eess.ASOct 6, 2026

A Novel Sentence Stress Detection Framework Leveraging Auxiliary Word-Stress Modeling and Loss Optimization

Authors: Tien-Hong Lo, Fong-Chun Tsai, Ting-An Hung, Yu-Hsuan Hsieh, Yao-Ting Sung, Berlin Chen

Organizations: National Taiwan Normal University, Taiwan

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Prosodic stress is a crucial aspect of automatic pronunciation assessment (APA), encompassing both sentence stress detection (SSD) and word stress detection (WSD). SSD highlights semantically salient words that shape discourse meaning, while WSD identifies the primary stressed syllable within each word to ensure lexical clarity. However, most prior work treats SSD and WSD as independent tasks, overlooking their shared reliance on prosodic cues such as pitch, duration, and intensity. To address this gap, we propose an effective SSD approach combining SSD with auxiliary WSD via a novel modeling paradigm. In addition, we introduce a word-span stress regularizer (WSR) that concentrates token-level SSD probabilities within each stressed word span. Experiments on the TinyStress-15K benchmark show that the proposed method outperforms strong baselines, with the complete configuration achieving the best SSD result.

Figures & tables

Figure 1: Illustration of stress detection at two linguistic levels. (a) SSD highlights the emphasized word within a sentence by shifting prominence across different lexical items. (b) WSD identifies the primary stressed syllable within a lexical item, where primary stress placement is indicated by the digit “1”.
Figure 2: Overall architecture of the proposed framework. A frozen Whisper backbone provides encoder representations to both heads and decoder representations to the SSD head. Sentence stress detection (SSD) operates at the token level with an additional word-span stress regularizer (WSR), while word stress detection (WSD) operates at the phone level. The SSD and WSD heads are task-specific and do not directly exchange hidden states or predictions.
 #Audios#Tokens#Stress#Unstress
Train13,500171,20622,537148,669
Valid1,50018,6762,54116,135
Test1,00012,5481,70510,843
Table 1: Statistics of the TinyStress-15K corpus.
ModelPrecisionRecallF1
GT alignment [ 14 ]0.8620.8530.858
MFA [ 14 ]0.7760.8590.815
WhiStress [ 14 ]0.9120.9060.909
STRAW0.9450.9240.934
- WSD0.9380.9060.922
- WSR0.9420.9170.929
Table 2: SSD performance on TinyStress-15K. STRAW denotes our proposed framework with auxiliary WSD and the WSR. Ablations remove the corresponding components.
ModelPrecisionRecallF1
STRAW0.9240.9160.920
- WSR0.9350.9070.921
Table 3: WSD performance of STRAW on TinyStress-15K.
Figure 3: Error analysis by part-of-speech (POS) on the test set. Bars show false negative rate (FNR) and false positive rate (FPR). The black line indicates the sample count per POS.

Explore similar work

CardsList

  1. Evaluating and Preserving Lexical Stress in English-to-Chinese Speech-to-Speech Translation

    Jun 13, 2026Yuchen Song, Xi Chen, Mingze Li +1Speech TranslationTranslation Quality

    The Chinese University of Hong Kong, Shenzhen, China · Shenzhen Loop Area Institute, China

  2. Prosodic ABX: A Language-Agnostic Method for Measuring Prosodic Contrast in Speech Representations

    Apr 2, 2026Haitong Sun, Stephen McIntosh, Kwanghee Choi +3ProsodySelf-Supervised Speech Models

    The University of Tokyo, Japan · University of Texas at Austin, USA

  3. Automatic Detection of Stress from Speech in the Trier Social Stress Test

    Jul 1, 2026Hanna Drimalla, Wieland R. Cremer, Christine Kraus +1StressAutomatic Identification System

    Human-Centered Artificial Intelligence Group, Faculty of Technology, Bielefeld University, Bielefeld, Germany. · Department of Cognitive Psychology, Faculty of Psychology, Ruhr University Bochum, Bochum, Germany.