Patrick Pantel

From frequency to meaning: vector space models of semantics

Computers understand very little of the meaning of human language. This profoundly limits our ability to give instructions to computers, the ability of computers to explain their actions to us, and the ability of computers to analyse and process text. Vector space models (VSMs) of semantics are beginning to address these limits. This paper surveys the use of VSMs for semantic processing of text. We organize the literature on VSMs according to the structure of the matrix in a VSM. There are currently three broad classes of VSMs, based on term-document, word-context, and pair-pattern matrices, yielding three classes of applications. We survey a broad range of applications in these three categories and we take a detailed look at a specific open source project in each category. Our goal in this survey is to show the breadth of applications of VSMs for semantics, to provide a new perspective on VSMs for those who are already familiar with the area, and to provide pointers into the literature for those who are less familiar with the field.

Discovering word senses from text

Inventories of manually compiled dictionaries usually serve as a source for word senses. However, they often include many rare senses while missing corpus/domain-specific senses. We present a clustering algorithm called CBC (Clustering By Committee) that automatically discovers word senses from text. It initially discovers a set of tight clusters called committees that are well scattered in the similarity space. The centroid of the members of a committee is used as the feature vector of the cluster. We proceed by assigning words to their most similar clusters. After assigning an element to a cluster, we remove their overlapping features from the element. This allows CBC to discover the less frequent senses of a word and to avoid discovering duplicate senses. Each cluster that a word belongs to represents one of its senses. We also present an evaluation methodology for automatically measuring the precision and recall of discovered senses.

Discovery of inference rules for question-answering

One of the main challenges in question-answering is the potential mismatch between the expressions in questions and the expressions in texts. While humans appear to use inference rules such as ‘X writes Y’ implies ‘X is the author of Y’ in answering questions, such rules are generally unavailable to question-answering systems due to the inherent difficulty in constructing them. In this paper, we present an unsupervised algorithm for discovering inference rules from text. Our algorithm is based on an extended version of Harris’ Distributional Hypothesis, which states that words that occurred in the same contexts tend to be similar. Instead of using this hypothesis on words, we apply it to paths in the dependency trees of a parsed corpus. Essentially, if two paths tend to link the same set of words, we hypothesize that their meanings are similar. We use examples to show that our system discovers many inference rules easily missed by humans.

Espresso: Leveraging Generic Patterns for Automatically Harvesting Semantic Relations

In this paper, we present Espresso, a weakly-supervised, general-purpose, and accurate algorithm for harvesting semantic relations. The main contributions are: i) a method for exploiting generic patterns by filtering incorrect instances using the Web; and ii) a principled measure of pattern and instance reliability enabling the filtering algorithm. We present an empirical comparison of Espresso with various state of the art systems, on different size and genre corpora, on extracting various general and specific relations. Experimental results show that our exploitation of generic patterns substantially increases system recall with small effect on overall precision.

Automatically Assessing Review Helpfulness

User-supplied reviews are widely and increasingly used to enhance e-commerce and other websites. Because reviews can be numerous and varying in quality, it is important to assess how helpful each review is. While review helpfulness is currently assessed manually, in this paper we consider the task of automatically assessing it. Experiments using SVM regression on a variety of features over Amazon.com product reviews show promising results, with rank correlations of up to 0.66. We found that the most useful features include the length of the review, its unigrams, and its product rating.

Web-Scale Distributional Similarity and Entity Set Expansion

Computing the pairwise semantic similarity between all words on the Web is a computationally challenging task. Parallelization and optimizations are necessary. We propose a highly scalable implementation based on distributional similarity, implemented in the MapReduce framework and deployed over a 200 billion word crawl of the Web. The pairwise similarity between 500 million terms is computed in 50 hours using 200 quad-core nodes. We apply the learned similarity matrix to the task of automatic set expansion and present a large empirical study to quantify the effect on expansion performance of corpus size, corpus quality, seed composition and seed size. We make public an experimental testbed for set expansion analysis that includes a large collection of diverse entity sets extracted from Wikipedia.

Document clustering with committees

Document clustering is useful in many information retrieval tasks: document browsing, organization and viewing of retrieval results, generation of Yahoo-like hierarchies of documents, etc. The general goal of clustering is to group data elements such that the intra-group similarities are high and the inter-group similarities are low. We present a clustering algorithm called CBC (Clustering By Committee) that is shown to produce higher quality clusters in document clustering tasks as compared to several well known clustering algorithms. It initially discovers a set of tight clusters (high intra-group similarity), called committees, that are well scattered in the similarity space (low inter-group similarity). The union of the committees is but a subset of all elements. The algorithm proceeds by assigning elements to their most similar committee. Evaluating cluster quality has always been a difficult task. We present a new evaluation methodology that is based on the editing distance between output clusters and manually constructed classes (the answer key). This evaluation measure is more intuitive and easier to interpret than previous evaluation measures.

Randomized Algorithms and NLP: Using Locality Sensitive Hash Functions for High Speed Noun Clustering

In this paper, we explore the power of randomized algorithm to address the challenge of working with very large amounts of data. We apply these algorithms to generate noun similarity lists from 70 million pages. We reduce the running time from quadratic to practically linear in the number of elements to be computed.

Concept discovery from text

Broad-coverage lexical resources such as WordNet are extremely useful. However, they often include many rare senses while missing domain-specific senses. We present a clustering algorithm called CBC (Clustering By Committee) that automatically discovers concepts from text. It initially discovers a set of tight clusters called committees that are well scattered in the similarity space. The centroid of the members of a committee is used as the feature vector of the cluster. We proceed by assigning elements to their most similar cluster. Evaluating cluster quality has always been a difficult task. We present a new evaluation methodology that is based on the editing distance between output clusters and classes extracted from WordNet (the answer key). Our experiments show that CBC outperforms several well-known clustering algorithms in cluster quality.

Towards terascale knowledge acquisition

Although vast amounts of textual data are freely available, many NLP algorithms exploit only a minute percentage of it. In this paper, we study the challenges of working at the terascale. We present an algorithm, designed for the teraxale, for mining is-a relations that achieves similar performance to a state-of-the-art linguistically-rich method. We focus on the accuracy of these two systems as a function of processing time and corpus size.