Inspired by the hierarchical hidden Markov models (HHMM), we present the hierarchical semi-Markov conditional random field (HSCRF), a generalisation of embedded undirectedMarkov chains tomodel complex hierarchical, nestedMarkov processes. It is parameterised in a discriminative framework and has polynomial time algorithms for learning and inference. Importantly, we consider partiallysupervised learning and propose algorithms for generalised partially-supervised learning and constrained inference. We demonstrate the HSCRF in two applications: (i) recognising human activities of daily living (ADLs) from indoor surveillance cameras, and (ii) noun-phrase chunking. We show that the HSCRF is capable of learning rich hierarchical models with reasonable accuracy in both fully and partially observed data cases.

The Aviation Safety Reporting System collects voluntarily submitted reports on aviation safety incidents to facilitate research work aiming to reduce such incidents. To effectively reduce these incidents, it is vital to accurately identify why these incidents occurred. More precisely, given a set of possible causes, or shaping factors, this task of cause identification involves identifying all and only those shaping factors that are responsible for the incidents described in a report. We investigate two approaches to cause identification. Both approaches exploit information provided by a semantic lexicon, which is automatically constructed via Thelen and Riloff's Basilisk framework augmented with our linguistic and algorithmic modifications. The first approach labels a report using a simple heuristic, which looks for the words and phrases acquired during the semantic lexicon learning process in the report. The second approach recasts cause identification as a text classification problem, employing supervised and transductive text classification algorithms to learn models from incident reports labeled with shaping factors and using the models to label unseen reports. Our experiments show that both the heuristic-based approach and the learning-based approach (when given sufficient training data) outperform the baseline system significantly.

Much of the ongoing explosion of digital content is in the form of text. This content is a virtual gold-mine of information that can inform a range of social, governmental, and business decisions. For example, using content available on blogs and social networking sites businesses can find out what its customers are saying about their products and services. In the digital age where customer is king, the business value of ascertaining consumer sentiment cannot be overstated. People express sentiments in myriad ways. At times, they use simple, direct assertions, but most often they use sentences involving comparisons, conjunctions expressing multiple and possibly opposing sentiments about multiple features and entities,and pronominal references whose resolution requires discourse level context. Frequently people use abbreviations, slang, SMSese, idioms and metaphors. Understanding the latter also requires common sense reasoning. In this paper, we present iSEE, a fully implemented sentiment extraction engine, which makes use of statistical methods, classical NLU techniques, common sense reasoning, and probabilistic inference to extract entity and feature specific sentiment from complex sentences and dialog. Most of the components of iSEE are domain independent and the system can be generalized to new domains by simply adding domain relevant lexicons.

Recognizing activities in real-world videos is a difficult problem exacerbated by background clutter, changes in camera angle & zoom, and rapid camera movements. Large corpora of labeled videos can be used to train automated activity recognition systems, but this requires expensive human labor and time. This paper explores how closed captions that naturally accompany many videos can act as weak supervision that allows automatically collecting "labeled" data for activity recognition. We show that such an approach can improve activity retrieval in soccer videos. Our system requires no manual labeling of video clips and needs minimal human supervision. We also present a novel caption classifier that uses additional linguistic information to determine whether a specific comment refers to an ongoing activity. We demonstrate that combining linguistic analysis and automatically trained activity recognizers can significantly improve the precision of video retrieval.

Parsing plays an important role in semantic role labeling (SRL) because most SRL systems infer semantic relations from 1-best parses. Therefore, parsing errors inevitably lead to labeling mistakes. To alleviate this problem, we propose to use packed forest, which compactly encodes all parses for a sentence. We design an algorithm to exploit exponentially many parses to learn semantic relations efciently. Experimental results on the CoNLL-2005 shared task show that using forests achieves an absolute improvement of 1.2% in terms of F1 score over using 1-best parses and 0.6% over using 50-best parses.

Building an intelligent agent, which simulates human-level learning appropriate for learning math, science, or a second language, could potentially benefit both education in understanding human learning, and artificial intelligence in creating human-level intelligence. Recently, we have proposed an efficient approach to acquiring procedural knowledge using transfer learning. However, it operated as a separate module. In this paper, we describe how to integrate this module into a machine-learning agent, SimStudent, that learns procedural knowledge from examples and through problem solving. We illustrate this method in the domain of algebra, after which we consider directions for future research in this area.

The quality and nature of knowledge that can be found by an automated knowledge-extraction system depends on its inputs. For systems that learn by reading text, the Web offers a breadth of topics and currency, but it also presents the problems of dealing with casual, unedited writing, non-textual inputs, and the mingling of languages. The results of extraction using the KNEXT system on two Web corpora — Wikipedia and a collection of weblog entries — indicate that, with automatic filtering of the output, even ungrammatical writing on arbitrary topics can yield an extensive knowledge base, which human judges find to be of good quality, with propositions receiving an average score across both corpora of 2.34 (where the range is 1 to 5 and lower is better) versus 3.00 for unfiltered output from the same sources.

Machine reading aims to automatically extract knowledge from text. It is a long-standing goal of AI and holds the promise of revolutionizing Web search and other fields. In this paper, we analyze the core challenges of machine reading and show that statistical relational AI is particularly well suited to address these challenges. We then propose a unifying approach to machine reading in which statistical relational AI plays a central role. Finally, we demonstrate the promise of this approach by presenting OntoUSP, an end-to-end machine reading system that builds on recent advances in statistical relational AI and greatly outperforms state-of-the-art systems in a task of extracting knowledge from biomedical abstracts and answering questions.

We propose automatically learning probabilistic Hierarchical Task Networks (pHTNs) in order to capture a user's preferences on plans, by observing only the user's behavior. HTNs are a common choice of representation for a variety of purposes in planning, including work on learning in planning. Our contributions are (a) learning structure and (b) representing preferences. In contrast, prior work employing HTNs considers learning method preconditions (instead of structure) and representing domain physics or search control knowledge (rather than preferences). Initially we will assume that the observed distribution of plans is an accurate representation of user preference, and then generalize to the situation where feasibility constraints frequently prevent the execution of preferred plans. In order to learn a distribution on plans we adapt an Expectation-Maximization (EM) technique from the discipline of (probabilistic) grammar induction, taking the perspective of task reductions as productions in a context-free grammar over primitive actions. To account for the difference between the distributions of possible and preferred plans we subsequently modify this core EM technique, in short, by rescaling its input.

Aiming at accurate recognition of relations, we introduce local alignment kernels and explore various possibilities of using them for this task. We give a definition of a local alignment (LA) kernel based on the Smith-Waterman score as a sequence similarity measure and proceed with a range of possibilities for computing similarity between elements of sequences. We show how distributional similarity measures obtained from unlabeled data can be incorporated into the learning task as semantic knowledge. Our experiments suggest that the LA kernel yields promising results on various biomedical corpora outperforming two baselines by a large margin. Additional series of experiments have been conducted on the data sets of seven general relation types, where the performance of the LA kernel is comparable to the current state-of-the-art results.