Towards Automatic Transcription of ILSE ― an Interdisciplinary Longitudinal Study of Adult Development and Aging
16688Authors
nan
Abstract
AbstractThe Interdisciplinary Longitudinal Study on Adult Development and Aging (ILSE) was created to facilitate the study of challenges posed by rapidly aging societies in developed countries such as Germany. ILSE contains over 8,000 hours of biographic interviews recorded from more than 1,000 participants over the course of 20 years. Investigations on various aspects of aging, such as cognitive decline, often rely on the analysis of linguistic features which can be derived from spoken content like these interviews. However, transcribing speech is a time and cost consuming manual process and so far only 380 hours of ILSE interviews have been transcribed. Thus, it is the aim of our work to establish technical systems to fully automatically transcribe the ILSE interview data. The joint occurrence of poor recording quality, long audio segments, erroneous transcriptions, varying speaking styles & crosstalk, and emotional & dialectal speech in these interviews presents challenges for automatic speech recognition (ASR). We describe our ongoing work towards the fully automatic transcription of all ILSE interviews and the steps we implemented in preparing the transcriptions to meet the interviews’ challenges. Using a recursive long audio alignment procedure 96 hours of the transcribed data have been made accessible for ASR training.
Keywords
Full Content
Towards Automatic Transcription of ILSE – an Interdisciplinary Longitudinal Study of Adult Development and Aging Jochen Weiner1, Claudia Frankenberg2, Dominic Telaar1, Britta Wendelstein2,3, Johannes Schr¨oder2,3, Tanja Schultz1 1Cognitive Systems Lab, University of Bremen (formerly Karlsruhe Institute of Technology), Germany 2Section of Geriatric Psychiatry, University Hospital Heidelberg, Germany 3Institute of Gerontology, Universit¨at Heidelberg, Germany jochen.weiner@uni-bremen.de Abstract The Interdisciplinary Longitudinal Study on Adult Development and Aging (ILSE) was created to facilitate the study of challenges posed by rapidly aging societies in developed countries such as Germany. ILSE contains over 8,000 hours of biographic interviews recorded from more than 1,000 participants over the course of 20 years. Investigations on various aspects of aging, such as cognitive decline, often rely on the analysis of linguistic features which can be derived from spoken content like these interviews. However, transcribing speech is a time and cost consuming manual process and so far only 380 hours of ILSE interviews have been transcribed. Thus, it is the aim of our work to establish technical systems to fully automatically transcribe the ILSE interview data. The joint occurrence of poor recording quality, long audio segments, erroneous transcriptions, varying speaking styles & crosstalk, and emotional & dialectal speech in these interviews presents challenges for automatic speech recognition (ASR). We describe our ongoing work towards the fully automatic transcription of all ILSE interviews and the steps we implemented in preparing the transcriptions to meet the interviews’ challenges. Using a recursive long audio alignment procedure 96 hours of the transcribed data have been made accessible for ASR training. Keywords: ILSE, ASR corpus, long audio alignment, dementia 1. Introduction The population in developed countries is aging rapidly. In Germany, for example, the most populous age group in the year 1950 were ten-year-olds and in 2000 forty-year-olds; for 2050 it is estimated to be the sixty-year-olds (Statis- tische ¨Amter des Bundes und der L¨ander, 2011). This de- mographic change is caused by and causes transitions in family structures, way of life, and work force and thus bears tremendous challenges for the society. To tackle these chal- lenges and mitigate the consequences, researchers of vari- ous disciplines such as physicians, psychologists, geron- tologists, sociologists, linguists, and engineers are joining forces in various ongoing interdisciplinary projects and ini- tiatives. One of these projects is the Interdisciplinary Lon- gitudinal Study on Adult Development and Aging (ILSE) (Martin and Martin, 2000; Lehr et al., 2000; Sch¨onknecht et al., 2005), a cohort-based longitudinal study similar to the Health and Retirement Study (Juster and Suzman, 1995) and the Victoria Longitudinal Study (Dixon and de Frias, 2004) in the USA, and PAQUID (Dartigues et al., 1992) in France. Within ILSE, over a thousand subjects participated in up to four biographic interviews over a time span of more than 20 years, resulting in over 8,000 hours of recorded speech. These interviews are complemented by data from psychological, cognitive, physical and dental examinations. Among other indicators, it includes for each participant a diagnosis about the existence and level of cognitive de- cline. In order …
Content truncated to 500 words
Labeled Text
AbstractThe Interdisciplinary Longitudinal Study on Adult Development and Aging (ILSE) was created to facilitate the study of challenges posed by rapidly aging societies in developed countries such as Germany. ILSE contains over 8,000 hours of biographic interviews recorded from more than 1,000 participants over the course of 20 years. Investigations on various aspects of aging, such as cognitive decline, often rely on the analysis of linguistic features which can be derived from spoken content like these interviews. However, transcribing speech is a time and cost consuming manual process and so far only 380 hours of ILSE interviews have been transcribed. Thus, it is the aim of our work to establish technical systems to fully automatically transcribe the ILSE interview data. The joint occurrence of poor recording quality, long audio segments, erroneous transcriptions, varying speaking styles & crosstalk, and emotional & dialectal speech in these interviews presents challenges for automatic speech recognition (ASR). We describe our ongoing work towards the fully automatic transcription of all ILSE interviews and the steps we implemented in preparing the transcriptions to meet the interviews’ challenges. Using a recursive long audio alignment procedure 96 hours of the transcribed data have been made accessible for ASR training.
Publication Details
- Paper ID:
- 16688
- Year:
- 2016
- Publication:
- TIPSTER
- Publication Type:
- conf
- DOI:
- nan