RESEARCH
Language shapes
understanding.
Our research asks how AI can help people connect events, compare perspectives, and examine evidence across sources and languages.
Events & framing
The same event can be told in very different ways.
One account may foreground a decision; another may focus on its consequences. We study how descriptions across documents refer to related events, and how wording, attribution, and context guide interpretation.
Our interests connect natural language processing with computational framing: studying the choices that shape how an event is presented.
When two articles discuss the same event, what do they each emphasize—and what does each leave in the background?
Meaning across languages
Understanding language means paying attention to its differences.
We are interested in how events and perspectives are expressed across languages and language varieties. This includes questions about linguistic meaning, multilingual language technologies, and whose ways of communicating are represented in AI.
Our research asks how computational approaches can preserve distinctions that matter to the people using a language.
How might two accounts in different languages describe the same event while expressing different perspectives?
Evidence & human understanding
People need ways to inspect the information behind an interpretation.
We are interested in language technologies that help people connect claims with relevant sources, compare accounts, and reason about uncertainty. We investigate how AI can support these activities while keeping the underlying language and evidence accessible.
This direction brings together language understanding and human-centered AI.
What would help a reader see which parts of an account support a claim, and which parts require further investigation?
RESEARCH BACKGROUND
Explore the work
behind our interests.
Jin Zhao’s personal publication list provides background on the research informing these interests, including work that predates the lab.
Jin Zhao’s publications