Ravi Kondadadi

Query-based opinion summarization for legal blog entries

We present the first report of automatic sentiment summarization in the legal domain. This work is based on processing a set of legal questions with a system consisting of a semi-automatic Web blog search module and FastSum, a fully automatic extractive multi-document sentiment summarization system. We provide quantitative evaluation results of the summaries using legal expert reviewers. We report baseline evaluation results for query-based sentiment summarization for legal blogs: on a five-point scale, average responsiveness and linguistic quality are slightly higher than 2 (with human inter-rater agreement at k = 0.75). To the best of our knowledge, this is the first evaluation of sentiment summarization in the legal blogosphere.

Fast tagging of medical terms in legal text

Medical terms occur across a wide variety of legal, medical, and news corpora. Documents containing these terms are of particular interest to legal professionals operating in such fields as medical malpractice, personal injury, and product liability. This paper describes a novel method of tagging medical terms in legal, medical, and news text that is very fast and also has high recall and precision. To date, most research in medical term spotting has been confined to medical text and has approached the problem by extracting noun phrases from sentences and mapping them to a list of medical concepts via a fuzzy lookup. The medical term tagging described in this paper relies on a fast finite state machine that finds within sentences the longest contiguous sets of words associated with medical terms in a medical term authority file, converts word sets into medical term hash keys, and looks up medical concept ids associated with the hash keys. Additionally our system relies on a probabilistic term classifier that uses local context to disambiguate terms being used in a medical sense from terms being used in a non-medical sense. Our method is two orders of magnitude faster than an approach based on noun phrase extraction and has better precision and recall for terms pertaining to injuries, diseases, drugs, medical procedures, and medical devices. The methods presented here have been implemented and are the core engines for a Thomson West product called the Medical Litigator. Thus far, the Medical Litigator has processed over 100 million documents and generated over 165 million tags representing approximately 164,000 unique medical concepts. The resulting system is very fast and posted a recall from 0.79 to 0.93 and precision between 0.94 and 0.97, depending on the document type.

A Metric for Automatically Evaluating Coherent Summaries via Context Chains

This paper introduces a new metric for automatically evaluation summaries called ContextChain. Based on an in-depth analysis of the TAC 2008 update summarization results, we show that previous automatic metrics such as ROUGE-2 and BE cannot reliably predict strong performing systems. We introduce two new terms called Correlation Recall and Correlation Precision and discuss how they cast more light on the coverage and the correctness of the respective metric. Our newly proposed metric called ContextChain incorporates findings from Giannakopoulos et al. (2008) and Barzilay and Lapata (2008) [2]. We show that our metric correlates with responsiveness scores even for the top n systems that participated in the TAC 2008 update summarization task, whereas ROUGE-2 and BE do not show a correlation for the top 25 systems.

Surrogate Learning - From Feature Independence to Semi-Supervised Classification

We consider the task of learning a classifier from the feature space X to the set of classes Y = {0, 1}, when the features can be partitioned into class-conditionally independent feature sets X1 and X2. We show that the class-conditional independence can be used to represent the original learning task in terms of 1) learning a classifier from X2 to X1 (in the sense of estimating the probability P(x1/x 2))and 2) learning the class-conditional distribution of the feature set X1. This fact can be exploited for semi-supervised learning because the former task can be accomplished purely from unlabeled samples. We present experimental evaluation of the idea in two real world applications.