Our research spans the full stack of that question — from how agents are built, to how they team up with people in consequential decisions, to how societies adapt when agents become everyday allies.
How IR techniques enhance agent capabilities, focusing on retrieval-augmented generation and knowledge grounding.
Designing and optimizing agentic frameworks powered by large language models for improved reasoning and task execution.
Benchmarking and enhancing IR-LLM integrated systems in real-world agent applications.
Enhancing human cognitive capabilities through agent collaboration — complex problem-solving and creative thinking.
Frameworks for human-agent collaboration in critical scenarios where decisions have significant consequences.
Seamless, intuitive interfaces that enable natural and effective communication between humans and agents.
How the integration of agents reshapes social structures, norms, and relationships.
The economic, cultural, and ethical implications of widespread agent adoption in everyday life.
Governance, policy, and regulatory approaches for responsible agent deployment at societal scale.
We ground our research in high-stakes domains — finance, law, policy, and scholarly communication — where getting human-agent collaboration right matters most.
We create the datasets and shared tasks the community evaluates against — FinNum, FinArg, NumEval, PromiseEval, RegCom, and more at NTCIR and SemEval.
Our work appears at ACL, EMNLP, SIGIR, WWW, and NAACL — spanning NLP, IR, and web research communities.
Joint projects on business ideation, live commentary generation, and human-centered AI with industrial partners.
We're always looking for passionate researchers and partners who share our vision of creating a better human-agent society.