On the importance of normative data in speech-based assessment

By Zeinab Noorian, Chloé Pou-Prom, Frank Rudzicz. 2017

27 citations recorded by the source corpus.

Data sets for identifying Alzheimer's disease (AD) are often relatively sparse, which limits their ability to train generalizable models. Here, we augment such a data set, DementiaBank, with each of two normative data sets, the Wisconsin Longitudinal Study and Talk2Me, each of which employs a speech-based picture-description assessment. Through minority class oversampling with ADASYN, we outperform state-of-the-art results in binary classification of people with and without AD in DementiaBank. This work highlights the effectiveness of combining sparse and difficult-to-acquire patient data with relatively large and easily accessible normative datasets.

Subfields: Artificial Intelligence, Experimental and Cognitive Psychology, Speech & Audio

Citations in the corpus

• BigSSL: Exploring the Frontier of Large-Scale Semi-Supervised Learning for Automatic Speech Recognition (2021, neutral): “Our results are worse than SoTA on DementiaBank, which has been achieved by using additional textual information along with the acoustic features [98].”
• Towards Learning a Universal Non-Semantic Representation of Speech (2020, unknown): “erage performance of the different embeddings on a number of downstream tasks. VoxCeleb1 VoxForge SpeechCommands CREMA-D SAVEE DementiaBank Prev SOTA 80.5[6] 89.0[25] 91.1 [26] 74.0 [27] 61.0[10] 80.6[28] Mel / MFCC 15.5 79.2 47.7 52.8 55.6 67.7 OpenSmile 2.3 78.0 36.5 53.7 62.6 68.8 Random Network ”