Continuous Temporal Representations of Event-Based Signals via Interference-Based Wave Modeling

2026-05-02

AI Summary

Traditional discrete or purely real-valued approaches struggle to effectively model the asynchronous yet structured temporal activation patterns inherent in event-driven biosignals such as surface electromyography (sEMG). This work proposes a continuous-time representation framework grounded in wave interference, mapping event signals into a complex-valued latent wave field. Temporal structure is encoded through phase modulation and interference among latent wave components, with the resulting representation projected into an energy domain to capture local dynamics and relational dependencies. Notably, the method eschews explicit recurrence or causal state propagation, introducing— for the first time—the wave interference mechanism to event-driven biosignal modeling to yield a continuous, differentiable, and structured representation. Experiments demonstrate that this approach significantly outperforms existing real-valued representations on sEMG data, achieving higher representational quality while maintaining computational efficiency, thereby proving well-suited for biomechanical control tasks such as prosthetics and exoskeletons.

Abstract

Spatio-temporal signals arising from event-driven biological processes, such as surface electromyography (sEMG), exhibit asynchronous and highly structured activation patterns that are challenging to model using conventional discrete or purely real-valued representations. In this work, we propose a continuous temporal modeling framework based on interference-based wave representations. The approach maps event-like input signals into a complex-valued latent wave field, where temporal structure is encoded through phase modulation and interactions between latent components. By projecting the resulting wave field onto an energy domain, the model induces structured activation patterns that capture both temporal localization and relational dependencies within finite observation windows, without relying on explicit recurrence or causal state propagation. The proposed formulation is particularly suited for event-driven biosignals, where continuous representations enable efficient gradient-based optimization and robust feature extraction. In particular, the method is designed to support learning from sEMG data for downstream control tasks in biomechanical systems, such as prosthetic devices and exoskeletons. Experimental results demonstrate that the proposed interference-based wave model provides improved representation quality compared to purely real-valued representations, while maintaining computational efficiency suitable for practical deployment.

Problem

• event-based signals
• temporal representation
• sEMG
• spatio-temporal modeling
• asynchronous activation

Innovation

• interference-based wave modeling
• continuous temporal representation
• event-based signals
• complex-valued latent field
• phase modulation