Self-play has powered breakthroughs in two-player and multi-player games. Here we show that self-play is a surprisingly effective strategy in another domain. We show that robust and naturalistic driving emerges entirely from self-play in simulation at unprecedented scale -- 1.6 billion km of driving. This is enabled by Gigaflow, a batched simulator that can synthesize and train on 42 years of subjective driving experience per hour on a single 8-GPU node. The resulting policy achieves state-of-the-art performance on three independent autonomous driving benchmarks. The policy outperforms the prior state of the art when tested on recorded real-world scenarios, amidst human drivers, without ever seeing human data during training. The policy is realistic when assessed against human references and achieves unprecedented robustness, averaging 17.5 years of continuous driving between incidents in simulation.

Estimating information-theoretic quantities such as entropy and mutual information is central to many problems in statistics and machine learning, but challenging in high dimensions. This paper presents estimators of entropy via inference (EEVI), which deliver upper and lower bounds on many information quantities for arbitrary variables in a probabilistic generative model. These estimators use importance sampling with proposal distribution families that include amortized variational inference and sequential Monte Carlo, which can be tailored to the target model and used to squeeze true information values with high accuracy. We present several theoretical properties of EEVI and demonstrate scalability and efficacy on two problems from the medical domain: (i) in an expert system for diagnosing liver disorders, we rank medical tests according to how informative they are about latent diseases, given a pattern of observed symptoms and patient attributes; and (ii) in a differential equation model of carbohydrate metabolism, we find optimal times to take blood glucose measurements that maximize information about a diabetic patient’s insulin sensitivity, given their meal and medication schedule.

A key design constraint when implementing Monte Carlo and variational inference algorithms is that it must be possible to cheaply and exactly evaluate the marginal densities of proposal distributions and variational families. This takes many interesting proposals off the table, such as those based on involved simulations or stochastic optimization. This paper broadens the design space, by presenting a framework for applying Monte Carlo and variational inference algorithms when proposal densities cannot be exactly evaluated. Our framework, recursive auxiliary-variable inference (RAVI), instead approximates the necessary densities using meta-inference: an additional layer of Monte Carlo or variational inference, that targets the proposal, rather than the model. RAVI generalizes and unifies several existing methods for inference with expressive approximating families, which we show correspond to specific choices of meta-inference algorithm, and provides new theory for analyzing their bias and variance. We illustrate RAVI’s design framework and theorems by using them to analyze and improve upon Salimans et al.’s Markov Chain Variational Inference, and to design a novel sampler for Dirichlet process mixtures, achieving state-of-the-art results on a standard benchmark dataset from astronomy and on a challenging datacleaning task with Medicare hospital data.

We present 3DP3, a framework for inverse graphics that uses inference in a structured generative model of objects, scenes, and images. 3DP3 uses (i) voxel models to represent the 3D shape of objects, (ii) hierarchical scene graphs to decompose scenes into objects and the contacts between them, and (iii) depth image likelihoods based on real-time graphics. Given an observed RGB-D image, 3DP3's inference algorithm infers the underlying latent 3D scene, including the object poses and a parsimonious joint parametrization of these poses, using fast bottom-up pose proposals, novel involutive MCMC updates of the scene graph structure, and, optionally, neural object detectors and pose estimators. We show that 3DP3 enables scene understanding that is aware of 3D shape, occlusion, and contact structure. Our results demonstrate that 3DP3 is more accurate at 6DoF object pose estimation from real images than deep learning baselines and shows better generalization to challenging scenes with novel viewpoints, contact, and partial observability.

Modern probabilistic programming languages aim to formalize and automate key aspects of probabilistic modeling and inference. Many languages provide constructs for programmable inference that enable developers to improve inference speed and accuracy by tailoring an algorithm for use with a particular model or dataset. Unfortunately, it is easy to use these constructs to write unsound programs that appear to run correctly but produce incorrect results. To address this problem, we present a denotational semantics for programmable inference in higher-order probabilistic programming languages, along with a type system that ensures that well-typed inference programs are sound by construction. A central insight is that the type of a probabilistic expression can track the space of its possible execution traces, not just the type of value that it returns, as these traces are often the objects that inference algorithms manipulate. We use our semantics and type system to establish soundness properties of custom inference programs that use constructs for variational, sequential Monte Carlo, importance sampling, and Markov chain Monte Carlo inference.

Although probabilistic programming is widely used for some restricted classes of statistical models, existing systems lack the flexibility and efficiency needed for practical use with more challenging models arising in fields like computer vision and robotics. This paper introduces Gen, a general-purpose probabilistic programming system that achieves modeling flexibility and inference efficiency via several novel language constructs: (i) the generative function interface for encapsulating probabilistic models; (ii) interoperable modeling languages that strike different flexibility/efficiency trade-offs; (iii) combinators that exploit common patterns of conditional independence; and (iv) an inference library that empowers users to implement efficient inference algorithms at a high level of abstraction. We show that Gen outperforms state-of-the-art probabilistic programming systems, sometimes by multiple orders of magnitude, on diverse problems including object tracking, estimating 3D body pose from a depth image, and inferring the structure of a time series.

We present new techniques for automatically constructing probabilistic programs for data analysis, interpretation, and prediction. These techniques work with probabilistic domain-specific data modeling languages that capture key properties of a broad class of data generating processes, using Bayesian inference to synthesize probabilistic programs in these modeling languages given observed data. We provide a precise formulation of Bayesian synthesis for automatic data modeling that identifies sufficient conditions for the resulting synthesis procedure to be sound. We also derive a general class of synthesis algorithms for domain-specific languages specified by probabilistic context-free grammars and establish the soundness of our approach for these languages. We apply the techniques to automatically synthesize probabilistic programs for time series data and multivariate tabular data. We show how to analyze the structure of the synthesized programs to compute, for key qualitative properties of interest, the probability that the underlying data generating process exhibits each of these properties. Second, we translate probabilistic programs in the domain-specific language into probabilistic programs in Venture, a general-purpose probabilistic programming system. The translated Venture programs are then executed to obtain predictions of new time series data and new multivariate data records. Experimental results show that our techniques can accurately infer qualitative structure in multiple real-world data sets and outperform standard data analysis methods in forecasting and predicting new data.

We present a novel approach for approximate sampling in probabilistic programs based on incremental inference. The key idea is to adapt the samples for a program P into samples for a program Q, thereby avoiding the expensive sampling computation for program Q. To enable incremental inference in probabilistic programming, our work: (i) introduces the concept of a trace translator which adapts samples from P into samples of Q, (ii) phrases this translation approach in the context of sequential Monte Carlo (SMC), which gives theoretical guarantees that the adapted samples converge to the distribution induced by Q, and (iii) shows how to obtain a concrete trace translator by establishing a correspondence between the random choices of the two probabilistic programs. We implemented our approach in two different probabilistic programming systems and showed that, compared to methods that sample the program Q from scratch, incremental inference can lead to orders of magnitude increase in efficiency, depending on how closely related P and Q are.

Approximate probabilistic inference algorithms are central to many fields. Examples include sequential Monte Carlo inference in robotics, variational inference in machine learning, and Markov chain Monte Carlo inference in statistics. A key problem faced by practitioners is measuring the accuracy of an approximate inference algorithm on a specific data set. This paper introduces the auxiliary inference divergence estimator (AIDE), an algorithm for measuring the accuracy of approximate inference algorithms. AIDE is based on the observation that inference algorithms can be treated as probabilistic models and the random variables used within the inference algorithm can be viewed as auxiliary variables. This view leads to a new estimator for the symmetric KL divergence between the approximating distributions of two inference algorithms. The paper illustrates application of AIDE to algorithms for inference in regression, hidden Markov, and Dirichlet process mixture models. The experiments show that AIDE captures the qualitative behavior of a broad class of inference algorithms and can detect failure modes of inference algorithms that are missed by standard heuristics.

OBJECTIVE: Social networks have been used in the study of outbreaks of infectious diseases, including in small group settings such as individual hospitals. Collecting the data needed to create such networks, however, can be time consuming, costly, and error prone. We sought to create a social network of hospital inpatients using electronic medical record (EMR) data already collected for other purposes, for use in simulating outbreaks of nosocomial infections. MATERIALS AND METHODS: We used the EMR data warehouse of a tertiary academic hospital to model contact among inpatients. Patient-to-patient contact due to shared rooms was inferred from admission-discharge-transfer data, and contact with healthcare workers was inferred from clinical documents. Contacts were used to generate a social network, which was then used to conduct probabilistic simulations of nosocomial outbreaks of methicillin-resistant Staphylococcus aureus and influenza. RESULTS: Simulations of infection transmission across the network reflected the staffing and patient flow practices of the hospital. Simulations modeling patient isolation, increased hand hygiene, and staff vaccination showed a decrease in the spread of infection. DISCUSSION: We developed a method of generating a social network of hospital inpatients from EMR data. This method allows the derivation of networks that reflect the local hospital environment, obviate the need for simulated or manually collected data, and can be updated in near real time. CONCLUSIONS: Inpatient social networks represent a novel secondary use of EMR data, and can be used to simulate nosocomial infections. Future work should focus on prospective validation of the simulations, and adapting such networks to other tasks.

We consider the problem of autonomously bringing an article of clothing into a desired configuration using a general-purpose two-armed robot. We propose a hidden Markov model (HMM) for estimating the identity of the article and tracking the article's configuration throughout a specific sequence of manipulations and observations. At the end of this sequence, the article's configuration is known, though not necessarily desired. The estimated identity and configuration of the article are then used to plan a second sequence of manipulations that brings the article into the desired configuration. We propose a relaxation of a strain limiting finite element model for cloth simulation that can be solved via convex optimization; this serves as the basis of the transition and observation models of the HMM. The observation model uses simple perceptual cues consisting of the height of the article when held by a single gripper and the silhouette of the article when held by two grippers. The model accurately estimates the identity and configuration of clothing articles, enabling our procedure to autonomously bring a variety of articles into desired configurations that are useful for other tasks, such as folding.

We present a novel vision-based grasp point detection algorithm that can reliably detect the corners of a piece of cloth, using only geometric cues that are robust to variation in texture. Furthermore, we demonstrate the effectiveness of our algorithm in the context of folding a towel using a general-purpose two-armed mobile robotic platform without the use of specialized end-effectors or tools. The robot begins by picking up a randomly dropped towel from a table, goes through a sequence of vision-based re-grasps and manipulations-partially in the air, partially on the table-and finally stacks the folded towel in a target location. The reliability and robustness of our algorithm enables for the first time a robot with general purpose manipulators to reliably and fully-autonomously fold previously unseen towels, demonstrating success on all 50 out of 50 single-towel trials as well as on a pile of 5 towels.