Communication complexity

About: Communication complexity is a research topic. Over the lifetime, 3870 publications have been published within this topic receiving 105832 citations.

CrypTFlow2: Practical 2-Party Secure Inference.

Abstract: We present CrypTFlow2, a cryptographic framework for secure inference over realistic Deep Neural Networks (DNNs) using secure 2-party computation. CrypTFlow2 protocols are both correct -- i.e., their outputs are bitwise equivalent to the cleartext execution -- and efficient -- they outperform the state-of-the-art protocols in both latency and scale. At the core of CrypTFlow2, we have new 2PC protocols for secure comparison and division, designed carefully to balance round and communication complexity for secure inference tasks. Using CrypTFlow2, we present the first secure inference over ImageNet-scale DNNs like ResNet50 and DenseNet121. These DNNs are at least an order of magnitude larger than those considered in the prior work of 2-party DNN inference. Even on the benchmarks considered by prior work, CrypTFlow2 requires an order of magnitude less communication and 20x-30x less time than the state-of-the-art.

GRECS: Graph Encryption for Approximate Shortest Distance Queries

Abstract: We propose graph encryption schemes that efficiently support approximate shortest distance queries on large-scale encrypted graphs. Shortest distance queries are one of the most fundamental graph operations and have a wide range of applications. Using such graph encryption schemes, a client can outsource large-scale privacy-sensitive graphs to an untrusted server without losing the ability to query it. Other applications include encrypted graph databases and controlled disclosure systems. We propose GRECS (stands for GRaph EnCryption for approximate Shortest distance queries) which includes three oracle encryption schemes that are provably secure against any semi-honest server. Our first construction makes use of only symmetric-key operations, resulting in a computationally-efficient construction. Our second scheme makes use of somewhat-homomorphic encryption and is less computationally-efficient but achieves optimal communication complexity (i.e. uses a minimal amount of bandwidth). Finally, our third scheme is both computationally-efficient and achieves optimal communication complexity at the cost of a small amount of additional leakage. We implemented and evaluated the efficiency of our constructions experimentally. The experiments demonstrate that our schemes are efficient and can be applied to graphs that scale up to 1.6 million nodes and 11 million edges.

A model of configuration complexity and its application to a change management system

Abstract: The complexity of configuring computing systems is a major impediment to the adoption of new information technology (IT) products and greatly increases the cost of IT services. This paper develops a model of configuration complexity and demonstrates its value for a change management system. The model represents systems as a set of nested containers with configuration controls. From this representation, we derive various metrics that indicate configuration complexity, including execution complexity, parameter complexity, and memory complexity. We apply this model to a J2EE-based enterprise application and its associated middleware stack to assess the complexity of the manual configuration process for this application. We then show how an automated change management system can greatly reduce configuration complexity.

Interactive PCP

Abstract: A central line of research in the area of PCPs is devoted to constructing short PCPs. In this paper, we show that if we allow an additional interactive verification phase, with very low communication complexity, then for some NP languages, one can construct PCPs that are significantly shorter than the known PCPs (without the additional interactive phase) for these languages. We give many cryptographical applications and motivations for our results and for the study of the new model in general. More specifically, we study a new model of proofs: interactive-PCP . Roughly speaking, an interactive-PCP (say, for the membership xi¾? L) is a proof-string that can be verified by reading only one of its bits, with the help of an interactive-proof with very small communication complexity. We show that for membership in some NP languages L, there are interactive-PCPs that are significantly shorter than the known (non-interactive) PCPs for these languages. Our main result is that for any constant depth Boolean formula i¾?(z 1 ,...,z k ) of size n(over the gates i¾? , i¾? , i¾? , ¬), a prover, Alice, can publish a proof-string for the satisfiability of i¾?, where the size of the proof-string is poly(k). Later on, any user who wishes to verify the published proof-string needs to interact with Alice via a short interactive protocol of communication complexity poly(logn), while accessing the proof-string at a single location. Note that the size of the published proof-string is poly(k), rather than poly(n), i.e., the size is polynomial in the size of the witness, rather than polynomial in the size of the instance. This compares to the known (non-interactive) PCPs that are of size polynomial in the size of the instance. By reductions, this result extends to many other central NP languages (e.g., SAT, k-clique, Vertex-Cover, etc.). More generally, we show that the satisfiability of $\bigwedge_{i=1}^n[\Phi_i(z_1,\ldots,z_k) =0]$, where each i¾? i (z 1 ,...,z k ) is an arithmetic formula of size n(say, over $\mathbb{GF}[2]$) that computes a polynomial of degree d, can be proved by a published proof-string of size poly(k,d). Later on, any user who wishes to verify the published proof-string needs to interact with the prover via an interactive protocol of communication complexity poly(d,logn), while accessing the proof-string at a single location. We give many applications and motivations for our results and for the study of the notion of interactive PCP in general. In particular, we have the following applications: Succinct zero knowledge proofs: We show that any interactive PCP, with certain properties, can be converted into a zero-knowledge interactive proof. We use this to construct zero-knowledge proofs of communication complexity polynomial in the size of the witness, rather than polynomial in the size of the instance, for many NP languages. Succinct probabilistically checkable arguments: In a subsequent paper, we study the new notion of probabilistically checkable argument, and show that any interactive PCP, with certain properties, translates into a probabilistically checkable argument [18]. We use this to construct probabilistically checkable arguments of size polynomial in the size of the witness, rather than polynomial in the size of the instance, for many NP languages. Commit-Reveal schemes: We show that Alice can commit to a string wof kbits, by a message of size poly(k), and later on, for any predicate i¾?of size n, whose satisfiability can be proved by an efficient enough interactive PCP with certain properties, Alice can prove the statement i¾?(w) = 1, by a zero-knowledge interactive proof with communication complexity poly(logn). (Surprisingly, the communication complexity may be significantly smaller than kand n).

Private information retrieval using trusted hardware

Abstract: Many theoretical PIR (Private Information Retrieval) constructions have been proposed in the past years. Though information theoretically secure, most of them are impractical to deploy due to the prohibitively high communication and computation complexity. The recent trend in outsourcing databases fuels the research on practical PIR schemes. In this paper, we propose a new PIR system by making use of trusted hardware. Our system is proven to be information theoretically secure. Furthermore, we derive the computation complexity lower bound for hardware-based PIR schemes and show that our construction meets the lower bounds for both the communication and computation costs, respectively.