2020/09/07 by Gizem Soğancıoğlu, Oxana Verkholyak, Soğancıoğlu, Gizem +11 · 2 citations
Computer Science · Medicine · Psychology · #Computation and Language (cs.CL) #Emotion and Mood Recognition #FOS: Computer and information sciences #Human-Computer Interaction (cs.HC) #Machine Learning (cs.LG) #Speech Recognition and Synthesis #Voice and Speech Disorders
paper · pdf · doi:10.48550/arxiv.2009.03432
openalex publication_date 2020/09/07 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Acoustic and linguistic analysis for elderly emotion recognition is an\nunder-studied and challenging research direction, but essential for the\ncreation of digital assistants for the elderly, as well as unobtrusive\ntelemonitoring of elderly in their residences for mental healthcare purposes.\nThis paper presents our contribution to the INTERSPEECH 2020 Computational\nParalinguistics Challenge (ComParE) - Elderly Emotion Sub-Challenge, which is\ncomprised of two ternary classification tasks for arousal and valence\nrecognition. We propose a bi-modal framework, where these tasks are modeled\nusing state-of-the-art acoustic and linguistic features, respectively. In this\nstudy, we demonstrate that exploiting task-specific dictionaries and resources\ncan boost the performance of linguistic models, when the amount of labeled data\nis small. Observing a high mismatch between development and test set\nperformances of various models, we also propose alternative training and\ndecision fusion strategies to better estimate and improve the generalization\nperformance.\n