Supervised Speech Representation Learning for Parkinson's Disease Classification

Abstract

Recently proposed automatic pathological speech classification techniques use unsupervised auto-encoders to obtain a high-level abstract representation of speech. Since these representations are learned based on reconstructing the input, there is no guarantee that they are robust to pathology-unrelated cues such as speaker identity information. Further, these representations are not necessarily discriminative for pathology detection. In this paper, we exploit supervised auto-encoders to extract robust and discriminative speech representations for Parkinson's disease classification. To reduce the influence of speaker variabilities unrelated to pathology, we propose to obtain speaker identity-invariant representations by adversarial training of an auto-encoder and a speaker identification task. To obtain a discriminative representation, we propose to jointly train an auto-encoder and a pathological speech classifier. Experimental results on a Spanish database show that the proposed supervised representation learning methods yield more robust and discriminative representations for automatically classifying Parkinson's disease speech, outperforming the baseline unsupervised representation learning system.

Interpretable Speech Features vs. DNN Embeddings: What to Use in the Automatic Assessment of Parkinson’s Disease in Multi-lingual Scenarios

“…Namely, the performances of our best IFMs on CzechPD and ItalianPVS outperformed those reported by Kovac et al [41,42] based on acoustic descriptors by more than 10 percentual absolute points in ACC. Similarly, TRILLsson reported best AUCs of 0.95, 0.87, 0.84 on Neurovoz, GITA, and GermanPD, which were 0.01, 0.03, and 0.16 higher than the x-vector approach of [31], the CNN approach of [28], and the mono-lingual CNN approach of [27], respectively. In this regard, HuBERT/Wav2Vec 2.0. reported best AUCs of 0.96, 0.90, and 0.88 on Neurovoz, GITA, and GermanPD, respectively, which were even higher than those achieved with TRILLsson.…”