A network-based SIR simulation of a COVID-19 outbreak. A network-based SIR simulation of a COVID-19 outbreak.

Research

My doctoral research focuses on pricing and auctions, specifically providing complexity bounds and designing effective algorithms for determining viability of mechanisms in the dual settings of social welfare and revenue maximisation. This stimulating and active area of research has an abundance of open problems and deep implications for the world outside of academia.

My work on the Product-Mix Auction (PMA), a multi-unit, multi good auction initially proposed by Paul Klemperer for the Bank of England to provide liquidity to financial institutions during the crisis of 2007-8, focuses on finding effective algorithms to solve the auction with respect to social welfare and revenue maximisation. An implementation for social welfare is available here. In related work, I also study how bidder demand can be learnt via queries to a demand or valuation oracle, allowing bidders to participate in PMAs without having to understand the bidding language used.

I am also interested in mechanisms that maximise the benefits of scarce resources in low to middle income countries (LMICs). In response to the COVID-19 pandemic, I co-founded the project Test and Contain, which considers the problem of utilising limited testing resources in an optimal way so as to minimise the impact on the health and livelihoods of those who are hardest hit in LMICs. This project is supported by an ACM SIGecom GCEC’20 grant. For more information, watch the 1 minute poster video submitted to GCEC’20 (received best poster video award).

Publications

1. Test and Contain: A Resource-Optimal Testing Strategy for COVID-19 link
Jakob Jonnerby, Philip Lazos, Edwin Lock, Francisco Marmolejo-Cossío, C. Bronk Ramsey and Divya Sridhar.
AI for Social Good '20, Harvard CRCS Workshop, 2020
2. Characterising and recognising game-perfect graphs link
Stephan Dominique Andres and Edwin Lock.
Discrete Mathematics and Theoretical Computer Science (DMTCS), vol. 21:1, #6, 2019

Working Papers

3. Maximising the Benefits of an Acutely Limited Number of COVID-19 Tests link
Jakob Jonnerby, Philip Lazos, Edwin Lock, Francisco Marmolejo-Cossío, C. Bronk Ramsey, Meghana Shukla and Divya Sridhar.
ArXiv preprint, 2020
4. Learning Strong Substitutes Demand via Queries link
Paul W Goldberg, Edwin Lock and Francisco Marmolejo-Cossío.
ArXiv preprint, 2020
5. Solving Strong-Substitutes Product-Mix Auctions link
Elizabeth Baldwin, Paul W Goldberg, Paul Klemperer and Edwin Lock.
ArXiv preprint, 2019

↳ Full CV