ScAI

Talk

Title: Trust in an Untrusted World: Private Access over Public Infrastructures

Abstract:

We are living in an era where our digital lives are increasingly interdependent

and deeply interconnected. These connections rely on a vast, layered ecosystem of

actors—many of whose trustworthiness is uncertain or outright suspect. Over the past three

decades, rapid advances in computing and communication technologies have brought

unprecedented access and connectivity to billions of users. Yet this digitization comes at a cost:

our interactions, queries, and data are increasingly vulnerable to privacy violations. Today,

threats to privacy come not just from malicious individuals, but also from powerful

institutions—ranging from service providers to nation-states.

In this reality of an untrusted world, we pose several foundational research questions: (i) Can

we design a scalable voice communication system that ensures absolute privacy? (ii) Can

we build an oblivious search engine over public document repositories? (iii) Can we develop

scalable private query processing over shared or public databases? (iv) And in the age of

large language models, can we enable private inference for user queries? These are not just

open problems — they are essential challenges if we are to build trusted services over

untrusted infrastructures. In this talk, I will present recent work that leverages Homomorphic

Encryption to address some of these questions. We explore the inherent performance and

scalability trade-offs in enabling private access, search, and inference. If nothing else, our

results underscore a critical insight: ensuring privacy at scale is not impossible, but it

comes at a high cost.

Speaker Bio:

Divy Agrawal is a Distinguished Professor and Chair of Computer Science at the University of California,

Santa Barbara (UCSB), where he also holds the Leadership Endowed Chair in the Department of

Computer Science. He received his B.E. (Hons.) in Electrical Engineering from BITS Pilani, followed by

M.S. and Ph.D. degrees in Computer Science from the State University of New York at Stony Brook.

Since joining UCSB, Professor Agrawal has established himself as a leading researcher in databases,

distributed systems, cloud computing, and large-scale data infrastructures and analytics. Over the course

of his career, he has published more than 400 research articles and mentored approximately 50 Ph.D.

students. He currently serves as Editor-in-Chief of both the Proceedings of the ACM on Modeling of Data

and the Springer journal Distributed and Parallel Databases. He has served on several editorial boards,

including ACM Transactions on Database Systems, IEEE Transactions on Knowledge and Data

Engineering, ACM Transactions on Spatial Algorithms and Systems, ACM Books, and the VLDB Journal.

Professor Agrawal is a former Trustee of the VLDB Endowment and recently served as Chair of the ACM

Special Interest Group on Management of Data (SIGMOD). His recognitions include the Gold Medal from

BITS Pilani, the UCSB Academic Senate Award for Outstanding Graduate Mentoring, and multiple paper

honors: Best Paper Awards (ICDE 2002, MDM 2011), an Influential Paper Award (NDSS 2024), and

Test-of-Time Awards (ICDT, MDM). He is a Fellow of the ACM, IEEE, and AAAS.

Events