Moritz Hardt
|
I'm a research staff member in the theory group at IBM Research Almaden. Previously, I was a postdoc in the same group. I completed my PhD in Computer Science at Princeton University in 2011. My advisor was Boaz Barak. Before coming to Princeton, I spent a year as a research scholar at Carnegie Mellon University and three years at Saarland University from where I obtained a B.Sc. and an M.Sc. in Computer Science in 2007. Research Interests: My research interests revolve around algorithms, machine learning and complexity theory. Much of my work addresses the problem of analyzing data that deals with human beings. The algorithmic challenges that arise include ensuring privacy, fairness, and, robustness. |
Contact
Address:
IBM Almaden Research |
News:
- I'm excited to have joined the theory group at IBM Almaden as a research staff member.
- From August 22 to December 20, 2013, I'll be a research fellow at the Simons Institute for the Theory of Computing at Berkeley participating in the program "Theoretical Foundations of Big Data Analysis".
Program comittees: FOCS 2013, STOC 2013, ICALP 2012
Publications (selected)
See selected publications.
-
STOC 2013 (to appear)
Links: Full version (PDF, arxiv) The arxiv paper is missing a plot. See PDF instead.
Abstract: Linear sketches are powerful algorithmic tools that turn an n-dimensional input into a concise lower-dimensional representation via a linear transformation. Such sketches have seen a wide range of applications including norm estimation over data streams, compressed sensing, and distributed computing. In almost any realistic setting, however, a linear sketch faces the possibility that its inputs are correlated with previous evaluations of the sketch. Known techniques no longer guarantee the correctness of the output in the presence of such correlations. We therefore ask: Are linear sketches inherently non-robust to adaptively chosen inputs? We give a strong affirmative answer to this question. Specifically, we show that no linear sketch approximates the Euclidean norm of its input to within an arbitrary multiplicative approximation factor on a polynomial number of adaptively chosen inputs. The result remains true even if the dimension of the sketch is d = n – o(n) and the sketch is given unbounded computation time. Our result is based on an algorithm with running time polynomial in d that adaptively finds a distribution over inputs on which the sketch is incorrect with constant probability. Our result implies several corollaries for related problems including lp-norm estimation and compressed sensing. Notably, we resolve an open problem in compressed sensing regarding the feasibility of l2/l2-recovery guarantees in the presence of computationally bounded adversaries.
-
STOC 2013 (to appear)
Links: Full version (arxiv)
Abstract: We consider differentially private approximate singular vector computation. Known worst-case lower bounds show that the error of any differentially private algorithm must scale polynomially with the dimension of the singular vector. We are able to replace this dependence on the dimension by a natural parameter known as the coherence of the matrix that is often observed to be significantly smaller than the dimension both theoretically and empirically. We also prove a matching lower bound showing that our guarantee is nearly optimal for every setting of the coherence parameter. Notably, we achieve our bounds by giving a robust analysis of the well-known power iteration algorithm, which may be of independent interest. Our algorithm also leads to improvements in worst-case settings and to better low-rank approximations in the spectral norm.
-
COLT 2013 (to appear)
Links: Full version (arxiv)
Abstract: We consider a fundamental problem in unsupervised learning: given a collection of m points in R^n, if many but not necessarily all of these points are contained in a d-dimensional subspace T can we find it? The points contained in T are called inliers and the remaining points are outliers. This problem has received considerable attention in computer science and in statistics. Yet efficient algorithms from computer science are not robust to adversarial outliers, and the estimators from robust statistics are hard to compute in high dimensions. This is a serious and persistent issue not just in this application, but for many other problems in unsupervised learning. Are there algorithms for linear regression that are both robust to outliers and efficient? We give an algorithm that finds T when it contains more than a d/n fraction of the points. Hence, for say d = n/2 this estimator is both easy to compute and well-behaved when there are a constant fraction of outliers. We prove that it is small set expansion hard to find T when the fraction of errors is any larger and so our estimator is an optimal compromise between efficiency and robustness. In fact, this basic problem has a surprising number of connections to other areas including small set expansion, matroid theory and functional analysis that we make use of here.
-
ISBI 2013 (to appear)
Abstract: Data is only as good as the similarity metric used to compare it. The all important notion of similarity allows us to leverage knowledge derived from prior observations to predict characteristics of new samples. In this paper we consider the problem of compiling a consistent and accurate view of similarity given its multiple incomplete and noisy approximations. We propose a new technique called Multiple Kernel Completion (MKC), which completes given similarity kernels as well as finds their best combination within a Support Vector Machine framework, so as to maximize the discrimination margin. We demonstrate the effectiveness of the proposed technique on datasets from UCI Machine Learning repository as well as for the task of heart valve disease discrimination using CW Doppler images. Our empirical results establish that MKC consistently outperforms existing data completion methods like 0-imputation, mean-imputation and matrix completion across datasets and training set sizes.
-
NIPS 2012: 2348–2356
Links: Full version (PDF, arxiv)
Abstract: We present a new algorithm for differentially private data release, based on a simple combination of the Exponential Mechanism with the Multiplicative Weights update rule. Our MWEM algorithm achieves what are the best known and nearly optimal theoretical guarantees, while at the same time being simple to implement and experimentally more accurate on actual data sets than existing techniques.
-
STOC 2012: 1255–1268
Links: Full version (arXiv), Proceedings (ACM)
Abstract: Computing accurate low rank approximations of large matrices is a fundamental data mining task. In many applications however the matrix contains sensitive information about individuals. In such case we would like to release a low rank approximation that satisfies a strong privacy guarantee such as differential privacy. Unfortunately, to date the best known algorithm for this task that satisfies differential privacy is based on naive input perturbation or randomized response: Each entry of the matrix is perturbed independently by a sufficiently large random noise variable, a low rank approximation is then computed on the resulting matrix.
We give (the first) significant improvements in accuracy over randomized response under the natural and necessary assumption that the matrix has low coherence. Our algorithm is also very efficient and finds a constant rank approximation of an m x n matrix in time O(mn). Note that even generating the noise matrix required for randomized response already requires time O(mn).
-
SODA 2012: 168–187
Links: Full version (arXiv), Proceedings (SIAM), Short talk (PDF), Blog post at windowsontheory.org
Abstract: This work considers computationally efficient privacy-preserving data release. We study the task of analyzing a database containing sensitive information about individual participants. Given a set of statistical queries on the data, we want to release approximate answers to the queries while also guaranteeing differential privacy–protecting each participant's sensitive data.
Our focus is on computationally efficient data release algorithms; we seek algorithms whose running time is polynomial, or at least sub-exponential, in the data dimensionality. Our primary contribution is a computationally efficient reduction from differentially private data release for a class of counting queries, to learning thresholded sums of predicates from a related class.
We instantiate this general reduction with a variety of algorithms for learning thresholds. These instantiations yield several new results for differentially private data release. As two examples, taking {0,1}^d to be the data domain (of dimension d), we obtain differentially private algorithms for:
- Releasing all k-way conjunction counting queries (or k-way contingency tables). For any given k, the resulting data release algorithm has bounded error as long as the database is of size at least d^{O(\sqrt{k\log(k\log d)})} (ignoring the dependence on other parameters). The running time is polynomial in the database size. The best sub-exponential time algorithms known prior to our work required a database of size \tilde{O}(d^{k/2}) [Dwork McSherry Nissim and Smith 2006].
- Releasing a (1-\gamma)-fraction of all 2^d parity counting queries. For any \gamma \geq \poly(1/d), the algorithm has bounded error as long as the database is of size at least \poly(d) (again ignoring the dependence on other parameters). The running time is polynomial in the database size.
Several other instantiations yield further results for privacy-preserving data release. Of the two results highlighted above, the first learning algorithm uses techniques for representing thresholded sums of predicates as low-degree polynomial threshold functions. The second learning algorithm is based on Jackson's Harmonic Sieve algorithm [Jackson 1997]. It utilizes Fourier analysis of the database viewed as a function mapping queries to answers.
-
ITCS 2012: 214–226
Links: Full vesion (arxiv), Proceedings (ACM)
Abstract: We study fairness in classification, where individuals are classified, e.g., admitted to a university, and the goal is to prevent discrimination against individuals based on their membership in some group, while maintaining utility for the classifier (the university). The main conceptual contribution of this paper is a framework for fair classification comprising (1) a (hypothetical) task-specific metric for determining the degree to which individuals are similar with respect to the classification task at hand; (2) an algorithm for maximizing utility subject to the fairness constraint, that similar individuals are treated similarly. We also present an adaptation of our approach to achieve the complementary goal of "fair affirmative action," which guarantees statistical parity (i.e., the demographics of the set of individuals receiving any classification are the same as the demographics of the underlying population), while treating similar individuals as similarly as possible. Finally, we discuss the relationship of fairness to privacy: when fairness implies privacy, and how tools developed in the context of differential privacy may be applied to fairness.
-
ITCS 2012: 380-392
Links: Full version (PDF), Proceedings (ACM)
Abstract: We initiate a principled study of graph densification. Given a graph G the goal of graph densification is to come up with another graph H that has significantly more edges than G but nevertheless approximates G well with respect to some set of test functions. In this paper we focus on the case of cut and spectral approximations. As it turns out graph densification exhibits rich connections to a set of interesting and sometimes seemingly unrelated questions in graph theory and metric embeddings. In particular we show the following results:
- A graph G has a multiplicative cut approximation with an asymptotically increased density if and only if it does not embed into ell_1 under a weak notion of embeddability. We demonstrate that all planar graphs as well as random geometric graphs possess such embeddings and thus do not have densifiers. On the other hand, expanders do have densifiers (namely, the complete graph) and as a result do not embed into ell_1 even under our weak notion of embedding.
- An analogous characterization is true for multiplicative spectral approximations where the embedding is into ell_2^2. Using this characterization we expose a surprisingly close connection between multiplicative spectral and multiplicative cut densifiers.
- We also consider additive cut and spectral approximations. We exhibit graphs that do not possess non-trivial additive densifiers.
Our results are mainly based on linear and semidefinite programs (and their duals) for computing the maximum weight densifier of a given graph. This also leads to efficient algorithms in the case of spectral densifiers and additive cut densifiers.
-
STOC 2011: 803–812; SICOMP (to appear)
Links: Full version (arXiv)
Abstract: Suppose we would like to know all answers to a set of statistical queries C on a data set up to small error, but we can only access the data itself using statistical queries. A trivial solution is to exhaustively ask all queries in C. Can we do any better?
• We show that the number of statistical queries necessary and sufficient for this task is---up to polynomial factors---equal to the agnostic learning complexity of C in Kearns' statistical query (SQ) model. This gives a complete answer to the question when running time is not a concern.
• We then show that the problem can be solved efficiently (allowing arbitrary error on a small fraction of queries) whenever the answers to C can be described by a submodular function. This includes many natural concept classes, such as graph cuts and Boolean disjunctions and conjunctions. In doing so we also give a new learning algorithm for submodular functions that improves upon recent results in a different context.
While interesting from a learning theoretic point of view, our main applications are in privacy-preserving data analysis:
Here, our second result leads to the first algorithm that efficiently releases differentially private answers to all Boolean conjunctions with 1% average error. This presents significant progress on a key open problem in privacy-preserving data analysis.
Our first result on the other hand gives unconditional lower bounds on any differentially private algorithm that admits a (potentially non-privacy-preserving) implementation using only statistical queries. Not only our algorithms, but also most known private algorithms can be implemented using only statistical queries, and hence are constrained by these lower bounds. Our result therefore isolates the complexity of agnostic learning in the SQ-model as a new barrier in the design of differentially private algorithms.
-
SODA 2011: 512–531
Links: Full version (ECCC), Talk (PDF)
Short abstract: We consider the following seemingly unrelated questions:
• Is the MaxCut problem hard on random geometric graphs of the type considered by Feige and Schechtman (2002)?
• Is the value of a mathematical relaxation for a constraint-satisfaction problem (CSP) preserved when one passes from an instance P to a random induced sub-formula of P?
It turns out that the answer to the first question is ``no'' and in fact this is intimately related to the second question. The answer to the second question is much more subtle, and, in contrast to the case of the objective value of the CSP, the answer strongly depends on the type of relaxation and CSP.
-
FOCS 2010: 61–70
Links: Full version (PDF), Talk (PDF, Video)
Short abstract: We consider privacy-preserving statistical data analysis with online queries. In this setting a trusted curator maintains a database of sensitive information about individual participants, and releases privacy-preserving answers to queries as they arrive online. Our primary contribution is a new differentially private multiplicative weights mechanism for answering a large number of counting queries that arrive online and may be adaptively chosen. Our mechanism is the first to achieve worst-case accuracy guarantees (accuracy on every input database) for a large number of online queries together with a runtime that is polynomial in the data universe. The error is optimal in its dependence on the size of the database and depends only logarithmically on the number of queries being answered. The runtime is nearly linear in the size of the data universe.
Our main technical contributions are a new application of multiplicative weights techniques to the differential privacy setting, and a new privacy analysis for multiplicative weights algorithms.
-
STOC 2010: 705–714
Links: Full version (PDF, arXiv), Proceedings (DOI)
Short abstract: We give a connection between convex geometry and differentially private data analysis. Specifically, we study the noise complexity of differentially private mechanisms in the setting where the data is represented by a histogram and the user asks a number of linear queries non-adaptively. We show that the amount of noise necessary and sufficient to achieve differential privacy is determined by two geometric parameters of a convex body associated with the set of queries. We use this connection to give tight upper and lower bounds for random linear queries. Assuming the truth of a deep conjecture from convex geometry, known as the Hyperplane conjecture, we can extend our results to arbitrary linear queries giving nearly matching upper and lower bounds.
-
SODA 2009: 1193–1200
-
FOCS 2008: 374–383
Short abstract: We show an essentially tight connection between the semidefinite programming relaxation of Unique Games and their behavior under parallel repetition. Our results generalize and help to explain Raz's refutation of the Strong Parallel Repetition Conjecture. Indeed, we show that a Unique Game is a counterexample to strong parallel repetition whenever its SDP value is significantly larger than its integral value.
-
ICALP (1) 2008: 345–356
-
In Inf. Process. Lett. 109(3): 187–192
See my DBLP and Google Scholar entry for more information.
Non-technical writings
-
Governing Algorithms, New York University, 2013
Links: PDF
Abstract: This short note is a response to Frank Pasquale's article "The Emperor's New Codes: Reputation and Search Algorithms in the Finance Sector" published as part of the conference Governing Algorithms, New York University, 2013. Pasquale describes how algorithms in the finance industry have failed to deliver on the promise of greater transparency, accountability, fairness, and efficiency. We discuss some of the steps that Computer Scientists have undertaken towards addressing these concerns. We focus on three problems: achieving fairness, designing algorithms that are robust to manipulation, and understanding the inherent complexity in financial products.
Thesis
-
Dissertation. Princeton University, 2011
Links: PDF
Abstract: In this thesis we consider the challenges arising in the design of algorithms that interact with sensitive personal data---such as medical records, online tracking data, or financial records.
One important goal is to protect the privacy of those individuals whose personal information contributed to the data set. We consider algorithms that satisfy the strong privacy guarantee known as differential privacy. A wide range of computational tasks reduces to the setting in which a trusted database curator responds to a number of statistical queries posed by an untrusted data analyst. The basic question is how accurately and efficiently the curator can release approximate answers to the given queries while satisfying differential privacy. We make the following main contributions to differentially private data analysis:
- We expose a connection between differential privacy and certain problems in convex geometry revolving around a deep conjecture known as the Hyperplane conjecture. Assuming the truth of this conjecture we give differentially private mechanisms with nearly optimal accuracy in the case where the queries are given all at once (non-interactively) and the number of queries does not exceed the database size.
- Multiplicative weights mechanisms are a powerful tool in algorithms, machine learning and optimization. We introduce a privacy-preserving multiplicative weights framework suitable for answering a huge number of queries even in the interactive setting. The accuracy of our algorithm in terms of database size and number of queries matches the statistical sampling error that already arises in the absence of any privacy concerns. Our algorithm can also be used to produce a differentially private synthetic data set encoding the curator's answers. For this task the runtime of our algorithm---which is linear in the universe size---is essentially optimal due to a prior cryptographic hardness result.
- We then consider avenues for obtaining differentially private algorithms with a runtime polynomial in the size of the data set or at least subexponential in the universe size. Based on a new learning algorithm for submodular functions, we present the first polynomial-time algorithm for answering a large number of Boolean conjunction queries (or contingency tables) with non-trivial accuracy guarantees. Conjunction queries are a widely used and important class of statistical queries.
- Furthermore, we exhibit an explicit and efficient reduction from the problem of privately releasing a class of queries to the problem of non-privately learning a related class of concepts. Instantiating this general reduction with new and existing learning algorithms yields several new results for privately releasing conjunctions and other queries.
Not all problems arising in the presence of sensitive data are a matter of privacy. In the final part of this thesis, we isolate fairness in classification as a formidable concern and thus initiate its formal study. The goal of fairness is to prevent discrimination against protected subgroups of the population in a classification system. We argue that fairness cannot be achieved by blindness to the attribute we would like to protect. Our main conceptual contribution is in asserting that fairness is achieved when similar individuals are treated similarly. Based on the goal of treating similar individuals similarly, we formalize and show how to achieve fairness in classification, given a similarity metric. We also observe that our notion of fairness can be seen as a generalization of differential privacy.
Some of my talks
Converted to PDF for compatability, some step effects missing. Powerpoint slides available on request.
- Differential Privacy and Spectral Analysis (PDF)
Invited Tutorial at Simons Workshop on "The Science Of Differential Privacy" - What is the complexity of ensuring differential privacy? (PDF)
Invited Plenary Talk at Oberwolfach Complexity Theory Workshop 2012 - Privacy by Learning the Database (PDF,
New:
YouTube)
Invited Tutorial at DIMACS Workshop on Differential Privacy - Beating Randomized Response on Incoherent Matrices (PDF)
STOC 2012 - A Tutorial on Differential Privacy in Health Care Analytics
and Medical Research (PDF)
IBM Almaden, Health Care Group Meeting, 2012 -
Private Data Release via Learning Thresholds
(PDF)
SODA 2012 - Graph Densification
(PDF)
ITCS 2012 - Fairness through Awareness (PDF)
ITCS 2012 - Multiplicative Weights in Privacy-Preserving Data
Analysis
(PDF)
Columbia University, 2011 - Subsampling Mathematical Relaxations and Average-case
Complexity (PDF)
SODA 2011 -
A Multiplicative Weights Approach to Privacy-Preserving Data
Analyis (video)
FOCS 2010 - An Algorithmic Proof of Forster's Lower Bound
(video)
Institute for Advanced Study
Miscellaneous
-
