An error correcting code is said to be locally testable if there is a test that checks whether a given string is a codeword, or rather far from the code, by reading only a constant number of symbols of the string. Locally Testable Codes (LTCs) were first systematically studied by ... more >>>

We introduce a 2-round stochastic constraint-satisfaction problem, and show that its approximation version is complete for (the promise version of) the complexity class $\mathsf{AM}$. This gives a `PCP characterization' of $\mathsf{AM}$ analogous to the PCP Theorem for $\mathsf{NP}$. Similar characterizations have been given for higher levels of the Polynomial Hierarchy, ... more >>>

We prove almost tight hardness results for finding independent sets in bounded degree graphs and hypergraphs that admit a good
coloring. Our specific results include the following (where $\Delta$, assumed to be a constant, is a bound on the degree, and
$n$ is the number of vertices):

... more >>>

Probabilistically checkable debate systems (PCDSs) are debates between two competing provers, in which a polynomial-time verifier inspects a constant number of bits of the debate. It was shown by Condon, Feigenbaum, Lund, and Shor that every language in PSPACE has a PCDS in which the debate length is polynomially bounded. ... more >>>

We construct a PCP based on the hyper-graph linearity test with 3 free queries. It has near-perfect completeness and soundness strictly less than 1/8. Such a PCP was known before only assuming the Unique Games Conjecture, albeit with soundness arbitrarily close to 1/16.

At a technical level, our ... more >>>

Locally correctable codes (LCCs) are error correcting codes C : \Sigma^k \to \Sigma^n which admit local algorithms that correct any individual symbol of a corrupted codeword via a minuscule number of queries. This notion is stronger than that of locally decodable codes (LDCs), where the goal is to only recover ... more >>>