Aarhus University Seal

Pouya Khani's PhD defence: Explainable Query Answering: drivers behind answers from databases and data streams

On Wednesday, November 19, Pouya Khani will defend his PhD Transparent Data Management and Verification

Info about event

Time

Thursday 19 November 2026,  at 09:00

Location

Building 5342, room 333 Ada, Department of Computer Science, Åbogade 34, Aarhus University, 8200 Aarhus N

Organizer

Department of Computer Science, Aarhus University

During his PhD studies, Pouya researched how to make database systems explain their own answers. Databases often turn millions of records into a single number – for example an average salary, total sales, or the average response time of an online service. Such summaries are useful, but they hide which parts of the data actually drive the result. Pouya developed methods that point to concrete groups in the data – for instance a particular job role, region or type of request – and show how much each group contributes to the observed number. One part of the work uses partial knowledge about cause and effect to separate factors that truly matter from misleading correlations. Another part makes the same kind of explanation fast enough for live data streams, where numbers keep changing as new information arrives. 

The research can help analysts, operators and decision-makers understand not only what the data says, but why.