Information Extraction methods can be used to automatically "fill-in" database forms from unstructured data such as Web documents or email. State-of-the-art methods have achieved low error rates but invariably make a number of errors. The goal of an interactive information extraction system is to assist the user in filling in database fields while giving the user confidence in the integrity of the data. The user is presented with an interactive interface that allows both the rapid verification of automatic field assignments and the correction of errors. In cases where there are multiple errors, our system takes into account user corrections, and immediately propagates these constraints such that other fields are often corrected automatically.

Common approaches to multi-label classification learn independent classifiers for each category, and employ ranking or thresholding schemes for classification. Because they do not exploit dependencies between labels, such techniques are only well-suited to problems in which categories are independent. However, in many domains labels are highly interdependent. This paper explores multilabel conditional random field (CRF) classification models that directly parameterize label co-occurrences in multi-label classification. Experiments show that the models outperform their singlelabel counterparts on standard text corpora. Even when multilabels are sparse, the models improve subset classification error by as much as 40%.

Conditional Random Fields (CRFs; Lafferty, McCallum, & Pereira, 2001) provide a flexible and powerful model for learning to assign labels to elements of sequences in such applications as part-of-speech tagging, text-to-speech mapping, protein and DNA sequence analysis, and information extraction from web pages. However, existing learning algorithms are slow, particularly in problems with large numbers of potential input features. This paper describes a new method...

Recent work on graphical models for relational data has demonstrated significant improvements in classification and inference when models represent the dependencies among instances. Despite its use in conventional statistical models, the assumption of instance independence is contradicted by most relational datasets. For example, in citation data there are dependencies among the topics of a paper’s references, and in genomic data there are dependencies among the functions of interacting proteins. In this paper, we present relational dependency networks (RDNs), graphical models that are capable of expressing and reasoning with such dependencies in a relational setting. We discuss RDNs in the context of relational Bayes networks and relational Markov networks and outline the relative strengths of RDNs—namely, the ability to represent cyclic dependencies, simple methods for parameter estimation, and efficient structure learning techniques. The strengths of RDNs are due to the use of pseudolikelihood learning techniques, which estimate an efficient approximation of the full joint distribution. We present learned RDNs for a number of real-world datasets and evaluate the models in a prediction context, showing that RDNs identify and exploit cyclic relational dependencies to achieve significant performance gains over conventional conditional models. In addition, we use synthetic data to explore model performance under various relational data characteristics, showing that RDN learning and inference techniques are accurate over a wide range of conditions.

Learning patterns of human behavior from sensor data is extremely important for high-level activity inference. We show how to extract and label a person’s activities and significant places from traces of GPS data. In contrast to existing techniques, our approach simultaneously detects and classifies the significant locations of a person and takes the high-level context into account. Our system uses relational Markov networks to represent the hierarchical activity model that encodes the complex relations among GPS readings, activities and significant places. We apply FFT-based message passing to perform efficient summation over large numbers of nodes in the networks. We present experiments that show significant improvements over existing techniques. 1

We propose Bayesian Conditional Random Fields (BCRFs) for classifying interdependent and structured data, such as sequences, images or webs. BCRFs are a Bayesian approach to training and inference with conditional random fields, which were previously trained by maximizing likelihood (ML) (Lafferty et al., 2001). Our framework eliminates the problem of overfitting, and offers the full advantages of a Bayesian treatment. Unlike the ML approach, we estimate the posterior distribution of the model parameters during training, and average over this posterior during inference. We apply an extension of EP method, the power EP method, to incorporate the partition function. For algorithmic stability and accuracy, we flatten the approximation structures to avoid two-level approximations. We demonstrate the superior prediction accuracy of BCRFs over conditional random fields trained with ML or MAP on synthetic and real datasets. 1

Abstract not found

We study the computational and sample complexity of parameter and structure learning in graphical models. Our main result shows that the class of factor graphs with bounded degree can be learned in polynomial time and from a polynomial number of training examples, assuming that the data is generated by a network in this class. This result covers both parameter estimation for a known network structure and structure learning. It implies as a corollary that we can learn factor graphs for both Bayesian networks and Markov networks of bounded degree, in polynomial time and sample complexity. Importantly, unlike standard maximum likelihood estimation algorithms, our method does not require inference in the underlying network, and so applies to networks where inference is intractable. We also show that the error of our learned model degrades gracefully when the generating distribution is not a member of the target class of networks. In addition to our main result, we show that the sample complexity of parameter learning in graphical models has an O(1) dependence on the number of variables in the model when using the KL-divergence normalized by the number of variables as the performance criterion.

Office workers everywhere are drowning in email—not only spam, but also large quantities of legitimate email to be read and organized for browsing. Although there have been extensive investigations of automatic document categorization, email gives rise to a number of unique challenges, and there has been relatively little study of classifying email into folders. This paper presents an extensive benchmark study of email foldering using two large corpora of real-world email messages and foldering schemes: one from former Enron employees, another from participants in an SRI research project. We discuss the challenges that arise from differences between email foldering and traditional document classification. We show experimental results from an array of automated classification methods and evaluation methodologies, including a new evaluation method of foldering results based on the email timeline, and including enhancements to the exponential gradient method Winnow, providing top-tier accuracy with a fraction the training time of alternative methods. We also establish that classification accuracy in many cases is relatively low, confirming the challenges of email data, and pointing toward email foldering as an important area for further research. 1.

There is an active debate within the field of phonology concerning the cognitive status of substantive phonetic factors such as ease of articulation and perceptual distinctiveness. A new framework is proposed in which substance acts as a bias, or prior, on phonological learning. Two experiments tested this framework with a method in which participants are first provided highly impoverished evidence of a new phonological pattern, and then tested on how they extend this pattern to novel contexts and novel sounds. Participants were found to generalize velar palatalization (e.g., the change from [k]asinkeep to [t�ʃ]asincheap) in a way that accords with linguistic typology, and that is predicted by a cognitive bias in favor of changes that relate perceptually similar sounds. Velar palatalization was extended from the mid front vowel context (i.e., before [e]asincape) to the high front vowel context (i.e., before [i]asin keep), but not vice versa. The key explanatory notion of perceptual similarity is quantified with a psychological model of categorization, and the substantively biased framework is formalized as a conditional random field. Implications of these results for the debate on substance, theories of phonological generalization, and the formalization of similarity are discussed.