DOCSLIB.ORG

Advances and Open Problems in Federated Learning

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Advances and Open Problems in Federated Learning Advances and Open Problems in Federated Learning Peter Kairouz7* H. Brendan McMahan7∗ Brendan Avent21 Aurelien´ Bellet9 Mehdi Bennis19 Arjun Nitin Bhagoji13 Kallista Bonawitz7 Zachary Charles7 Graham Cormode23 Rachel Cummings6 Rafael G.L. D’Oliveira14 Hubert Eichner7 Salim El Rouayheb14 David Evans22 Josh Gardner24 Zachary Garrett7 Adria` Gascon´ 7 Badih Ghazi7 Phillip B. Gibbons2 Marco Gruteser7;14 Zaid Harchaoui24 Chaoyang He21 Lie He 4 Zhouyuan Huo 20 Ben Hutchinson7 Justin Hsu25 Martin Jaggi4 Tara Javidi17 Gauri Joshi2 Mikhail Khodak2 Jakub Konecnˇ y´7 Aleksandra Korolova21 Farinaz Koushanfar17 Sanmi Koyejo7;18 Tancrede` Lepoint7 Yang Liu12 Prateek Mittal13 Mehryar Mohri7 Richard Nock1 Ayfer Ozg¨ ur¨ 15 Rasmus Pagh7;10 Hang Qi7 Daniel Ramage7 Ramesh Raskar11 Mariana Raykova7 Dawn Song16 Weikang Song7 Sebastian U. Stich4 Ziteng Sun3 Ananda Theertha Suresh7 Florian Tramer` 15 Praneeth Vepakomma11 Jianyu Wang2 Li Xiong5 Zheng Xu7 Qiang Yang8 Felix X. Yu7 Han Yu12 Sen Zhao7 1Australian National University, 2Carnegie Mellon University, 3Cornell University, 4Ecole´ Polytechnique Fed´ erale´ de Lausanne, 5Emory University, 6Georgia Institute of Technology, 7Google Research, 8Hong Kong University of Science and Technology, 9INRIA, 10IT University of Copenhagen, 11Massachusetts Institute of Technology, 12Nanyang Technological University, 13Princeton University, 14Rutgers University, 15Stanford University, 16University of California Berkeley, 17 University of California San Diego, 18University of Illinois Urbana-Champaign, 19University of Oulu, 20University of Pittsburgh, 21University of Southern California, 22University of Virginia, 23University of Warwick, 24University of Washington, 25University of Wisconsin–Madison arXiv:1912.04977v3 [cs.LG] 9 Mar 2021 Abstract Federated learning (FL) is a machine learning setting where many clients (e.g. mobile devices or whole organizations) collaboratively train a model under the orchestration of a central server (e.g. service provider), while keeping the training data decentralized. FL embodies the principles of focused data collection and minimization, and can mitigate many of the systemic privacy risks and costs resulting from traditional, centralized machine learning and data science approaches. Motivated by the explosive growth in FL research, this paper discusses recent advances and presents an extensive collection of open problems and challenges. *Peter Kairouz and H. Brendan McMahan conceived, coordinated, and edited this work. Correspondence to kairouz@ google.com and [email protected]. 1 Contents 1 Introduction 4 1.1 The Cross-Device Federated Learning Setting . .5 1.1.1 The Lifecycle of a Model in Federated Learning . .7 1.1.2 A Typical Federated Training Process . .8 1.2 Federated Learning Research . .9 1.3 Organization . 10 2 Relaxing the Core FL Assumptions: Applications to Emerging Settings and Scenarios 11 2.1 Fully Decentralized / Peer-to-Peer Distributed Learning . 11 2.1.1 Algorithmic Challenges . 12 2.1.2 Practical Challenges . 14 2.2 Cross-Silo Federated Learning . 14 2.3 Split Learning . 16 2.4 Executive summary . 17 3 Improving Efficiency and Effectiveness 18 3.1 Non-IID Data in Federated Learning . 18 3.1.1 Strategies for Dealing with Non-IID Data . 19 3.2 Optimization Algorithms for Federated Learning . 20 3.2.1 Optimization Algorithms and Convergence Rates for IID Datasets . 21 3.2.2 Optimization Algorithms and Convergence Rates for Non-IID Datasets . 25 3.3 Multi-Task Learning, Personalization, and Meta-Learning . 28 3.3.1 Personalization via Featurization . 28 3.3.2 Multi-Task Learning . 28 3.3.3 Local Fine Tuning and Meta-Learning . 29 3.3.4 When is a Global FL-trained Model Better? . 30 3.4 Adapting ML Workflows for Federated Learning . 30 3.4.1 Hyperparameter Tuning . 31 3.4.2 Neural Architecture Design . 31 3.4.3 Debugging and Interpretability for FL . 32 3.5 Communication and Compression . 32 3.6 Application To More Types of Machine Learning Problems and Models . 34 3.7 Executive summary . 34 4 Preserving the Privacy of User Data 36 4.1 Actors, Threat Models, and Privacy in Depth . 37 4.2 Tools and Technologies . 38 4.2.1 Secure Computations . 40 4.2.2 Privacy-Preserving Disclosures . 44 4.2.3 Verifiability . 46 4.3 Protections Against External Malicious Actors . 48 4.3.1 Auditing the Iterates and Final Model . 49 4.3.2 Training with Central Differential Privacy . 49 4.3.3 Concealing the Iterates . 51 4.3.4 Repeated Analyses over Evolving Data . 52 4.3.5 Preventing Model Theft and Misuse . 52 4.4 Protections Against an Adversarial Server . 53 4.4.1 Challenges: Communication Channels, Sybil Attacks, and Selection . 53 4.4.2 Limitations of Existing Solutions . 54 4.4.3 Training with Distributed Differential Privacy . 55 4.4.4 Preserving Privacy While Training Sub-Models . 58 2 4.5 User Perception . 59 4.5.1 Understanding Privacy Needs for Particular Analysis Tasks . 59 4.5.2 Behavioral Research to Elicit Privacy Preferences . 60 4.6 Executive Summary . 60 5 Defending Against Attacks and Failures 62 5.1 Adversarial Attacks on Model Performance . 62 5.1.1 Goals and Capabilities of an Adversary . 63 5.1.2 Model Update Poisoning . 66 5.1.3 Data Poisoning Attacks . 67 5.1.4 Inference-Time Evasion Attacks . 69 5.1.5 Defensive Capabilities from Privacy Guarantees . 70 5.2 Non-Malicious Failure Modes . 71 5.3 Exploring the Tension between Privacy and Robustness . 73 5.4 Executive Summary . 73 6 Ensuring Fairness and Addressing Sources of Bias 75 6.1 Bias in Training Data . 75 6.2 Fairness Without Access to Sensitive Attributes . 76 6.3 Fairness, Privacy, and Robustness . 77 6.4 Leveraging Federation to Improve Model Diversity . 78 6.5 Federated Fairness: New Opportunities and Challenges . 79 6.6 Executive Summary . 79 7 Addressing System Challenges 81 7.1 Platform Development and Deployment Challenges . 81 7.2 System Induced Bias . 82 7.2.1 Device Availability Profiles . 82 7.2.2 Examples of System Induced Bias . 83 7.2.3 Open Challenges in Quantifying and Mitigating System Induced Bias . 84 7.3 System Parameter Tuning . 85 7.4 On-Device Runtime . 86 7.5 The Cross-Silo Setting . 87 7.6 Executive Summary . 88 8 Concluding Remarks 89 A Software and Datasets for Federated Learning 119 3 1 Introduction Federated learning (FL) is a machine learning setting where many clients (e.g. mobile devices or whole or- ganizations) collaboratively train a model under the orchestration of a central server (e.g. service provider), while keeping the training data decentralized. It embodies the principles of focused collection and data minimization, and can mitigate many of the systemic privacy risks and costs resulting from traditional, cen- tralized machine learning. This area has received significant interest recently, both from research and applied perspectives. This paper describes the defining characteristics and challenges of the federated learning set- ting, highlights important practical constraints and considerations, and then enumerates a range of valuable research directions. The goals of this work are to highlight research problems that are of significant theo- retical and practical interest, and to encourage research on problems that could have significant real-world impact. The term federated learning was introduced in 2016 by McMahan et al. [337]: “We term our approach Federated Learning, since the learning task is solved by a loose federation of participating devices (which we refer to as clients) which are coordinated by a central server.” An unbalanced and non-IID (identically and independently distributed) data partitioning across a massive number of unreliable devices with limited communication bandwidth was introduced as the defining set of challenges. Significant related work predates the introduction of the term federated learning. A longstanding goal pursued by many research communities (including cryptography, databases, and machine learning) is to ana- lyze and learn from data distributed among many owners without exposing that data. Cryptographic methods for computing on encrypted data were developed starting in the early 1980s [396, 492], and Agrawal and Srikant [11] and Vaidya et al. [457] are early examples of work that sought to learn from local data using a centralized server while preserving privacy. Conversely, even since the introduction of the term federated learning, we are aware of no single work that directly addresses the full set of FL challenges. Thus, the term federated learning provides a convenient shorthand for a set of characteristics, constraints, and challenges that often co-occur in applied ML problems on decentralized data where privacy is paramount. This paper originated at the Workshop on Federated Learning and Analytics held June 17–18th, 2019, hosted at Google’s Seattle office. During the course of this.
Load more

Recommended publications
  • ARTIFICIAL INTELLIGENCE and ALGORITHMIC LIABILITY a Technology and Risk Engineering Perspective from Zurich Insurance Group and Microsoft Corp
    White Paper ARTIFICIAL INTELLIGENCE AND ALGORITHMIC LIABILITY A technology and risk engineering perspective from Zurich Insurance Group and Microsoft Corp. July 2021 TABLE OF CONTENTS 1. Executive summary 03 This paper introduces the growing notion of AI algorithmic 2. Introduction 05 risk, explores the drivers and A. What is algorithmic risk and why is it so complex? ‘Because the computer says so!’ 05 B. Microsoft and Zurich: Market-leading Cyber Security and Risk Expertise 06 implications of algorithmic liability, and provides practical 3. Algorithmic risk : Intended or not, AI can foster discrimination 07 guidance as to the successful A. Creating bias through intended negative externalities 07 B. Bias as a result of unintended negative externalities 07 mitigation of such risk to enable the ethical and responsible use of 4. Data and design flaws as key triggers of algorithmic liability 08 AI. A. Model input phase 08 B. Model design and development phase 09 C. Model operation and output phase 10 Authors: D. Potential complications can cloud liability, make remedies difficult 11 Zurich Insurance Group 5. How to determine algorithmic liability? 13 Elisabeth Bechtold A. General legal approaches: Caution in a fast-changing field 13 Rui Manuel Melo Da Silva Ferreira B. Challenges and limitations of existing legal approaches 14 C. AI-specific best practice standards and emerging regulation 15 D. New approaches to tackle algorithmic liability risk? 17 Microsoft Corp. Rachel Azafrani 6. Principles and tools to manage algorithmic liability risk 18 A. Tools and methodologies for responsible AI and data usage 18 Christian Bucher B. Governance and principles for responsible AI and data usage 20 Franziska-Juliette Klebôn C.
    [Show full text]
  • On the Randomness Complexity of Interactive Proofs and Statistical Zero-Knowledge Proofs*
    On the Randomness Complexity of Interactive Proofs and Statistical Zero-Knowledge Proofs* Benny Applebaum† Eyal Golombek* Abstract We study the randomness complexity of interactive proofs and zero-knowledge proofs. In particular, we ask whether it is possible to reduce the randomness complexity, R, of the verifier to be comparable with the number of bits, CV , that the verifier sends during the interaction. We show that such randomness sparsification is possible in several settings. Specifically, unconditional sparsification can be obtained in the non-uniform setting (where the verifier is modelled as a circuit), and in the uniform setting where the parties have access to a (reusable) common-random-string (CRS). We further show that constant-round uniform protocols can be sparsified without a CRS under a plausible worst-case complexity-theoretic assumption that was used previously in the context of derandomization. All the above sparsification results preserve statistical-zero knowledge provided that this property holds against a cheating verifier. We further show that randomness sparsification can be applied to honest-verifier statistical zero-knowledge (HVSZK) proofs at the expense of increasing the communica- tion from the prover by R−F bits, or, in the case of honest-verifier perfect zero-knowledge (HVPZK) by slowing down the simulation by a factor of 2R−F . Here F is a new measure of accessible bit complexity of an HVZK proof system that ranges from 0 to R, where a maximal grade of R is achieved when zero- knowledge holds against a “semi-malicious” verifier that maliciously selects its random tape and then plays honestly.
    [Show full text]
  • Practical Homomorphic Encryption and Cryptanalysis
    Practical Homomorphic Encryption and Cryptanalysis Dissertation zur Erlangung des Doktorgrades der Naturwissenschaften (Dr. rer. nat.) an der Fakult¨atf¨urMathematik der Ruhr-Universit¨atBochum vorgelegt von Dipl. Ing. Matthias Minihold unter der Betreuung von Prof. Dr. Alexander May Bochum April 2019 First reviewer: Prof. Dr. Alexander May Second reviewer: Prof. Dr. Gregor Leander Date of oral examination (Defense): 3rd May 2019 Author's declaration The work presented in this thesis is the result of original research carried out by the candidate, partly in collaboration with others, whilst enrolled in and carried out in accordance with the requirements of the Department of Mathematics at Ruhr-University Bochum as a candidate for the degree of doctor rerum naturalium (Dr. rer. nat.). Except where indicated by reference in the text, the work is the candidates own work and has not been submitted for any other degree or award in any other university or educational establishment. Views expressed in this dissertation are those of the author. Place, Date Signature Chapter 1 Abstract My thesis on Practical Homomorphic Encryption and Cryptanalysis, is dedicated to efficient homomor- phic constructions, underlying primitives, and their practical security vetted by cryptanalytic methods. The wide-spread RSA cryptosystem serves as an early (partially) homomorphic example of a public- key encryption scheme, whose security reduction leads to problems believed to be have lower solution- complexity on average than nowadays fully homomorphic encryption schemes are based on. The reader goes on a journey towards designing a practical fully homomorphic encryption scheme, and one exemplary application of growing importance: privacy-preserving use of machine learning.
    [Show full text]
  • Oacerts: Oblivious Attribute Certificates
    OACerts: Oblivious Attribute Certificates Jiangtao Li and Ninghui Li CERIAS and Department of Computer Science, Purdue University 250 N. University Street, West Lafayette, IN 47907 {jtli,ninghui}@cs.purdue.edu Abstract. We propose Oblivious Attribute Certificates (OACerts), an attribute certificate scheme in which a certificate holder can select which attributes to use and how to use them. In particular, a user can use attribute values stored in an OACert obliviously, i.e., the user obtains a service if and only if the attribute values satisfy the policy of the service provider, yet the service provider learns nothing about these attribute values. This way, the service provider’s access con- trol policy is enforced in an oblivious fashion. To enable the oblivious access control using OACerts, we propose a new cryp- tographic primitive called Oblivious Commitment-Based Envelope (OCBE). In an OCBE scheme, Bob has an attribute value committed to Alice and Alice runs a protocol with Bob to send an envelope (encrypted message) to Bob such that: (1) Bob can open the envelope if and only if his committed attribute value sat- isfies a predicate chosen by Alice, (2) Alice learns nothing about Bob’s attribute value. We develop provably secure and efficient OCBE protocols for the Pedersen commitment scheme and predicates such as =, ≥, ≤, >, <, 6= as well as logical combinations of them. 1 Introduction In Trust Management and certificate-based access control Systems [3, 40, 19, 11, 34, 33], access control decisions are based on attributes of requesters, which are estab- lished by digitally signed certificates. Each certificate associates a public key with the key holder’s identity and/or attributes such as employer, group membership, credit card information, birth-date, citizenship, and so on.
    [Show full text]
  • Explainable Federated Learning for Taxi Travel Time Prediction
    Explainable Federated Learning for Taxi Travel Time Prediction Jelena Fiosina a Institute of Informatics, Clausthal Technical University, Julius-Albert Str. 4, D-38678, Clausthal-Zellerfeld, Germany Keywords: FCD Trajectories, Traffic, Travel Time Prediction, Federated Learning, Explainability. Abstract: Transportation data are geographically scattered across different places, detectors, companies, or organisations and cannot be easily integrated under data privacy and related regulations. The federated learning approach helps process these data in a distributed manner, considering privacy concerns. The federated learning archi- tecture is based mainly on deep learning, which is often more accurate than other machine learning models. However, deep-learning-based models are intransparent unexplainable black-box models, which should be explained for both users and developers. Despite the fact that extensive studies have been carried out on investigation of various model explanation methods, not enough solutions for explaining federated models exist. We propose an explainable horizontal federated learning approach, which enables processing of the dis- tributed data while adhering to their privacy, and investigate how state-of-the-art model explanation methods can explain it. We demonstrate this approach for predicting travel time on real-world floating car data from Brunswick, Germany. The proposed approach is general and can be applied for processing data in a federated manner for other prediction and classification tasks. 1 INTRODUCTION transparent. This feature of AI-driven systems com- plicates the user acceptance and can be troublesome Most real-world data are geographically scattered even for model developers. across different places, companies, or organisations, Therefore, the contemporary AI technologies and, unfortunately, cannot be easily integrated under should be capable of processing these data in a de- data privacy and related regulations.
    [Show full text]
  • Exploiting Shared Representations for Personalized Federated Learning
    Exploiting Shared Representations for Personalized Federated Learning Liam Collins 1 Hamed Hassani 2 Aryan Mokhtari 1 Sanjay Shakkottai 1 Abstract effective models for each client by leveraging the local com- putational power, memory, and data of all clients (McMahan Deep neural networks have shown the ability to et al., 2017). The task of coordinating between the clients extract universal feature representations from data is fulfilled by a central server that combines the models such as images and text that have been useful received from the clients at each round and broadcasts the for a variety of learning tasks. However, the updated information to them. Importantly, the server and fruits of representation learning have yet to be clients are restricted to methods that satisfy communica- fully-realized in federated settings. Although data tion and privacy constraints, preventing them from directly in federated settings is often non-i.i.d. across applying centralized techniques. clients, the success of centralized deep learning suggests that data often shares a global feature However, one of the most important challenges in federated representation, while the statistical heterogene- learning is the issue of data heterogeneity, where the under- ity across clients or tasks is concentrated in the lying data distribution of client tasks could be substantially labels. Based on this intuition, we propose a different from each other. In such settings, if the server novel federated learning framework and algorithm and clients learn a single shared model (e.g., by minimizing for learning a shared data representation across average loss), the resulting model could perform poorly for clients and unique local heads for each client.
    [Show full text]
  • Architectural Patterns for the Design of Federated Learning Systems
    Architectural Patterns for the Design of Federated Learning Systems a,b < a,b a,b b a,b a Sin Kit Lo , , Qinghua Lu , Liming Zhu , Hye-Young Paik , Xiwei Xu and Chen Wang aData61, CSIRO, Australia bUniversity of New South Wales, Australia ARTICLEINFO ABSTRACT Federated learning has received fast-growing interests from academia and industry to tackle the chal- Keywords: lenges of data hungriness and privacy in machine learning. A federated learning system can be viewed as a large-scale distributed system with different components and stakeholders as numerous client de- Federated Learning vices participate in federated learning. Designing a federated learning system requires software system design thinking apart from the machine learning knowledge. Although much effort has been put into Pattern federated learning from the machine learning technique aspects, the software architecture design con- Software Architecture cerns in building federated learning systems have been largely ignored. Therefore, in this paper, we present a collection of architectural patterns to deal with the design challenges of federated learning Machine Learning systems. Architectural patterns present reusable solutions to a commonly occurring problem within a Artificial Intelligence given context during software architecture design. The presented patterns are based on the results of a systematic literature review and include three client management patterns, four model management patterns, three model training patterns, and four model aggregation patterns. The patterns are asso- ciated to the particular state transitions in a federated learning model lifecycle, serving as a guidance for effective use of the patterns in the design of federated learning systems. Central Model 1. Introduction Server aggregation Federated Learning Overview The ever-growing use of big data systems, industrial-scale IoT platforms, and smart devices contribute to the exponen- tial growth in data dimensions [30].
    [Show full text]
  • Information Guide
    INFORMATION GUIDE 7 ALL-PRO 7 NFL MVP LAMAR JACKSON 2018 - 1ST ROUND (32ND PICK) RONNIE STANLEY 2016 - 1ST ROUND (6TH PICK) 2020 BALTIMORE DRAFT PICKS FIRST 28TH SECOND 55TH (VIA ATL.) SECOND 60TH THIRD 92ND THIRD 106TH (COMP) FOURTH 129TH (VIA NE) FOURTH 143RD (COMP) 7 ALL-PRO MARLON HUMPHREY FIFTH 170TH (VIA MIN.) SEVENTH 225TH (VIA NYJ) 2017 - 1ST ROUND (16TH PICK) 2020 RAVENS DRAFT GUIDE “[The Draft] is the lifeblood of this Ozzie Newsome organization, and we take it very Executive Vice President seriously. We try to make it a science, 25th Season w/ Ravens we really do. But in the end, it’s probably more of an art than a science. There’s a lot of nuance involved. It’s Joe Hortiz a big-picture thing. It’s a lot of bits and Director of Player Personnel pieces of information. It’s gut instinct. 23rd Season w/ Ravens It’s experience, which I think is really, really important.” Eric DeCosta George Kokinis Executive VP & General Manager Director of Player Personnel 25th Season w/ Ravens, 2nd as EVP/GM 24th Season w/ Ravens Pat Moriarty Brandon Berning Bobby Vega “Q” Attenoukon Sarah Mallepalle Sr. VP of Football Operations MW/SW Area Scout East Area Scout Player Personnel Assistant Player Personnel Analyst Vincent Newsome David Blackburn Kevin Weidl Patrick McDonough Derrick Yam Sr. Player Personnel Exec. West Area Scout SE/SW Area Scout Player Personnel Assistant Quantitative Analyst Nick Matteo Joey Cleary Corey Frazier Chas Stallard Director of Football Admin. Northeast Area Scout Pro Scout Player Personnel Assistant David McDonald Dwaune Jones Patrick Williams Jenn Werner Dir.
    [Show full text]
  • A Survey on Federated Learning and Its Applications for Accelerating Industrial Internet of Things
    A Survey on Federated Learning and its Applications for Accelerating Industrial Internet of Things Jiehan Zhou Shouhua Zhang Qinghua Lu [email protected] [email protected] [email protected] University of Oulu University of Oulu CSIRO, Data61 13 Garden Street Eveleigh SYDNEY Wenbin Dai Min Chen Xin Liu [email protected] [email protected] [email protected] Shanghai Jiao Tong University Huazhong Univ. of Science China University of and Technology, China Petroleum Huadong Susanna Pirttikangas Yang Shi Weishan Zhang [email protected] [email protected] [email protected] University of Oulu University of Victoria China University of Petroleum Huadong - Qingdao Campus Enrique Herrera-Viedma [email protected] (IoT) is being widely applied in mobile services. There are few Abstract—Federated learning (FL) brings collaborative reports on applying large-scale data and deep learning (DL) to intelligence into industries without centralized training data to implement large-scale enterprise intelligence. One of the accelerate the process of Industry 4.0 on the edge computing level. reasons is lack of machine learning (ML) approaches which can FL solves the dilemma in which enterprises wish to make the use make distributed learning available while not infringing the of data intelligence with security concerns. To accelerate user’s data privacy. Clearly, FL trains a model by enabling the industrial Internet of things with the further leverage of FL, existing achievements on FL are developed from three aspects: 1) individual devices to act as local learners and send local model define terminologies and elaborate a general framework of FL for parameters to a federal server (defined in section 2) instead of accommodating various scenarios; 2) discuss the state-of-the-art training data.
    [Show full text]
  • Homomorphic Encryption Library
    Ludwig-Maximilians-Universit¨at Munchen¨ Prof. Dr. D. Kranzlmuller,¨ Dr. N. gentschen Felde Data Science & Ethics { Microsoft SEAL { Exercise 1: Library for Matrix Operations Microsoft's Simple Encrypted Arithmetic Library (SEAL)1 is a publicly available homomorphic encryption library. It can be found and downloaded at http://sealcrypto.codeplex.com/. Implement a library 13b MS-SEAL.h supporting matrix operations using homomorphic encryption on basis of the MS SEAL. due date: 01.07.2018 (EOB) no. of students: 2 deliverables: 1. Implemenatation (including source code(s)) 2. Documentation (max. 10 pages) 3. Presentation (10 { max. 15 minutes) 13b MS-SEAL.h #i n c l u d e <f l o a t . h> #i n c l u d e <s t d b o o l . h> typedef struct f double ∗ e n t r i e s ; unsigned int width; unsigned int height; g matrix ; /∗ Initialize new matrix: − reserve memory only ∗/ matrix initMatrix(unsigned int width, unsigned int height); /∗ Initialize new matrix: − reserve memory − set any value to 0 ∗/ matrix initMatrixZero(unsigned int width, unsigned int height); /∗ Initialize new matrix: − reserve memory − set any value to random number ∗/ matrix initMatrixRand(unsigned int width, unsigned int height); 1https://www.microsoft.com/en-us/research/publication/simple-encrypted-arithmetic-library-seal-v2-0/# 1 /∗ copy a matrix and return its copy ∗/ matrix copyMatrix(matrix toCopy); /∗ destroy matrix − f r e e memory − set any remaining value to NULL ∗/ void freeMatrix(matrix toDestroy); /∗ return entry at position (xPos, yPos), DBL MAX in case of error
    [Show full text]
  • Simple Doubly-Efficient Interactive Proof Systems for Locally
    Electronic Colloquium on Computational Complexity, Revision 3 of Report No. 18 (2017) Simple doubly-efficient interactive proof systems for locally-characterizable sets Oded Goldreich∗ Guy N. Rothblumy September 8, 2017 Abstract A proof system is called doubly-efficient if the prescribed prover strategy can be implemented in polynomial-time and the verifier’s strategy can be implemented in almost-linear-time. We present direct constructions of doubly-efficient interactive proof systems for problems in P that are believed to have relatively high complexity. Specifically, such constructions are presented for t-CLIQUE and t-SUM. In addition, we present a generic construction of such proof systems for a natural class that contains both problems and is in NC (and also in SC). The proof systems presented by us are significantly simpler than the proof systems presented by Goldwasser, Kalai and Rothblum (JACM, 2015), let alone those presented by Reingold, Roth- blum, and Rothblum (STOC, 2016), and can be implemented using a smaller number of rounds. Contents 1 Introduction 1 1.1 The current work . 1 1.2 Relation to prior work . 3 1.3 Organization and conventions . 4 2 Preliminaries: The sum-check protocol 5 3 The case of t-CLIQUE 5 4 The general result 7 4.1 A natural class: locally-characterizable sets . 7 4.2 Proof of Theorem 1 . 8 4.3 Generalization: round versus computation trade-off . 9 4.4 Extension to a wider class . 10 5 The case of t-SUM 13 References 15 Appendix: An MA proof system for locally-chracterizable sets 18 ∗Department of Computer Science, Weizmann Institute of Science, Rehovot, Israel.
    [Show full text]
  • Session Types and Game Semantics a Synchronous Side and an Asynchronous Side
    Two Sides of the Same Coin: Session Types and Game Semantics A Synchronous Side and an Asynchronous Side SIMON CASTELLAN, Imperial College London, United Kingdom NOBUKO YOSHIDA, Imperial College London, United Kingdom Game semantics and session types are two formalisations of the same concept: message-passing open programs following certain protocols. Game semantics represents protocols as games, and programs as strategies; while session types specify protocols, and well-typed π-calculus processes model programs. Giving faithful models of the π-calculus and giving a precise description of strategies as a programming language are two difficult 27 problems. In this paper, we show how these two problems can be tackled at the same time by building an accurate game semantics model of the session π-calculus. Our main contribution is to fill a semantic gap between the synchrony of the (session) π-calculus and the asynchrony of game semantics, by developing an event-structure based game semantics for synchronous concurrent computation. This model supports the first truly concurrent fully abstract (for barbed congruence) interpretation of the synchronous (session) π-calculus. We further strengthen this correspondence, establishing finite definability of asynchronous strategies by the internal session π-calculus. As an application of these results, we propose a faithful encoding of synchronous strategies into asynchronous strategies by call-return protocols, which induces automatically an encoding at the level of processes. Our results bring session types and game semantics into the same picture, proposing the session calculus as a programming language for strategies, and strategies as a very accurate model of the session calculus.
    [Show full text]