Dr. Ang Li, an Assistant Professor in the Computer Science Department, and his research lab have recently published their work in the prestigious Thirty-Eighth AAAI Conference on Artificial Intelligence (AAAI-24). AAAI is recognized as one of the top conferences in the field of artificial intelligence. Their research paper, titled “Unit Selection with Nonbinary Treatment and Effect,” co-authored with Judea Pearl, a Turing Award recipient and Dr. Li’s Ph.D. advisor, marks a significant contribution.

The binary unit selection model, introduced by Li and Pearl in 2019, evaluates the weighted composition of four counterfactual response types: complier, always-taker, never-taker, and defier, in a given population. They demonstrated the model’s superiority over the widely-used A/B testing heuristic in various applications, including vaccination effectiveness, advertisement promotion, and customer relationship management. This latest paper extends the unit selection model and its objective function in a more general form, not limited to binary cases. It provides two algorithms for identifying and bounding the objective function, enabling the model’s application to real datasets, which often involve non-binary treatments and effects.

The paper was presented at the AAAI 2024 conference in Vancouver, Canada, in February and is set to be published by the AAAI Press in the conference proceedings.