7 papers
How to Avoid Debate: Scalable AI Safety via Doubly-Efficient Interactive Proofs
Liyan Chen, Yael Tauman Kalai, Zoe Xi
As AI models continue to develop powerful capabilities, it becomes critical that we are able to verify that their output is aligned with our intentions. A recent line of work focus…
Towards a Doubly Efficient IP=PSPACE
Liyan Chen, Matthew M. Hong, Yael Tauman Kalai +1
We show that every language in PSPACE decidable by a Turing machine in time admits a doubly efficient interactive proof system: the prover runs in time polynom…
Consensus Sampling for Safer Generative AI
Adam Tauman Kalai, Yael Tauman Kalai, Or Zamir
Motivated by undetectable risks in generative AI, we study a general robust aggregation problem: how to aggregate several probability distributions to boost safety. We present cons…
How to Classically Verify a Quantum Cat without Killing It
Yael Tauman Kalai, Dakshita Khurana, Justin Raizes
Existing protocols for classical verification of quantum computation (CVQC) consume the prover's witness state, requiring a new witness state for each invocation. Because QMA witne…
Efficiently Batching Unambiguous Interactive Proofs
Bonnie Berger, Rohan Goyal, Matthew M. Hong +1
We show that if a language admits a public-coin unambiguous interactive proof (UIP) with round complexity , where bits are communicated per round, then the batch lang…
Compiling Any into a (Succinct) Classical Interactive Argument
Andrew Huang, Yael Tauman Kalai
We present a generic compiler that converts any protocol into a succinct interactive argument where the communication and the verifier are classical, and where p…