MPLS and Medical Sciences Graduate Seminar

All College members are welcome to come along to this term's MPLS and Medical Sciences Graduate Seminar, where two of LMH's DPhil students will present their research. 

Kai Sandbrink, DPhil in Experimental Psychology

Modelling Cognitive Flexibility with Deep Neural Networks

Humans are capable of adapting quickly to changes in their environment. This ability is called cognitive flexibility. In contrast, to adapt to change, typical artificial neural systems respond by updating a large number of parameters, a slow process that involves forgetting past information. In this talk, we consider a simple model of how rapid behavioural adaptation is possible in linear neural networks with fast and bounded units. These units form abstract task representations gating existing knowledge. After studying the formation of these task abstractions, we will briefly consider how these principles can be extended to understanding adaptation in more complex architectures.

Khadija Sarguroh, DPhil in Engineering Science

Engineering Entanglement in Quantum Dots for Quantum Computing

Quantum dots are a promising platform for quantum computing, where both electronic and a large register of nuclear spins can be used as qubits. This system offers fast optical control over electronic spins and provides access to a long-lived quantum memory in the nuclear spins, which can store and process quantum information. The always-on nature of interactions between electronic and nuclear states makes controlling individual states challenging. By using dynamical decoupling techniques, we can refine control over these interactions. Additionally, using an approach called one-tangles, we can measure and optimize entanglement to help design reliable quantum operations.