Jacob Steinhardt

Learning from untrusted data

The vast majority of theoretical results in machine learning and statistics assume that the training data is a reliable reflection of the phenomena to be learned. Similarly, most learning techniques used in practice are brittle to the presence of large amounts of biased or malicious data. Motivated by this, we consider two frameworks for studying estimation, learning, and optimization in the presence of significant fractions of arbitrary data. The first framework, list-decodable learning, asks whether it is possible to return a list of answers such that at least one is accurate. For example, given a dataset of n points for which an unknown subset of αn points are drawn from a distribution of interest, and no assumptions are made about the remaining (1 - α)n points, is it possible to return a list of poly(1/α) answers? The second framework, which we term the semi-verified model, asks whether a small dataset of trusted data (drawn from the distribution in question) can be used to extract accurate information from a much larger but untrusted dataset (of which only an α-fraction is drawn from the distribution). We show strong positive results in both settings, and provide an algorithm for robust learning in a very general stochastic optimization setting. This result has immediate implications for robustly estimating the mean of distributions with bounded second moments, robustly learning mixtures of such distributions, and robustly finding planted partitions in random graphs in which significant portions of the graph have been perturbed by an adversary.

Finite-time regional verification of stochastic non-linear systems

Recent trends pushing robots into unstructured environments with limited sensors have motivated considerable work on planning under uncertainty and stochastic optimal control, but these methods typically do not provide guaranteed performance. Here we consider the problem of bounding the probability of failure (defined as leaving a finite region of state space) over a finite time for stochastic non-linear systems with continuous state. Our approach searches for exponential barrier functions that provide bounds using a variant of the classical supermartingale result. We provide a relaxation of this search to a semidefinite program, yielding an efficient algorithm that provides rigorous upper bounds on the probability of failure for the original non-linear system. We give a number of numerical examples in both discrete and continuous time that demonstrate the effectiveness of the approach.

The Malicious Use of Artificial Intelligence: Forecasting, Prevention, and Mitigation

The following organisations are named on the report: Future of Humanity Institute, University of Oxford, Centre for the Study of Existential Risk, University of Cambridge, Center for a New American Security, Electronic Frontier Foundation, OpenAI. The Future of Life Institute is acknowledged as a funder.

Robust moment estimation and improved clustering via sum of squares

We develop efficient algorithms for estimating low-degree moments of unknown distributions in the presence of adversarial outliers and design a new family of convex relaxations for k-means clustering based on sum-of-squares method. As an immediate corollary, for any γ > 0, we obtain an efficient algorithm for learning the means of a mixture of k arbitrary distributions in d in time dO(1/γ) so long as the means have separation Ω(kγ). This in particular yields an algorithm for learning Gaussian mixtures with separation Ω(kγ), thus partially resolving an open problem of Regev and Vijayaraghavan regev2017learning. The guarantees of our robust estimation algorithms improve in many cases significantly over the best previous ones, obtained in the recent works. We also show that the guarantees of our algorithms match information-theoretic lower-bounds for the class of distributions we consider. These improved guarantees allow us to give improved algorithms for independent component analysis and learning mixtures of Gaussians in the presence of outliers. We also show a sharp upper bound on the sum-of-squares norms for moment tensors of any distribution that satisfies the Poincare inequality. The Poincare inequality is a central inequality in probability theory, and a large class of distributions satisfy it including Gaussians, product distributions, strongly log-concave distributions, and any sum or uniformly continuous transformation of such distributions. As a consequence, this yields that all of the above algorithmic improvements hold for distributions satisfying the Poincare inequality.