Responsible Human-Machine Teaming
Tuesday 01 Oct 2019
Time: 09:45 - 17:30
Bio
Prof. Sarvapali Ramchurn is a Professor of Artificial Intelligence in the Agents, Interaction, and Complexity research group where he carries out research into the design of autonomous agents and multi-agents for real-world socio-technical applications including energy systems, disaster management, and crowdsourcing. He has won multiple best paper awards for his work and is a recipient of the prestigious AXA Research Fund Award for Responsible Artificial Intelligence. He works closely with industry and his research touches on a number of fields including Machine Learning, Data Science, and Game Theory. Specifically, he has pioneered the development of agent-based coordination algorithms for distributed task allocation that have been deployed on real-world unmanned aerial vehicles and in the Premier League’s Fantasy Football game where his approach has been shown to outperform more than 5M human players. His papers have been cited more than 6000 times (according to Google scholar) and his work has featured in various media including BBC News, New Scientist, Sky News, BBC Click, and Wired.
Research interests
To support the management of large numbers of robots, Artificial Intelligence (AI) algorithms have been developed to automate the actions of such robot swarms and enable to act in a cohesive and coordinated way. AI thus allows swarms to allocate tasks to each other, react to losses in communication or resources (e.g., other members of the swarm) and therefore reduce the workload of their human counterparts. However, it has been shown that, in some situations, operators are overwhelmed with information coming from robots, may not completely trust their decisions, and therefore override them; by doing so, they may cause the system to fail. In other situations, humans may completely rely on the automation and fail to notice obvious errors in the system. Moreover, to manage such large fleets of robots in a safe manner, previous studies suggest there should be shifts in autonomy levels to allow humans to take corrective action to recover from failures or to optimize task performance. Understanding when such shifts should occur without losing out on the fault-tolerance benefits of a decentralized swarm, what levels of workload these shifts induce, and how the team of operators should enact such shifts are key questions that will be addressed in this project.
Achievements and awards
- Winner of the AXA Research Award for achievements in Responsible Artificial Intelligence (2017)
- Runner up for the best paper award at the 2016 Humanitarian Technology conference (with my student Elliot Salisbury and Seb Stein).
- Honourable mention for the IJCAI-JAIR best paper 2016 award
- Winner of the Best Paper (Applications Track) at AAMAS 2015
- Winner of the best student paper award (with my student Muddasser Alam) at AAMAS 2013 for work on energy exchange in rural communities (out of 685 papers submitted)
- Runner up of the TREC 2012 (Text Retrieval Conference) Crowdsourcing challenge (AUC metric).
- Runner up for the best paper award at AAMAS 2011 (Applications) for work on agent-based optimisation for demand-side management.
- Winner of the best paper award at AAMAS 2010 for work on agent-based energy storage.
- Winner of the Infrastructure Competition of the RoboCupRescue league at RoboCup 2007 – RoboCup is an international competition held every year and attracts more than a thousand participants from more than 30 different countries.
- Winner of the International Prisoner’s Dilemma Competition (Comp 2) at the IEEE Symposium on Computational Intelligence and Games 2005)
- Winner of the International Prisoner’s Dilemma competition at the Congress on Evolutionary Computing (CEC 2004) – we designed the first strategy to beat Tit-For-Tat in 20 years since the original Axelrod experiment.
