuff. Over here in the ABRG, I will use more realistic models to investigate how we can develop novel "actions," i.e., discover the specific behavior that we just did that resulted in something interesting that just happened. A possible mechanism by which this can happen is through a very short-latency and short-lasting dopamine signal that reinforces behaviors, as discussed in Redgrave et al. (2008) and Redgrave and Gurney (2006). The simulations we will do are closely related to experimental work being done by Peter Redgrave and colleagues here at Sheffield.
- I was a postdoc working primarily with Kevin Gurney on a project called Intrinsically-Motivated Cumulative Learning Versatile Robots (IM-CLeVeR).
- Research interests include computational neuroscience, action discovery / skill acquisition, reinforcement learning, decision making, and motor control.
- Ashvin Shah and Kevin N. Gurney (2014), Emergent Structured Transition from Variation to Repetition in a Biologically-Plausible Model of Learning in Basal Ganglia, Frontiers in Psychology (Cognitive Science), 5:91.
- A. Shah, K. N. Gurney, (2014), Finding minimal action sequences with a simple evaluation of actions, Frontiers in Computational Neuroscience, 8, 151.
- Ashvin Shah, Andrew G. Barto, and Andrew H. Fagg (2013), A Dual Process Account of Coarticulation in Motor Skill Acquisition, Journal of Motor Behavior, volume 45, pages 531--549.
- Martin Thirkettle, Thomas Walton, Ashvin Shah, Kevin Gurney, Peter Redgrave, Tom Stafford (2013), The path to learning: Action acquisition is impaired when visual reinforcement signals must first access cortex, Behavioural Brain Research, volume 243, pages 267--272.
- Gurney K, Lepora N, Shah A, Koene A and Redgrave P (2013), Action Discovery and Intrinsic Motivation: A Biologically Constrained Formalisation, Intrinsically Motivated Learning in Natural and Artificial Systems, G Baldassarre and M Mirolli, editors. Chapter 7, pages 151--181. Published by Springer-Verlag, Berlin Heidelberg. .
- Ashvin Shah (2012), Psychological and Neuroscientific Connections with Reinforcement Learning , in Marco Wiering and Martijn van Otterlo, editors, Reinforcement Learning: State of the Art, Chapter 16, pages 507--537, Springer-Verlag, Berlin Heidelberg.
- Ashvin Shah and Kevin Gurney (2011), Dopamine-mediated action discovery promotes optimal behavior ‘for free’, Twentieth Annual Computational Neuroscience Meeting (CNS*2011), Poster presentation. Abstract available in BMC Neuroscience 2011, 12 (Suppl 1):P138 (18 July 2011). .