Uri Stemmer
I am interested in privacy-preserving data analysis, computational learning theory, and algorithms. Typically my research is theoretical. I am also a faculty member at the school of computer science of Tel Aviv University.
Authored Publications
Sort By
Preview abstract
One of the most basic problems for studying the "price of privacy over time" is the so called private counter problem, introduced by Dwork et al. (2010) and Chan et al. (2011). In this problem, we aim to track the number of events that occur over time, while hiding the existence of every single event. More specifically, in every time step $t\in[T]$ we learn (in an online fashion) that $\Delta_t\geq 0$ new events have occurred, and must respond with an estimate $n_t\approx\sum_{j=1}^t \Delta_j$. The privacy requirement is that all of the outputs together, across all time steps, satisfy event level differential privacy.
The main question here is how our error needs to depend on the total number of time steps $T$ and the total number of events $n$. Dwork et al. (2015) showed an upper bound of $O\left(\log(T)+\log^2(n)\right)$, and Henzinger et al. (2023) showed a lower bound of $\Omega\left(\min\{\log n, \log T\}\right)$. We show a new lower bound of $\Omega\left(\min\{n,\log T\}\right)$, which is tight w.r.t. the dependence on $T$, and is tight in the sparse case where $\log^2 n=O(\log T)$. Our lower bound has the following implications:
* We show that our lower bound extends to the online thresholds problem, where the goal is to privately answer many "quantile queries" when these queries are presented one-by-one. This resolves an open question of Bun et al. (2017).
* Our lower bound implies, for the first time, a separation between the number of mistakes obtainable by a private online learner and a non-private online learner. This partially resolves a COLT'22 open question published by Sanyal and Ramponi.
* Our lower bound also yields the first separation between the standard model of private online learning and a recently proposed relaxed variant of it, called private online prediction.
View details
Preview abstract
In the current digital world, large organizations (sometimes referred to as tech giants) provide service to extremely large numbers of users. The service provider is often interested in computing various data analyses over the private data of its users, which in turn have their incentives to cooperate, but do not necessarily trust the service provider.
In this work, we introduce the Gulliver multi-party computation model (GMPC) to realistically capture the above scenario. The GMPC model considers a single highly powerful party, called the server or Gulliver, that is connected to n users over a star topology network (alternatively formulated as a full network, where the server can block any message). The users are significantly less powerful than the server, and, in particular, should have both computation and communication complexities that are polylogarithmic in n. Protocols in the GMPC model should be secure against malicious adversaries that may corrupt a subset of the users and/or the server.
Designing protocols in the GMPC model is a delicate task, since users can only hold information about polylog(n) other users (and, in particular, can only communicate with polylog(n) other users). In addition, the server can block any message between any pair of honest parties. Thus, reaching an agreement becomes a challenging task. Nevertheless, we design generic protocols in the GMPC model, assuming that <1/8 fraction of the users may be corrupted (in addition to the server). Our main contribution is a variant of Feige's committee election protocol [FOCS 1999] that is secure in the GMPC model. Given this tool we show:
* Assuming fully homomorphic encryption (FHE), any computationally efficient function with O(n polylog(n))-size output can be securely computed in the GMPC model.
* Any function that can be computed by a circuit of O(polylog(n)) depth, O(n polylog(n))$ size, and bounded fan-in and fan-out can be securely computed in the GMPC model assuming vector commitment schemes (without assuming FHE).
* In particular, sorting can be securely computed in the GMPC model assuming vector commitment schemes. This has important applications for the shuffle model of differential privacy, and resolves an open question of Bell et al. [CCS 2020].
View details
Preview abstract
Private Everlasting Prediction (PEP), recently introduced by Naor et al. [2023], is a model for differentially private learning in which the learner never publicly releases a hypothesis. Instead, it provides a black-box access to a ``prediction oracle'' that can predict the labels of an endless stream of unlabeled examples drawn from the underlying distribution. Importantly, PEP provides privacy both for the initial training set and for the endless stream of classification queries. We present two conceptual modifications to the definition of PEP, as well as new constructions exhibiting significant improvements over prior work. Specifically, our contributions include:
(1) Robustness: PEP only guarantees accuracy provided that all the classification queries are drawn from the correct underlying distribution. A few out-of-distribution queries might break the validity of the prediction oracle for future queries, even for future queries which are sampled from the correct distribution. We incorporate robustness against such poisoning attacks into the definition of PEP, and show how to obtain it.
(2) Dependence of the privacy parameter delta in the time horizon: We present a relaxed privacy definition, suitable for PEP, that allows us to disconnect the privacy parameter delta from the number of total time steps T. This allows us to obtain algorithms for PEP whose sample complexity is independent from T, thereby making them "truly everlasting". This is in contrast to prior work where the sample complexity grows with polylog(T).
(3) New constructions: Prior constructions for PEP exhibit sample complexity that is quadratic in the VC dimension of the target class. We present new constructions of PEP for axis-aligned rectangles and for decision-stumps, that exhibit sample complexity linear in the dimension (instead of quadratic). We show that our constructions satisfy very strong robustness properties.
View details
Preview abstract
The problem of learning threshold functions is a fundamental one in machine learning. Classical learning theory implies sample complexity of $O(\xi^{-1} \log(1/\beta))$ (for generalization error $\xi$ with confidence $1-\beta$). The private version of the problem, however, is more challenging and in particular, the sample complexity must depend on the size $|X|$ of the domain. Progress on quantifying this dependence, via lower and upper bounds, was made in a line of works over the past decade. In this paper, we finally close the gap for approximate-DP and provide a nearly tight upper bound of $\widetilde{O}(\log^* |X|)$, which matches a lower bound by Alon et al (that applies even with improper learning) and improves over a prior upper bound of $\widetilde{O}((\log^* |X|)^{1.5})$ by Kaplan et al. We also provide matching upper and lower bounds of $\tilde{\Theta}(2^{\log^*|X|})$ for the additive error of private quasi-concave optimization (a related and more general problem). Our improvement is achieved via the novel Reorder-Slice-Compute paradigm for private data analysis which we believe will have further applications.
View details
Preview abstract
We introduce the concurrent shuffle model of differential privacy. In this model we have multiple concurrent shufflers permuting messages from different, possibly overlapping, batches of users. Similarly to the standard (single) shuffle model, the privacy requirement is that the concatenation of all shuffled messages should be differentially private. We study the private continual summation problem (a.k.a. the counter problem) and show that
the concurrent shuffle model allows for significantly improved error compared to a standard (single) shuffle model. Specifically, we give a summation algorithm with error $\Tilde{O}(n^{1/(2k+1)})$ with $k$ concurrent shufflers on a sequence of length $n$. Furthermore, we prove that this bound is tight for any $k$, even if the algorithm can choose the sizes of the batches adaptively. For $k=\log n$ shufflers, the resulting error is polylogarithmic, much better than $\Tilde{\Theta}(n^{1/3})$ which we show is the smallest possible with a single shuffler.
We use our online summation algorithm to get algorithms with improved regret bounds for the contextual linear bandit problem. In particular we get optimal $\Tilde{O}(\sqrt{n})$ regret with $k= \Tilde{\Omega}(\log n)$ concurrent shufflers.
View details
Preview abstract
A private learner is trained on a sample of labeled points and generates a hypothesis that can be used for predicting the labels of newly sampled points while protecting the privacy of the training set [Kasiviswannathan et al., FOCS 2008]. Research uncovered that private learners may need to exhibit significantly higher sample complexity than non-private learners as is the case with, e.g., learning of one-dimensional threshold functions [Bun et al., FOCS 2015, Alon et al., STOC 2019].
We explore prediction as an alternative to learning. Instead of putting forward a hypothesis, a predictor answers a stream of classification queries. Earlier work has considered a private prediction model with just a single classification query [Dwork and Feldman, COLT 2018]. We observe that when answering a stream of queries, a predictor must modify the hypothesis it uses over time, and, furthermore, that it must use the queries for this modification, hence introducing potential privacy risks with respect to the queries themselves.
We introduce private everlasting prediction taking into account the privacy of both the training set and the (adaptively chosen) queries made to the predictor. We then present a generic construction of private everlasting predictors in the PAC model. The sample complexity of the initial training sample in our construction is quadratic (up to polylog factors) in the VC dimension of the concept class. Our construction allows prediction for all concept classes with finite VC dimension, and in particular threshold functions with constant size initial training sample, even when considered over infinite domains, whereas it is known that the sample complexity of privately learning threshold functions must grow as a function of the domain size and hence is impossible for infinite domains.
View details
Preview abstract
CountSketch and Feature Hashing (the "hashing trick") are popular randomized dimensionality reduction methods that support recovery of $\ell_2$-heavy hitters (keys $i$ where $v_i^2 > \epsilon \|\boldsymbol{v}\|_2^2$) and approximate inner products. When the inputs are {\em not adaptive} (do not depend on prior outputs), classic estimators applied to a sketch of size $O(\ell/\epsilon)$ are accurate for a number of queries that is exponential in $\ell$. When inputs are adaptive, however, an adversarial input can be constructed after $O(\ell)$ queries with the classic estimator and the best known robust estimator only supports $\tilde{O}(\ell^2)$ queries. In this work we show that this quadratic dependence is in a sense inherent: We design an attack that after $O(\ell^2)$ queries produces an adversarial input vector whose sketch is highly biased. Our attack uses "natural" non-adaptive inputs (only the final adversarial input is chosen adaptively) and universally applies with any correct estimator, including one that is unknown to the attacker. In that, we expose inherent vulnerability of this fundamental method.
View details
Preview abstract
We study the space complexity of the two related fields of differential privacy and adaptive data analysis. Specifically,
(1) Under standard cryptographic assumptions, we show that there exists a problem P that requires exponentially more space to be solved efficiently with differential privacy, compared to the space needed without privacy. To the best of our knowledge, this is the first separation between the space complexity of private and non-private algorithms.
(2) The line of work on adaptive data analysis focuses on understanding the number of samples needed for answering a sequence of adaptive queries. We revisit previous lower bounds at a foundational level, and show that they are a consequence of a space bottleneck rather than a sampling bottleneck.
To obtain our results, we define and construct an encryption scheme with multiple keys that is built to withstand a limited amount of key leakage in a very particular way.
View details
Preview abstract
In this work we revisit an interactive variant of joint differential privacy, recently introduced by Naor et al. [2023], and generalize it towards handling online processes in which existing privacy definitions seem too restrictive. We study basic properties of this definition and demonstrate that it satisfies (suitable variants) of group privacy, composition, and post processing.
In order to demonstrate the advantages of this privacy definition compared to traditional forms of differential privacy, we consider the basic setting of online classification. We show that any (possibly non-private) learning rule can be effectively transformed to a private learning rule with only a polynomial overhead in the mistake bound. This demonstrates a stark difference with traditional forms of differential privacy, such as the one studied by Golowich and Livni [2021], where only a double exponential overhead in the mistake bound is known (via an information theoretic upper bound).
View details
Preview abstract
In adaptive data analysis, a mechanism gets n i.i.d. samples from an unknown distribution D, and is required to provide accurate estimations to a sequence of adaptively chosen statistical queries with respect to D. Hardt and Ullman [2014] and Steinke and Ullman [2015] showed that, in general, it is computationally hard to answer more than n^2 adaptive queries, assuming the existence of one-way functions.
However, these negative results strongly rely on an adversarial model that significantly advantages the adversarial analyst over the mechanism, as the analyst, who chooses the adaptive queries, also chooses the underlying distribution D. This imbalance raises questions with respect to the applicability of the obtained hardness results -- an analyst who has complete knowledge of the underlying distribution D would have little need, if at all, to issue statistical queries to a mechanism which only holds a finite number of samples from D.
We consider more restricted adversaries, called balanced, where each such adversary consists of two separated algorithms: The sampler who is the entity that chooses the distribution and provides the samples to the mechanism, and the analyst who chooses the adaptive queries, but does not have a prior knowledge of the underlying distribution. We improve the quality of previous lower bounds by revisiting them using an efficient balanced adversary, under standard public-key cryptography assumptions. We show that these stronger hardness assumptions are unavoidable in the sense that any computationally bounded balanced adversary that has the structure of all known attacks, implies the existence of public-key cryptography.
View details
