A recent line of work initiated by Chiesa and Gur and further developed by Herman and Rothblum investigates the sample and communication complexity of verifying properties of distributions with the assistance of a powerful, knowledgeable, but untrusted prover. In this work, we initiate the study of differentially private (DP) distribution property testing. After all, if we do not trust the prover to help us with verification, why should we trust it with our sensitive sample? We map a landscape of DP prover-aided proofs of properties of distributions. In the non-private case it is known that one-round (two message) private-coin protocols can have substantially lower complexity than public-coin AM protocols, but in the private case, the possibility for improvement depends on the parameter regime and privacy model. Drawing on connections to replicability and techniques for amplification, we show: (1) There exists a reduction from any one-round (ε,δ)-DP private-coin interactive proof to a one-round public-coin DP interactive proof with the same privacy parameters, for the parameter regime ε=O(1/√(n)) and δ=O(1/n^5/2), and with the same sample and communication complexities. (2) If the verifier's message in the private-coin interactive proof is O(1/√(log n)) locally DP – a far more relaxed privacy parameter regime in a different model – then applying one additional transformation again yields a one-round public-coin protocol with the same privacy bound and the same sample and computational complexities. (3) However, when the privacy guarantee is very relaxed (ε(log n)), private coins indeed reduce complexity. We also obtain a Merlin-Arthur (one-message) proof for privately testing whether samples are drawn from a product distribution, and prove that its sample complexity is optimal.

Professional networks are a key determinant of individuals' labor market outcomes. They may also play a role in either exacerbating or ameliorating inequality of opportunity across demographic groups. In a theoretical model of professional network formation, we show that inequality can increase even without exogenous in-group preferences, confirming and complementing existing theoretical literature. Increased inequality emerges from the differential leverage privileged and unprivileged individuals have in forming connections due to their asymmetric ex ante prospects. This is a formalization of a source of inequality in the labor market which has not been previously explored. We next show how inequality-aware platforms may reduce inequality by subsidizing connections, through link recommendations that reduce costs, between privileged and unprivileged individuals. Indeed, mixed-privilege connections turn out to be welfare improving, over all possible equilibria, compared to not recommending links or recommending some smaller fraction of cross-group links. Taken together, these two findings reveal a stark reality: professional networking platforms that fail to foster integration in the link formation process risk reducing the platform's utility to its users and exacerbating existing labor market inequality.

Individuals with similar qualifications and skills may vary in their demeanor, or outward manner: some tend toward self-promotion while others are modest to the point of omitting crucial information. Comparing the self-descriptions of equally qualified job-seekers with different self-presentation styles is therefore problematic. We build an interactive AI for skill elicitation that provides accurate determination of skills while simultaneously allowing individuals to speak in their own voice. Such a system can be deployed, for example, when a new user joins a professional networking platform, or when matching employees to needs during a company reorganization. To obtain sufficient training data, we train an LLM to act as synthetic humans. Elicitation mitigates endogenous bias arising from individuals' own self-reports. To address systematic model bias we enforce a mathematically rigorous notion of equitability ensuring that the covariance between self-presentation manner and skill evaluation error is small.

To meet its dual burdens of providing useful statistics and ensuring privacy of individual respondents, the US Census Bureau has for decades introduced some form of "noise" into published statistics. Initially, they used a method known as "swapping" (1990-2010). In 2020, they switched to an algorithm called TopDown that ensures a form of Differential Privacy. While the TopDown algorithm has been made public, no implementation of swapping has been released and many details of the deployed swapping methodology deployed have been kept secret. Further, the Bureau has not published (even a synthetic) "original" dataset and its swapped version. It is therefore difficult to evaluate the effects of swapping, and to compare these effects to those of other privacy technologies. To address these difficulties we describe and implement a parameterized swapping algorithm based on Census publications, court documents, and informal interviews with Census employees. With this implementation, we characterize the impacts of swapping on a range of statistical quantities of interest. We provide intuition for the types of shifts induced by swapping and compare against those introduced by TopDown. We find that even when swapping and TopDown introduce errors of similar magnitude, the direction in which statistics are biased need not be the same across the two techniques. More broadly, our implementation provides researchers with the tools to analyze and potentially correct for the impacts of disclosure avoidance systems on the quantities they study.

We initiate the study of differentially private learning in the proportional dimensionality regime, in which the number of data samples n and problem dimension d approach infinity at rates proportional to one another, meaning that d / n →δ as n →∞ for an arbitrary, given constant δ∈ (0, ∞ ) . This setting is significantly more challenging than that of all prior theoretical work in high-dimensional differentially private learning, which, despite the name, has assumed that δ = 0 or is sufficiently small for problems of sample complexity O(d), a regime typically considered “low-dimensional” or “classical” by modern standards in high-dimensional statistics. We provide sharp theoretical estimates of the error of several well-studied differentially private algorithms for robust linear regression and logistic regression, including output perturbation, objective perturbation, and noisy stochastic gradient descent, in the proportional dimensionality regime. The 1 + o(1) factor precision of our error estimates enables a far more nuanced understanding of the price of privacy of these algorithms than that afforded by existing, coarser analyses, which are essentially vacuous in the regime we consider. Using our estimates, we discover a previously unobserved “double descent”-like phenomenon in the training error of objective perturbation for robust linear regression. We also identify settings in which output perturbation outperforms objective perturbation on average, and vice versa. To prove our main theorems, we introduce—and strengthen, to handle perturbations required for privacy—several probabilistic tools that have not previously been used to analyze differentially private learning algorithms, such as a modern Gaussian comparison inequality and recent universality laws with origins in statistical physics.

This study investigates the ease with which third parties can access and exploit intimate health data, including mental health- and abortion-related information after the Dobbs decision. Our findings demonstrate incongruities surrounding the perceived protections of HIPAA and the ready availability for the purchase of medical data.

Given a small random sample of n-bit strings labeled by an unknown Boolean function, which properties of this function can be tested computationally efficiently? We show an equivalence between properties that are efficiently testable from few samples and properties with structured symmetry, which depend only on the function's average values on an efficiently computable partition of the domain. Without the efficiency constraint, a similar characterization in terms of unstructured symmetry was obtained by Blais and Yoshida (2019). We also give a function testing analogue of the classic characterization of testable graph properties in terms of regular partitions, as well as a sublinear time and differentially private algorithm to compute concise summaries of such partitions of graphs. Finally, we tighten a recent characterization of the computational indistinguishability of product distributions, which encompasses the related task of efficiently testing which of two candidate functions labeled the observed samples. Essential to our proofs is the following observation of independent interest: Every randomized Boolean function, no matter how complex, admits a supersimulator: a randomized polynomial-size circuit whose output on random inputs cannot be efficiently distinguished from reality with constant advantage, even by polynomially larger distinguishers. This surprising fact is implicit in a theorem of Dwork et al. (2021) in the context of algorithmic fairness, but its complexity-theoretic implications were not previously explored. We give a new proof of this lemma using an iteration technique from the graph regularity literature, and we observe that a subtle quantifier switch allows it to powerfully circumvent known barriers to improving the landmark complexity-theoretic regularity lemma of Trevisan, Tulsiani, and Vadhan (2009).

We study a fundamental question of domain generalization: given a family of domains (i.e., data distributions), how many randomly sampled domains do we need to collect data from in order to learn a model that performs reasonably well on every seen and unseen domain in the family? We model this problem in the PAC framework and introduce a new combinatorial measure, which we call the domain shattering dimension. We show that this dimension characterizes the domain sample complexity. Furthermore, we establish a tight quantitative relationship between the domain shattering dimension and the classic VC dimension, demonstrating that every hypothesis class that is learnable in the standard PAC setting is also learnable in our setting.

Given a differentially private unbiased estimate (q) over tilde = q(D) | nu of a statistic q(D), we wish to obtain unbiased estimates of functions of q(D), such as 1/q(D), solely through post-processing of (q) over tilde, with no further access to the confidential dataset D. To this end, we adapt the deconvolution method used for unbiased estimation in the statistical literature, deriving unbiased estimators for a broad family of twice-differentiable functions - those that are tempered distributions - when the privacy-preserving noise nu is drawn from the Laplace distribution (Dwork et al., 2006). We further extend this technique to functions other than tempered distributions, deriving approximately optimal estimators that are unbiased for values in a user-specified interval (possibly extending to +/-infinity). We use these results to derive an unbiased estimator for private means when the size n of the dataset is not publicly known. In a numerical application, we find that a mechanism that uses our estimator to return an unbiased sample size and mean outperforms a mechanism that instead uses the previously known unbiased privacy mechanism for such means (Kamath et al., 2023). We also apply our estimators to develop unbiased transformation mechanisms for per-record differential privacy, a privacy concept in which the privacy guarantee is a public function of a record's value (Seeman et al., 2024). Our mechanisms provide stronger privacy guarantees than those in prior work (Finley et al., 2024) by using Laplace, rather than Gaussian, noise. Finally, using a different approach, we go beyond Laplace noise by deriving unbiased estimators for polynomials under the weak condition that the noise distribution has sufficiently many moments.

Professional networks provide invaluable entree to opportunity through referrals and introductions. A rich literature shows they also serve to entrench and even exacerbate a status quo of privilege and disadvantage. Hiring platforms, equipped with the ability to nudge link formation, provide a tantalizing opening for beneficial structural change. We anticipate that key to this prospect will be the ability to estimate the likelihood of edge formation in an evolving graph. Outcome-indistinguishable prediction algorithms ensure that the modeled world is indistinguishable from the real world by a family of statistical tests. Omnipredictors ensure that predictions can be post-processed to yield loss minimization competitive with respect to a benchmark class of predictors for many losses simultaneously, with appropriate post-processing. We begin by observing that, by combining a slightly modified form of the online K29* algorithm of Vovk (2007) with basic facts from the theory of reproducing kernel Hilbert spaces, one can derive simple and efficient online algorithms satisfying outcome indistinguishability and omniprediction, with guarantees that improve upon, or are complementary to, those currently known. This is of independent interest; for example, we obtain efficient outcome indistinguishability for some interesting infinite collections of tests, as well as for any bounded function - including those computable by deep (graph) neural networks. We apply these techniques to evolving graphs by designing efficient kernel functions that capture socially meaningful features of nodes and their neighborhoods. We obtain online outcome-indistinguishable omnipredictors for rich - possibly infinite - sets of distinguishers yielding, inter alia, multicalibrated predictions of edge formation with respect to pairs of demographic groups, and the ability to simultaneously optimize loss as measured by a variety of social welfare functions.

We prove that every randomized Boolean function admits a supersimulator: a randomized polynomial-size circuit whose output on random inputs cannot be efficiently distinguished from reality with constant advantage, even by polynomially larger distinguishers. Our result builds on the landmark complexity-theoretic regularity lemma of Trevisan, Tulsiani and Vadhan (2009), which, in contrast, provides a simulator that fools smaller distinguishers. We circumvent lower bounds for the simulator size by letting the distinguisher size bound vary with the target function, while remaining below an absolute upper bound independent of the target function. This dependence on the target function arises naturally from our use of an iteration technique originating in the graph regularity literature. The simulators provided by the regularity lemma and recent refinements thereof, known as multiaccurate and multicalibrated predictors, respectively, as per Hebert-Johnson et al. (2018), have previously been shown to have myriad applications in complexity theory, cryptography, learning theory, and beyond. We first show that a recent multicalibration-based characterization of the computational indistinguishability of product distributions actually requires only (calibrated) multiaccuracy. We then show that supersimulators yield an even tighter result in this application domain, closing a complexity gap present in prior versions of the characterization.

We describe the launching of the Society for Privacy and Confidentiality Research (SPCR). SPCR is the new owner of the Journal of Privacy and Confidentiality, with the goal of ensuring a sustainable future for the Journal, and continuing to publish the multidisciplinary, open access journal that JPC has been since its founding in 2008.

We study the impact of content moderation policies in online communities. In our theoretical model, a platform chooses a content moderation policy and individuals choose whether or not to participate in the community according to the fraction of user content that aligns with their preferences. The effects of content moderation, at first blush, might seem obvious: it restricts speech on a platform. However, when user participation decisions are taken into account, its effects can be more subtle $\unicode{x2013}$ and counter-intuitive. For example, our model can straightforwardly demonstrate how moderation policies may increase participation and diversify content available on the platform. In our analysis, we explore a rich set of interconnected phenomena related to content moderation in online communities. We first characterize the effectiveness of a natural class of moderation policies for creating and sustaining stable communities. Building on this, we explore how resource-limited or ideological platforms might set policies, how communities are affected by differing levels of personalization, and competition between platforms. Our model provides a vocabulary and mathematically tractable framework for analyzing platform decisions about content moderation.

Professional networks -- the social networks among people in a given line of work -- can serve as a conduit for job prospects and other opportunities. Here we propose a model for the formation of such networks and the transfer of opportunities within them. In our theoretical model, individuals strategically connect with others to maximize the probability that they receive opportunities from them. We explore how professional networks balance connectivity, where connections facilitate opportunity transfers to those who did not get them from outside sources, and congestion, where some individuals receive too many opportunities from their connections and waste some of them. We show that strategic individuals are over-connected at equilibrium relative to a social optimum, leading to a price of anarchy for which we derive nearly tight asymptotic bounds. We also show that, at equilibrium, individuals form connections to those who provide similar benefit to them as they provide to others. Thus, our model provides a microfoundation in professional networking contexts for the fundamental sociological principle of homophily, that "similarity breeds connection," which in our setting is realized as a form of status homophily based on alignment in individual benefit. We further explore how, even if individuals are a priori equally likely to receive opportunities from outside sources, equilibria can be unequal, and we provide nearly tight bounds on how unequal they can be. Finally, we explore the ability for online platforms to intervene to improve social welfare and show that natural heuristics may result in adverse effects at equilibrium. Our simple model allows for a surprisingly rich analysis of coordination problems in professional networks and suggests many directions for further exploration.

The development of generative language models that can create long and coherent textual outputs via autoregression has lead to a proliferation of uses and a corresponding sweep of analyses as researches work to determine the limitations of this new paradigm. Unlike humans, these '*Large Language Models*' (LLMs) are highly sensitive to small changes in their inputs, leading to unwanted inconsistency in their behavior. One problematic inconsistency when LLMs are used to answer multiple-choice questions or analyze multiple inputs is *order dependency*: the output of an LLM can (and often does) change significantly when sub-sequences are swapped, despite both orderings being semantically identical. In this paper we present , a technique that *guarantees* the output of an LLM will not have order dependence on a specified set of sub-sequences. We show that this method *provably* eliminates order dependency, and that it can be applied to *any* transformer-based LLM to enable text generation that is unaffected by re-orderings. Delving into the implications of our method, we show that, despite our inputs being out of distribution, the impact on expected accuracy is small, where the expectation is over the order of uniformly chosen shuffling of the candidate responses, and usually significantly less in practice. Thus, can be used as a '*dropped-in*' method on fully trained models. Finally, we discuss how our method's success suggests that other strong guarantees can be obtained on LLM performance via modifying the input representations. Code is available at [github.com/reidmcy/set-based-prompting](https://github.com/reidmcy/set-based-prompting.).

The US Decennial Census provides valuable data for both research and policy purposes. Census data are subject to a variety of disclosure avoidance techniques prior to release in order to preserve respondent confidentiality. While many are interested in studying the impacts of disclosure avoidance methods on downstream analyses, particularly with the introduction of differential privacy in the 2020 Decennial Census, these efforts are limited by a critical lack of data: The underlying "microdata," which serve as necessary input to disclosure avoidance methods, are kept confidential. In this work, we aim to address this limitation by providing tools to generate synthetic microdata solely from published Census statistics, which can then be used as input to any number of disclosure avoidance algorithms for the sake of evaluation and carrying out comparisons. We define a principled distribution over microdata given published Census statistics and design algorithms to sample from this distribution. We formulate synthetic data generation in this context as a knapsack-style combinatorial optimization problem and develop novel algorithms for this setting. While the problem we study is provably hard, we show empirically that our methods work well in practice, and we offer theoretical arguments to explain our performance. Finally, we verify that the data we produce are "close" to the desired ground truth.

We present connections between the recent literature on multigroup fairness for prediction algorithms and classical results in computational complexity. Multiaccurate predictors are correct in expectation on each member of an arbitrary collection of pre-specified sets. Multicalibrated predictors satisfy a stronger condition: they are calibrated on each set in the collection. Multiaccuracy is equivalent to a regularity notion for functions defined by Trevisan, Tulsiani, and Vadhan (2009). They showed that, given a class $F$ of (possibly simple) functions, an arbitrarily complex function $g$ can be approximated by a low-complexity function $h$ that makes a small number of oracle calls to members of $F$, where the notion of approximation requires that $h$ cannot be distinguished from $g$ by members of $F$. This complexity-theoretic Regularity Lemma is known to have implications in different areas, including in complexity theory, additive number theory, information theory, graph theory, and cryptography. Starting from the stronger notion of multicalibration, we obtain stronger and more general versions of a number of applications of the Regularity Lemma, including the Hardcore Lemma, the Dense Model Theorem, and the equivalence of conditional pseudo-min-entropy and unpredictability. For example, we show that every boolean function (regardless of its hardness) has a small collection of disjoint hardcore sets, where the sizes of those hardcore sets are related to how balanced the function is on corresponding pieces of an efficient partition of the domain.

We identify a novel connection between the recent literature on multi-group fairness for prediction algorithms and well-established notions of graph regularity from extremal graph theory. We frame our investigation using new, statistical distance-based variants of multi-calibration that are closely related to the concept of outcome indistinguishability . Adopting this perspective leads us naturally not only to our graph theoretic results, but also to new multi-calibration algorithms with improved complexity in certain parameter regimes, and to a generalization of a state-of-the-art result on omniprediction . Along the way, we also unify several prior algorithms for achieving multi-group fairness, as well as their analyses, through the lens of no-regret learning.