Chaos based activations and attention improve EMG to speech synthesis
Lyapunov-Inspired LyRIC Activation and GLARE Attention in Chaos-Guided State Space Modeling for EMG-To-Speech (ETS) Synthesis
Sound
Summary
Converting muscle signals to speech is a complex and chaotic process that previous methods struggled to model accurately. The authors introduce new techniques inspired by chaos theory to better capture the unpredictable nature of human speech production. Their approach uses special functions and attention mechanisms to improve speech clarity with fewer model parameters. Tests show the new system produces much clearer speech than older methods and runs more efficiently. This is the first time chaotic behavior has been explicitly used to improve this kind of speech synthesis.
What this means in practice
- •For assistive technology developers: Develop real-time EMG-to-speech devices with improved clarity and smaller model size for people with speech impairments.$Commercial implications: This paper enables more accurate, efficient speech synthesis hardware for disabled users through chaos-inspired models reducing size and improving intelligibility.
- •For mobile device engineers: Integrate efficient ETS synthesis to add hands-free speech communication for wearable or embedded computing devices.
Authors
Sajid Fardin Dipto, Tarikul Islam Tamiti, Luke Baja-Ricketts, David Vergano, Anomadarshi Barua
Abstract
Electromyography-to-Speech (ETS) synthesis is typically a non-linear, chaotic dynamical system. However, no prior work has studied the chaotic behavior of ETS synthesis to date. Yet, prior works strictly rely on standard reconstruction metrics with parameter-heavy transformers that systematically over-smooth natural acoustic dynamics. To close this gap, for the first time, we propose a chaos-inspired Lyapunov-derived activation function (LyRIC) with two novel chaotic loss functions, Lyapunov Exponent Regularization and Multi-Scale Detrended Fluctuation Analysis, to explicitly capture the deterministic chaos of human phonation. In addition, we introduce a compressed novel encoder, GLAME, which synergizes global Mamba state-space modeling with localized GLARE attention. We comprehensively perform frame-level acoustic evaluation in a multilingual and multi-speaker setup using English and Mandarin datasets. The proposed system outperforms the established baseline with a 4.69x increase in objective intelligibility (STOI: 0.61 vs. 0.13) and a 2.08x improvement in spectral reconstruction (LSD: 1.08 vs. 2.25). Importantly, this improvement is achieved with 73.49% fewer parameters (14.34M vs. 54.10M), establishing a new baseline for ETS synthesis. To the best of our knowledge, this is the first work demonstrating that integrating non-linear chaotic physics into neural networks yields superior yet compact inductive biases for real-time ETS synthesis.