Abstract. In this paper we report on a series of experiments investigating the path from text-summarisation to style-specific summarisation of spoken news stories. We show that the portability of traditional text summarisation features to broadcast news is dependent on the diffusiveness of the information in the broadcast news story. An analysis of two categories of news stories (containing only read speech or some spontaneous speech) demonstrates the importance of the style and the quality of the transcript, when extracting the summary-worthy information content. Further experiments indicate the advantages of doing style-specific summarisation of broadcast news.

Chemistry text mining tools should be interoperable and adaptable regardless of system-level implementation, installation or even programming issues. We aim to abstract the functionality of these tools from the underlying implementation via reconfigurable workflows for automatically identifying chemical names. To achieve this, we refactored an established named entity recogniser (in the chemistry domain), OSCAR and studied the impact of each component on the net performance. We developed two reconfigurable workflows from OSCAR using an interoperable text mining framework, U-Compare. These workflows can be altered using the drag-&-drop mechanism of the graphical user interface of U-Compare. These workflows also provide a platform to study the relationship between text mining components such as tokenisation and named entity recognition (using maximum entropy Markov model (MEMM) and pattern recognition based classifiers). Results indicate that, for chemistry in particular, eliminating noise generated by tokenisation techniques lead to a slightly better performance than others, in terms of named entity recognition (NER) accuracy. Poor tokenisation translates into poorer input to the classifier components which in turn leads to an increase in Type I or Type II errors, thus, lowering the overall performance. On the Sciborg corpus, the workflow based system, which uses a new tokeniser whilst retaining the same MEMM component, increases the F-score from 82.35% to 84.44%. On the PubMed corpus, it recorded an F-score of 84.84% as against 84.23% by OSCAR.

Research on specialized biological systems is often hampered by a lack of consistent terminology, especially across species. In bacterial Type IV secretion systems genes within one set of orthologs may have over a dozen different names. Classifying research publications based on biological processes, cellular components, molecular functions, and microorganism species should improve the precision and recall of literature searches allowing researchers to keep up with the exponentially growing literature, through resources such as the Pathosystems Resource Integration Center (PATRIC, patricbrc.org). We developed named entity recognition (NER) tools for four entities related to Type IV secretion systems: 1) bacteria names, 2) biological processes, 3) molecular functions, and 4) cellular components. These four entities are important to pathogenesis and virulence research but have received less attention than other entities, e.g., genes and proteins. Based on an annotated corpus, large domain terminological resources, and machine learning techniques, we developed recognizers for these entities. High accuracy rates (>80%) are achieved for bacteria, biological processes, and molecular function. Contrastive experiments highlighted the effectiveness of alternate recognition strategies; results of term extraction on contrasting document sets demonstrated the utility of these classes for identifying T4SS-related documents.

In this paper we report on a series of experiments which compare the effect of individual features on both text and speech summarisation, the effect of basing the speech summaries on automatic speech recognition transcripts with varying word error rates, and the effect of summarisation approach and transcript source on summary quality. We show that classical text summarisation features (based on stylistic and content information) are portable to broadcast news. However, the quality of the speech transcripts as well as the difference in information structure between broadcast and newspaper news affect the usability of the individual features.

This paper presents an automatic system for structuring and preparing a news broadcast for applications such as speech summarization, browsing, archiving and information retrieval. This process comprises transcribing the audio using an automatic speech recognizer and subsequently segmenting the text into utterances and topics. A maximum entropy approach is used to build statistical models for both utterance and topic segmentation. The experimental work addresses the effect on performance of the topic boundary detector of three factors: the types of feature used, the quality of the ASR transcripts, and the quality of the utterance boundary detector. The results show that the topic segmentation is not affected severely by transcripts errors, whereas errors in the utterance segmentation are more devastating.