I am an assistant professor at Princeton University (on leave) and a
senior scientist at NTT Research.
My primary research focus is cryptography, though I am broadly interested in all aspects of
computer science theory. I received my PhD from Stanford University
and graduated with Highest Honors from UC Berkeley.
|Bounded Storage Model|
Cryptography typically models adversaries as time-bounded, but what about adversaries that are space-bounded? The space-bounded model allows for unconditional and everlasting protocols, sometimes far simpler than their time-bounded counterparts. If we bound time and space, we can also achieve never-before-possible functionalities, such as ciphertexts that effectively disappear after transmission.
Intractible mathematical problems are the heart of modern cryptography. Unfortunately, until someone proves that P≠NP, the intractability of such problems cannot be proven unconditionally and can only be conjectured. Then how do we discover novel mathematical structures, figure out how to use them, and gain confidence in their security? Through extensive study, devising new applications, attacks, and mitigations.
Many of the most practical cryptosystems lack a full security proof in the standard model. Nevertheless, we can gain confidence in their security by heuristically treating one or more of the building blocks as an "ideal" object implemented as an oracle. Prominent examples include random oracles, ideal ciphers, generic groups, etc. Proofs in idealized models are often very different from standard crypto proofs, requiring both reductions and query complexity arguments.
Can you hide secrets in software code? Program obfuscation aims to do exactly that. In addition to direct applications such as protecting intellectual property in software, obfuscation also has numerous applications in cryptography. In fact, it is now widely considered to be "Crypto Complete." Yet numerous quesitons still remain: can obfuscation be made practical? Can it be based on "standard" computational assumptions?
|Property Preserving Encryption|
Property Preserving Encryption (PPE) deliberately preserves certain relations on the plaintext data (e.g. equalities in the case of deterministic encryption, or order in the case of order revealing encryption). In addition to applications such as encrypted databases, PPE also has other interesting connections, such as security under bad randomness and differential privacy. The question is then: how to reveal such information without revealing other sensitive data, and what security, if any, remains.
Quantum computers harnesses the strange features of quantum mechanics, such as superpositions, entanglement, etc. Post-Quantum Cryptography aims to secure cryptosystems against the enhanced power of such computers. On the other hand, Quantum Cryptography uses quantum computing to achieve never-before-possible cryptographic functionalities, such as programs that cannot be copied.
Traitor tracing systems seek to deter piracy by enabling content distributors to identify the origin of pirate decryption boxes. The "usual" goal in traitor tracing is to achieve the shortest ciphertexts, secret keys, and public keys possible. But there is also a rich set of questions beyond parameter sizes: how to embed arbitrary information into a secret key? How to keep honest users' information private, while exposing traitors'? What happens when the decoder uses a quantum computer?