Projects

SESiL: Social, Evolutionary Supported, Learning

Social learning is characterised by agents observing, adapting and absorbing the skill and knowledge of others. But what if an agent cannot directly absorb the desired capabilities? The answer: agree with the source of the attractive skill/knowledge and pass a combination of your capabilities to your children.

Our first paper:

"SESiL: Social, Evolutionary Supported Learning", by Tianshu Zhao and Zinovi Rabinovich, in AAMAS 2026

Frankenmandering

Initially funded by an CU Development Grant, this project investigates the interference between two common modes of interaction between opinions: social networks and (district based) voting. In particular, we are interested in how commonplace manipulations can be boosted or suppressed, when these interaction modes co-exist.

Our motivational paper:

"Frankenmandering: Repeated Social Graph Gerrymandering", by Sahil Agarwal, Svetlana Obraztsova , Zinovi Rabinovich and Alan Tsang, COMSOC 2025

Choice Manipulation and Security Games

This was an umbrella project of our lab @NTU, under which other ideas germinated before they grew onto their own path. The original theme of the project was the use of information disclosure as means of control in Security Game-like scenarios.

Mixed Initiative DCOPs: Incorporating decision support by humans into AI-oriented algorithms for distributed constraint optimisation

This was a project under an AcRF Tier-1 grant from the Ministry of Education, Singapore. We sought to enable distributed constraint optimisation algorithms to handle heterogeneity of its computational nodes. Of particular concern was the ability of computational nodes to process and communicate information. This was a small-scale proof-of-idea project.

Report abuse