This paper reports on the performance of a new system for recognizing textual entailment. We present a brief overview of the CERES system and summarize its semantic alignment method for determining entailment. Then, we describe the experiments that were performed, both official and subsequently, against the RTE4 test dataset. We also report on comparison runs against the previous RTE3 developme...

this study was motivated by three factors, which also contribute to its significance for today’s academic writing. first, research articles are the significant means of communication between the writers all over the world. second, persuasion and organization are crucial notions in academic writing where the authors have to consider the academic audiences and their needs.  third, some writers ar...

This paper presents a robust textual entailment system using the principle of just noticeable difference in psychology, which we call a local graph matching-based system with active learning. First, although an early textual entailment task often involved two rather simple sentences of T (“Text”) and H (“Hypothesis”), the recent textual entailment task often involves multiple / complex / compou...

Textual process descriptions are widely used in organizations since they can be created and understood by virtually everyone. The inherent ambiguity of natural language, however, impedes the automated analysis of textual process descriptions. While human readers can use their context knowledge to correctly understand statements with multiple possible interpretations, automated analysis techniqu...

Web Information Extraction (WIE) is the task of automatically extracting knowledge from Web content. On-demand WIE systems such as KNOWITNOW [1] and TEXTRUNNER [2] allow users to query the Web for a textual context that indicates a desired relation. For example, the context “$x invented $y” indicates the Invented(x, y) relation. The WIE system is tasked with responding in real-time with a list ...

Overview: During Phase I of the TIPSTER program, HNC developed a unique approach to machine learning of similarity of meaning. This approach, embodied in a system called "MatchPlus", exploits this learned similarity of meaning for concept-based text retrieval, routing and visualization of textual information. MatchPlus uses an information representation scheme called "context vectors" to encode...

In this paper we investigate the relation between positive and negative pairs in Textual Entailment (TE), in order to highlight the role of contradiction in TE datasets. We base our analysis on the decomposition of Text-Hypothesis pairs into monothematic pairs, i.e. pairs where only one linguistic phenomenon at a time is responsible for entailment judgment and we argue that such a deeper inspec...

Context-specific document recommender systems rely on the accurate identification of context descriptors from unstructured textual information to identify highly relevant documents. In this paper, we propose two term-weighting measures, “normal distance” and “adjusted inverse polysemy”, to enable the retrieval of relevant documents with higher precision. We analyze the performance of the propos...