We resolve a 30-year-old open problem concerning the power of unlabeled data in online learning by tightly quantifying the gap between transductive and standard online learning. We prove that for every concept class $\mathcal{H}$ with Littlestone dimension $d$, the transductive mistake bound is at least $\Omega(\sqrt{d})$. This establishes an exponential improvement over previous lower bounds of $\Omega(\log \log d)$, $\Omega(\sqrt{\log d})$, and $\Omega(\log d)$, respectively due to Ben-David, Kushilevitz, and Mansour (1995, 1997) and Hanneke, Moran, and Shafer (2023). We also show that our bound is tight: for every $d$, there exists a class of Littlestone dimension $d$ with transductive mistake bound $O(\sqrt{d})$. Our upper bound also improves the previous best known upper bound of $(2/3) \cdot d$ from Ben-David et al. (1997). These results demonstrate a quadratic gap between transductive and standard online learning, thereby highlighting the benefit of advanced access to the unlabeled instance sequence. This stands in stark contrast to the PAC setting, where transductive and standard learning exhibit similar sample complexities.

We study protocols for verifying approximate optimality of strategies in multi-armed bandits and normal-form games. As the number of actions available to each player is often large, we seek protocols where the number of queries to the utility oracle is sublinear in the number of actions. We prove that such verification is possible for sufficiently smooth strategies that do not put too much probability mass on any specific action and provide protocols for verifying that a smooth policy for a multi-armed bandit is close to optimal. Our verification protocols require provably fewer arm queries than learning. Furthermore, we show how to use cryptographic tools to reduce the communication cost of our protocols. We complement our protocol by proving a nearly tight lower bound on the query complexity of verification in our settings. As an application, we use our bandit verification protocol to build a protocol for verifying approximate optimality of a strong smooth Nash equilibrium, with sublinear query complexity.

We study which machine learning algorithms have tight generalization bounds with respect to a given collection of population distributions. Our results build on and extend the recent work of Gastpar et al. (2023). First, we present conditions that preclude the existence of tight generalization bounds. Specifically, we show that algorithms that have certain inductive biases that cause them to be unstable do not admit tight generalization bounds. Next, we show that algorithms that are sufficiently loss-stable do have tight generalization bounds. We conclude with a simple characterization that relates the existence of tight generalization bounds to the conditional variance of the algorithm's loss.

We study the notion of a generalization bound being _uniformly tight_, meaning that the difference between the bound and the population loss is small for all learning algorithms and all population distributions. Numerous generalization bounds have been proposed in the literature as potential explanations for the ability of neural networks to generalize in the overparameterized setting. However, in their paper "Fantastic Generalization Measures and Where to Find Them," Jiang et al. (2020) examine more than a dozen generalization bounds, and show empirically that none of them are uniformly tight. This raises the question of whether uniformly-tight generalization bounds are at all possible in the overparameterized setting. We consider two types of generalization bounds: (1) bounds that may depend on the training set and the learned hypothesis (e.g., margin bounds). We prove mathematically that no such bound can be uniformly tight in the overparameterized setting; (2) bounds that may in addition also depend on the learning algorithm (e.g., stability bounds). For these bounds, we show a trade-off between the algorithm's performance and the bound's tightness. Namely, if the algorithm achieves good accuracy on certain distributions, then no generalization bound can be uniformly tight for it in the overparameterized setting. We explain how these formal results can, in our view, inform research on generalization bounds for neural networks, while stressing that other interpretations of these results are also possible.

Learning curves plot the expected error of a learning algorithm as a function of the number of labeled samples it receives from a target distribution. They are widely used as a measure of an algorithm’s performance, but classic PAC learning theory cannot explain their behavior. As observed by Antos and Lugosi (1996, 1998), the classic ‘No Free Lunch’ lower bounds only trace the upper envelope above all learning curves of specific target distributions. For a concept class with VC dimension $d$ the classic bound decays like $d/n$, yet it is possible that the learning curve for \emph{every} specific distribution decays exponentially. In this case, for each $n$ there exists a different ‘hard’ distribution requiring $d/n$ samples. Antos and Lugosi asked which concept classes admit a ‘strong minimax lower bound’ – a lower bound of $d’/n$ that holds for a fixed distribution for infinitely many $n$.We solve this problem in a principled manner, by introducing a combinatorial dimension called VCL that characterizes the best $d’$ for which $d’/n$ is a strong minimax lower bound. Conceptually, the VCL dimension determines the asymptotic rate of decay of the minimax learning curve, which we call the ‘distribution-free trail’ of the class. Our characterization strengthens the lower bounds of Bousquet, Hanneke, Moran, van Handel, and Yehudayoff (2021), and it refines their analysis of learning curves, by showing that for classes with finite VCL the learning rate can be decomposed into a linear component that depends only on the hypothesis class and a faster (e.g., exponential) component that depends also on the target distribution. As a corollary, we recover the lower bound of Antos and Lugosi (1996, 1998) for half-spaces in $\mathbb{R}^d$.Finally, to provide another viewpoint on our work and how it compares to traditional PAC learning bounds, we also present an alternative formulation of our results in a language that is closer to the PAC setting.

Goldwasser et al. (2021) recently proposed the setting of PAC verification, where a hypothesis (machine learning model) that purportedly satisfies the agnostic PAC learning objective is verified using an interactive proof. In this paper we develop this notion further in a number of ways. First, we prove a lower bound of $\Omega(\sqrt{d}/\varepsilon^2)$ i.i.d. samples for PAC verification of hypothesis classes of VC dimension $d$. Second, we present a protocol for PAC verification of unions of intervals over $\mathbb{R}$ that improves upon their proposed protocol for that task, and matches our lower bound’s dependence on $d$. Third, we introduce a natural generalization of their definition to verification of general statistical algorithms, which is applicable to a wider variety of settings beyond agnostic PAC learning. Showcasing our proposed definition, our final result is a protocol for the verification of statistical query algorithms that satisfy a combinatorial constraint on their queries.

We present new upper and lower bounds on the number of learner mistakes in the `transductive' online learning setting of Ben-David, Kushilevitz and Mansour (1997). This setting is similar to standard online learning, except that the adversary fixes a sequence of instances $x_1,\dots,x_n$ to be labeled at the start of the game, and this sequence is known to the learner. Qualitatively, we prove a \emph{trichotomy}, stating that the minimal number of mistakes made by the learner as $n$ grows can take only one of precisely three possible values: $n$, $\Theta\left(\log (n)\right)$, or $\Theta(1)$. Furthermore, this behavior is determined by a combination of the VC dimension and the Littlestone dimension. Quantitatively, we show a variety of bounds relating the number of mistakes to well-known combinatorial dimensions. In particular, we improve the known lower bound on the constant in the $\Theta(1)$ case from $\Omega\left(\sqrt{\log(d)}\right)$ to $\Omega(\log(d))$ where $d$ is the Littlestone dimension. Finally, we extend our results to cover multiclass classification and the agnostic setting.

We show that many definitions of stability found in the learning theory literature are equivalent to one another. We distinguish between two families of definitions of stability: distribution-dependent and distribution-independent Bayesian stability. Within each family, we establish equivalences between various definitions, encompassing approximate differential privacy, pure differential privacy, replicability, global stability, perfect generalization, TV stability, mutual information stability, KL-divergence stability, and Rényi-divergence stability. Along the way, we prove boosting results that enable the amplification of the stability of a learning rule. This work is a step towards a more systematic taxonomy of stability notions in learning theory, which can promote clarity and an improved understanding of an array of stability concepts that have emerged in recent years.

We consider the following question: using a source of labeled data and interaction with an untrusted prover, what is the complexity of verifying that a given hypothesis is "approximately correct"? We study interactive proof systems for PAC verification, where a verifier that interacts with a prover is required to accept good hypotheses, and reject bad hypotheses. Both the verifier and the prover are efficient and have access to labeled data samples from an unknown distribution. We are interested in cases where the verifier can use significantly less data than is required for (agnostic) PAC learning, or use a substantially cheaper data source (e.g., using only random samples for verification, even though learning requires membership queries). We believe that today, when data and data-driven algorithms are quickly gaining prominence, the question of verifying purported outcomes of data analyses is very well-motivated. We show three main results. First, we prove that for a specific hypothesis class, verification is significantly cheaper than learning in terms of sample complexity, even if the verifier engages with the prover only in a single-round (NP-like) protocol. Moreover, for this class we prove that single-round verification is also significantly cheaper than testing closeness to the class. Second, for the broad class of Fourier-sparse boolean functions, we show a multi-round (IP-like) verification protocol, where the prover uses membership queries, and the verifier is able to assess the result while only using random samples. Third, we show that verification is not always more efficient. Namely, we show a class of functions where verification requires as many samples as learning does, up to a logarithmic factor.

We study learning algorithms that are restricted to using a small amount of information from their input sample. We introduce a category of learning algorithms we term {\em $d$-bit information learners}, which are algorithms whose output conveys at most $d$ bits of information of their input. A central theme in this work is that such algorithms generalize. We focus on the learning capacity of these algorithms, and prove sample complexity bounds with tight dependencies on the confidence and error parameters. We also observe connections with well studied notions such as sample compression schemes, Occam’s razor, PAC-Bayes and differential privacy. We discuss an approach that allows us to prove upper bounds on the amount of information that algorithms reveal about their inputs, and also provide a lower bound by showing a simple concept class for which every (possibly randomized) empirical risk minimizer must reveal a lot of information. On the other hand, we show that in the distribution-dependent setting every VC class has empirical risk minimizers that do not reveal a lot of information.

How many bits of information are required to PAC learn a class of hypotheses of VC dimension $d$? The mathematical setting we follow is that of Bassily et al., where the value of interest is the mutual information $\mathrm{I}(S;A(S))$ between the input sample $S$ and the hypothesis outputted by the learning algorithm $A$. We introduce a class of functions of VC dimension $d$ over the domain $\mathcal{X}$ with information complexity at least $\Omega \left(d\log \log \frac{|\mathcal{X}|}{d}\right)$ bits for any consistent and proper algorithm (deterministic or random). Bassily et al. proved a similar (but quantitatively weaker) result for the case $d=1$. The above result is in fact a special case of a more general phenomenon we explore. We define the notion of {\em information complexity} of a given class of functions $\cH$. Intuitively, it is the minimum amount of information that an algorithm for $\mathcal{X}$ must retain about its input to ensure consistency and properness. We prove a direct sum result for information complexity in this context; roughly speaking, the information complexity sums when combining several classes.