We consider the problem of speaker diarization, the problem of segmenting an audio recording of a meeting into temporal segments corresponding to individual speakers. The problem is rendered particularly difficult by the fact that we are not allowed to assume knowledge of the number of people participating in the meeting. To address this problem, we take a Bayesian nonparametric approach to speaker diarization that builds on the hierarchical Dirichlet process hidden Markov model (HDP-HMM) of Teh et al. [J. Amer. Statist. Assoc. 101 (2006) 1566--1581]. Although the basic HDP-HMM tends to over-segment the audio data---creating redundant states and rapidly switching among them---we describe an augmented HDP-HMM that provides effective control over the switching rate. We also show that this augmentation makes it possible to treat emission distributions nonparametrically. To scale the resulting architecture to realistic diarization problems, we develop a sampling algorithm that employs a truncated approximation of the Dirichlet process to jointly resample the full state sequence, greatly improving mixing rates. Working with a benchmark NIST data set, we show that our Bayesian nonparametric architecture yields state-of-the-art speaker diarization results.

Inference in popular nonparametric Bayesian models typically relies on sampling or other approximations. This paper presents a general methodology for constructing novel tractable nonparametric Bayesian methods by applying the kernel trick to inference in a parametric Bayesian model. For example, Gaussian process regression can be derived this way from Bayesian linear regression. Despite the success of the Gaussian process framework, the kernel trick is rarely explicitly considered in the Bayesian literature. In this paper, we aim to fill this gap and demonstrate the potential of applying the kernel trick to tractable Bayesian parametric models in a wider context than just regression. As an example, we present an intuitive Bayesian kernel machine for density estimation that is obtained by applying the kernel trick to a Gaussian generative model in feature space.

We present a first theoretical analysis of the power of polynomial-time preprocessing for important combinatorial problems from various areas in AI. We consider problems from Constraint Satisfaction, Global Constraints, Satisfiability, Nonmonotonic and Bayesian Reasoning. We show that, subject to a complexity theoretic assumption, none of the considered problems can be reduced by polynomial-time preprocessing to a problem kernel whose size is polynomial in a structural problem parameter of the input, such as induced width or backdoor size. Our results provide a firm theoretical boundary for the performance of polynomial-time preprocessing algorithms for the considered problems.

We develop the distance dependent Chinese restaurant process (CRP), a flexible class of distributions over partitions that allows for non-exchangeability. This class can be used to model many kinds of dependencies between data in infinite clustering models, including dependencies across time or space. We examine the properties of the distance dependent CRP, discuss its connections to Bayesian nonparametric mixture models, and derive a Gibbs sampler for both observed and mixture settings. We study its performance with three text corpora. We show that relaxing the assumption of exchangeability with distance dependent CRPs can provide a better fit to sequential data. We also show its alternative formulation of the traditional CRP leads to a faster-mixing Gibbs sampling algorithm than the one based on the original formulation.

Graphical Gaussian models have proven to be useful tools for exploring network structures based on multivariate data. Applications to studies of gene expression have generated substantial interest in these models, and resulting recent progress includes the development of fitting methodology involving penalization of the likelihood function. In this paper we advocate the use of multivariate $t$-distributions for more robust inference of graphs. In particular, we demonstrate that penalized likelihood inference combined with an application of the EM algorithm provides a computationally efficient approach to model selection in the $t$-distribution case. We consider two versions of multivariate $t$-distributions, one of which requires the use of approximation techniques. For this distribution, we describe a Markov chain Monte Carlo EM algorithm based on a Gibbs sampler as well as a simple variational approximation that makes the resulting method feasible in large problems.

Automatic detection of persuasion is essential for machine interaction on the social web. To facilitate automated persuasion detection, we present a novel microtext corpus derived from hostage negotiation transcripts as well as a detailed manual (codebook) for persuasion annotation. Our corpus, called the NPS Persuasion Corpus, consists of 37 transcripts from four sets of hostage negotiation transcriptions. Each utterance in the corpus is hand annotated for one of nine categories of persuasion based on Cialdini’s model: reciprocity, commitment, consistency, liking, authority, social proof, scarcity, other, and not persuasive. Initial results using three supervised learning algorithms (Na ̈ve Bayes, Maximum Entropy, and Support Vector Machines) combined with gappy and orthogonal sparse bigram feature expansion techniques show that the annotation process did capture machine learnable features of persuasion with F-scores better than baseline.

Skill discovery algorithms in reinforcement learning typically identify single states or regions in state space that correspond to potential task-specific subgoals. However, such methods do not directly address the question of how many distinct skills are appropriate for solving the tasks that the agent faces. This can be highly inefficient when many identified subgoals correspond to the same underlying skill, but are all used in- dividually as skill goals. Furthermore, skills created in this manner are often only transferable to tasks that share iden- tical state spaces, since corresponding subgoals across tasks are not merged into a single skill goal. We show that these problems can be overcome by clustering subgoal data defined in an agent-space and using the resulting clusters as templates for skill termination conditions. Clustering via a Dirichlet process mixture model is used to discover a minimal, suffi- cient collection of portable skills.

Recently, information gain has been proposed as a candidate intrinsic motivation for lifelong learning agents that may not always have a specific task.  In the InfoMax control framework, reinforcement learning is used to find a control policy for a POMDP in which movement and sensing actions are selected to reduce Shannon entropy as quickly as possible. In this study, we implement InfoMax control on a robot which can move between objects and perform sound-producing manipulations on them.  We formulate a novel latent variable mixture model for acoustic similarities and learn InfoMax polices that allow the robot to rapidly reduce uncertainty about the categories of the objects in a room. We find that InfoMax with our improved acoustic model leads to policies which lead to high classification accuracy.  Interestingly, we also find that with an insufficient model, the InfoMax policy eventually learns to "bury its head in the sand" to avoid getting additional evidence that might increase uncertainty.  We discuss the implications of this finding for InfoMax as a principle of intrinsic motivation in lifelong learning agents.

We present a model of interactive activity recognition and prompting for use in an assistive system for persons with cognitive disabilities. The system can determine the user’s state by interpreting sensor data and/or by explicitly querying the user, and can prompt the user to begin or end tasks. The objective of the system is to help the user maintain a daily schedule of activities while minimizing interruptions from questions or prompts. The model is built upon an option-based hierarchical POMDP. Options can be programmed and customized to specify complex routines for prompting or questioning. Novel aspects of the model include (1) the introduction of adaptive options, which employ a lightweight user model and are able to provide near-optimal performance with little exploration; (2) a restricted-inquiry dual-control algorithm that can appeal for help from the user when sensor data is ambiguous; and (3) a combined filtering / most likely-sequence algorithm for activities determining the beginning and ending time points of the user’s activities. Experiments show that each of these features contributes to the robustness of the model.

Crowdsourcing platforms, such as Amazon Mechanical Turk, have enabled the construction of scalable applications for tasks ranging from product categorization and photo tagging to audio transcription and translation. These vertical applications are typically realized with complex, self-managing workflows that guarantee quality results. But constructing such workflows is challenging, with a huge number of alternative decisions for the designer to consider. We argue the thesis that “Artificial intelligence methods can greatly simplify the process of creating and managing complex crowdsourced workflows.” We present the design of CLOWDER, which uses machine learning to continually refine models of worker performance and task difficulty. Using these models, CLOWDER uses decision-theoretic optimization to 1) choose between alternative workflows, 2) optimize parameters for a workflow, 3) create personalized interfaces for individual workers, and 4) dynamically control the workflow. Preliminary experience suggests that these optimized workflows are significantly more economical (and return higher quality output) than those generated by humans.