arXiv subjects

Ran Raz

Publications and source records attributed to Ran Raz.

29 records · Page 2Linked to original sources

arXiv2017

Extractor-Based Time-Space Lower Bounds for Learning

A matrix $M: A \times X \rightarrow \{-1,1\}$ corresponds to the following learning problem: An unknown element $x \in X$ is chosen uniformly at random. A learner tries to learn $x$ from a stream of samples, $(a_1, b_1), (a_2, b_2) \ldots$, where for every $i$, $a_i \in A$ is chosen uniformly at random and $b_i = M(a_i,x)$. Assume that $k,\ell, r$ are such that any submatrix of $M$ of at least $2^{-k} \cdot |A|$ rows and at least $2^{-\ell} \cdot |X|$ columns, has a bias of at most $2^{-r}$. We show that any learning algorithm for the learning problem corresponding to $M$ requires either a memory of size at least $Ω\left(k \cdot \ell \right)$, or at least $2^{Ω(r)}$ samples. The result holds even if the learner has an exponentially small success probability (of $2^{-Ω(r)}$). In particular, this shows that for a large class of learning problems, any learning algorithm requires either a memory of size at least $Ω\left((\log |X|) \cdot (\log |A|)\right)$ or an exponential number of samples, achieving a tight $Ω\left((\log |X|) \cdot (\log |A|)\right)$ lower bound on the size of the memory, rather than a bound of $Ω\left(\min\left\{(\log |X|)^2,(\log |A|)^2\right\}\right)$ obtained in previous works [R17,MM17b]. Moreover, our result implies all previous memory-samples lower bounds, as well as a number of new applications. Our proof builds on [R17] that gave a general technique for proving memory-samples lower bounds.

cs.LG↗

arXiv2016

Fast Learning Requires Good Memory: A Time-Space Lower Bound for Parity Learning

We prove that any algorithm for learning parities requires either a memory of quadratic size or an exponential number of samples. This proves a recent conjecture of Steinhardt, Valiant and Wager and shows that for some learning problems a large storage space is crucial. More formally, in the problem of parity learning, an unknown string $x \in \{0,1\}^n$ was chosen uniformly at random. A learner tries to learn $x$ from a stream of samples $(a_1, b_1), (a_2, b_2) \ldots$, where each~$a_t$ is uniformly distributed over $\{0,1\}^n$ and $b_t$ is the inner product of $a_t$ and $x$, modulo~2. We show that any algorithm for parity learning, that uses less than $\frac{n^2}{25}$ bits of memory, requires an exponential number of samples. Previously, there was no non-trivial lower bound on the number of samples needed, for any learning problem, even if the allowed memory size is $O(n)$ (where $n$ is the space needed to store one sample). We also give an application of our result in the field of bounded-storage cryptography. We show an encryption scheme that requires a private key of length $n$, as well as time complexity of $n$ per encryption/decription of each bit, and is provenly and unconditionally secure as long as the attacker uses less than $\frac{n^2}{25}$ memory bits and the scheme is used at most an exponential number of times. Previous works on bounded-storage cryptography assumed that the memory size used by the attacker is at most linear in the time needed for encryption/decription.

cs.LG↗

arXiv2015

Welfare Maximization with Limited Interaction

We continue the study of welfare maximization in unit-demand (matching) markets, in a distributed information model where agent's valuations are unknown to the central planner, and therefore communication is required to determine an efficient allocation. Dobzinski, Nisan and Oren (STOC'14) showed that if the market size is $n$, then $r$ rounds of interaction (with logarithmic bandwidth) suffice to obtain an $n^{1/(r+1)}$-approximation to the optimal social welfare. In particular, this implies that such markets converge to a stable state (constant approximation) in time logarithmic in the market size. We obtain the first multi-round lower bound for this setup. We show that even if the allowable per-round bandwidth of each agent is $n^{ε(r)}$, the approximation ratio of any $r$-round (randomized) protocol is no better than $Ω(n^{1/5^{r+1}})$, implying an $Ω(\log \log n)$ lower bound on the rate of convergence of the market to equilibrium. Our construction and technique may be of interest to round-communication tradeoffs in the more general setting of combinatorial auctions, for which the only known lower bound is for simultaneous ($r=1$) protocols [DNO14].

cs.GT↗

arXiv2013

Arthur-Merlin Streaming Complexity

We study the power of Arthur-Merlin probabilistic proof systems in the data stream model. We show a canonical $\mathcal{AM}$ streaming algorithm for a wide class of data stream problems. The algorithm offers a tradeoff between the length of the proof and the space complexity that is needed to verify it. As an application, we give an $\mathcal{AM}$ streaming algorithm for the \emph{Distinct Elements} problem. Given a data stream of length $m$ over alphabet of size $n$, the algorithm uses $\tilde O(s)$ space and a proof of size $\tilde O(w)$, for every $s,w$ such that $s \cdot w \ge n$ (where $\tilde O$ hides a $\polylog(m,n)$ factor). We also prove a lower bound, showing that every $\mathcal{MA}$ streaming algorithm for the \emph{Distinct Elements} problem that uses $s$ bits of space and a proof of size $w$, satisfies $s \cdot w = Ω(n)$. As a part of the proof of the lower bound for the \emph{Distinct Elements} problem, we show a new lower bound of $Ω(\sqrt n)$ on the $\mathcal{MA}$ communication complexity of the \emph{Gap Hamming Distance} problem, and prove its tightness.

cs.CC↗

arXiv2012

Label Cover instances with large girth and the hardness of approximating basic k-spanner

We study the well-known Label Cover problem under the additional requirement that problem instances have large girth. We show that if the girth is some $k$, the problem is roughly $2^{\log^{1-ε} n/k}$ hard to approximate for all constant $ε> 0$. A similar theorem was claimed by Elkin and Peleg [ICALP 2000], but their proof was later found to have a fundamental error. We use the new proof to show inapproximability for the basic $k$-spanner problem, which is both the simplest problem in graph spanners and one of the few for which super-logarithmic hardness was not known. Assuming $NP \not\subseteq BPTIME(2^{polylog(n)})$, we show that for every $k \geq 3$ and every constant $ε> 0$ it is hard to approximate the basic $k$-spanner problem within a factor better than $2^{(\log^{1-ε} n) / k}$ (for large enough $n$). A similar hardness for basic $k$-spanner was claimed by Elkin and Peleg [ICALP 2000], but the error in their analysis of Label Cover made this proof fail as well. Thus for the problem of Label Cover with large girth we give the first non-trivial lower bound. For the basic $k$-spanner problem we improve the previous best lower bound of $Ω(\log n)/k$ by Kortsarz [Algorithmica 1998]. Our main technique is subsampling the edges of 2-query PCPs, which allows us to reduce the degree of a PCP to be essentially equal to the soundness desired. This turns out to be enough to essentially guarantee large girth.

cs.DS↗

arXiv2010

The Surprise Examination Paradox and the Second Incompleteness Theorem

We give a new proof for Godel's second incompleteness theorem, based on Kolmogorov complexity, Chaitin's incompleteness theorem, and an argument that resembles the surprise examination paradox. We then go the other way around and suggest that the second incompleteness theorem gives a possible resolution of the surprise examination paradox. Roughly speaking, we argue that the flaw in the derivation of the paradox is that it contains a hidden assumption that one can prove the consistency of the mathematical theory in which the derivation is done; which is impossible by the second incompleteness theorem.

math.LO↗

arXiv2008

Exponential separations for one-way quantum communication complexity, with applications to cryptography

We give an exponential separation between one-way quantum and classical communication protocols for a partial Boolean function (a variant of the Boolean Hidden Matching Problem of Bar-Yossef et al.) Earlier such an exponential separation was known only for a relational problem. The communication problem corresponds to a \emph{strong extractor} that fails against a small amount of \emph{quantum} information about its random source. Our proof uses the Fourier coefficients inequality of Kahn, Kalai, and Linial. We also give a number of applications of this separation. In particular, we show that there are privacy amplification schemes that are secure against classical adversaries but not against quantum adversaries; and we give the first example of a key-expansion scheme in the model of bounded-storage cryptography that is secure against classical memory-bounded adversaries but not against quantum ones.

quant-ph↗

arXiv2007

Resolution over Linear Equations and Multilinear Proofs

We develop and study the complexity of propositional proof systems of varying strength extending resolution by allowing it to operate with disjunctions of linear equations instead of clauses. We demonstrate polynomial-size refutations for hard tautologies like the pigeonhole principle, Tseitin graph tautologies and the clique-coloring tautologies in these proof systems. Using the (monotone) interpolation by a communication game technique we establish an exponential-size lower bound on refutations in a certain, considerably strong, fragment of resolution over linear equations, as well as a general polynomial upper bound on (non-monotone) interpolants in this fragment. We then apply these results to extend and improve previous results on multilinear proofs (over fields of characteristic 0), as studied in [RazTzameret06]. Specifically, we show the following: 1. Proofs operating with depth-3 multilinear formulas polynomially simulate a certain, considerably strong, fragment of resolution over linear equations. 2. Proofs operating with depth-3 multilinear formulas admit polynomial-size refutations of the pigeonhole principle and Tseitin graph tautologies. The former improve over a previous result that established small multilinear proofs only for the \emph{functional} pigeonhole principle. The latter are different than previous proofs, and apply to multilinear proofs of Tseitin mod p graph tautologies over any field of characteristic 0. We conclude by connecting resolution over linear equations with extensions of the cutting planes proof system.

cs.CC↗

arXiv2006

The one-way communication complexity of the Boolean Hidden Matching Problem

We give a tight lower bound of Omega(\sqrt{n}) for the randomized one-way communication complexity of the Boolean Hidden Matching Problem [BJK04]. Since there is a quantum one-way communication complexity protocol of O(\log n) qubits for this problem, we obtain an exponential separation of quantum and classical one-way communication complexity for partial functions. A similar result was independently obtained by Gavinsky, Kempe, de Wolf [GKdW06]. Our lower bound is obtained by Fourier analysis, using the Fourier coefficients inequality of Kahn Kalai and Linial [KKL88].

quant-ph↗

arXiv2005

Quantum Information and the PCP Theorem

We show how to encode $2^n$ (classical) bits $a_1,...,a_{2^n}$ by a single quantum state $|Ψ>$ of size O(n) qubits, such that: for any constant $k$ and any $i_1,...,i_k \in \{1,...,2^n\}$, the values of the bits $a_{i_1},...,a_{i_k}$ can be retrieved from $|Ψ>$ by a one-round Arthur-Merlin interactive protocol of size polynomial in $n$. This shows how to go around Holevo-Nayak's Theorem, using Arthur-Merlin proofs. We use the new representation to prove the following results: 1) Interactive proofs with quantum advice: We show that the class $QIP/qpoly$ contains ALL languages. That is, for any language $L$ (even non-recursive), the membership $x \in L$ (for $x$ of length $n$) can be proved by a polynomial-size quantum interactive proof, where the verifier is a polynomial-size quantum circuit with working space initiated with some quantum state $|Ψ_{L,n} >$ (depending only on $L$ and $n$). Moreover, the interactive proof that we give is of only one round, and the messages communicated are classical. 2) PCP with only one query: We show that the membership $x \in SAT$ (for $x$ of length $n$) can be proved by a logarithmic-size quantum state $|Ψ>$, together with a polynomial-size classical proof consisting of blocks of length $polylog(n)$ bits each, such that after measuring the state $|Ψ>$ the verifier only needs to read {\bf one} block of the classical proof. While the first result is a straight forward consequence of the new representation, the second requires an additional machinery of quantum low-degree-test that may be interesting in its own right.

quant-ph↗

arXiv2003

$P \ne NP$, propositional proof complexity, and resolution lower bounds for the weak pigeonhole principle

Recent results established exponential lower bounds for the length of any Resolution proof for the weak pigeonhole principle. More formally, it was proved that any Resolution proof for the weak pigeonhole principle, with $n$ holes and any number of pigeons, is of length $Ω(2^{n^ε})$, (for a constant $ε= 1/3$). One corollary is that certain propositional formulations of the statement $P \ne NP$ do not have short Resolution proofs. After a short introduction to the problem of $P \ne NP$ and to the research area of propositional proof complexity, I will discuss the above mentioned lower bounds for the weak pigeonhole principle and the connections to the hardness of proving $P \ne NP$.

cs.CC↗