No abstract

This paper presents a conditional random field-based approach for identifying speaker-produced disfluencies (i.e. if and where they occur) in spontaneous speech transcripts. We emphasize false start regions, which are often missed in current disfluency identification approaches as they lack lexical or structural similarity to the speech immediately following. We find that combining lexical, syntactic, and language model-related features with the output of a state-of-the-art disfluency identification system improves overall word-level identification of these and other errors. Improvements are reinforced under a stricter evaluation metric requiring exact matches between cleaned sentences annotator-produced reconstructions, and altogether show promise for general reconstruction efforts.

This paper presents an approach to partial parse selection for robust deep processing. The work is based on a bottom-up chart parser for HPSG parsing. Following the definition of partial parses in (Kasper et al., 1999), different partial parse selection methods are presented and evaluated on the basis of multiple metrics, from both the syntactic and semantic viewpoints. The application of the partial parsing in spontaneous speech texts processing shows promising competence of the method.

While speaking spontaneously, speakers often make errors such as self-correction or false starts which interfere with the successful application of natural language processing techniques like summarization and machine translation to this data. There is active work on reconstructing this errorful data into a clean and fluent transcript by identifying and removing these simple errors. Previous research has approximated the potential benefit of conducting word-level reconstruction of simple errors only on those sentences known to have errors. In this work, we explore new approaches for automatically identifying speaker construction errors on the utterance level, and quantify the impact that this initial step has on word-and sentence-level reconstruction accuracy.

Spontaneously produced speech text often includes disfluencies which make it difficult to analyze underlying structure. Successful reconstruction of this text would transform these errorful utterances into fluent strings and offer an alternate mechanism for analysis.Our investigation of naturally-occurring spontaneous speaker errors aligned to corrected text with manual semanticosyntactic analysis yields new insight into the syntactic and structural semantic differences between spoken and reconstructed language.