Technology Category
Cognitive Modeling & Cognitive Systems: Neural Spike CodingNatural Language Processing: Sentence-level Semantics, Textual Inference, etc.Machine Learning: Mixture of Experts (MoE)
SplashNet: Split-and-Share Encoders for Accurate and Efficient Typing with Surface Electromyography
🤖 AI Summary
To address three key challenges in wrist-based sEMG-driven keyboard-free text input—poor cross-user generalization, fragile reliance on high-order features, and absence of bilateral physiological priors—this paper proposes SplashNet-mini. The method innovatively encodes hand bilateral symmetry as a structured inductive bias and introduces a Split-and-Share dual-stream Transformer encoder. It integrates rolling temporal normalization, aggressive channel masking, and lightweight spectral dimensionality reduction (33 → 6 frequency bands). In zero-shot evaluation, character error rate drops to 36.4% (−31% relative improvement); after user-specific fine-tuning, it further decreases to 5.9% (−21%). With only one-quarter the parameters and 40% fewer FLOPs compared to baseline models, SplashNet-mini significantly enhances generalization, robustness, and computational efficiency.
Technology Category
Application Category
📝 Abstract
Surface electromyography (sEMG) at the wrists could enable natural, keyboard-free text entry, yet the state-of-the-art emg2qwerty baseline still misrecognizes $51.8%$ of characters in the zero-shot setting on unseen users and $7.0%$ after user-specific fine-tuning. We trace many of these errors to mismatched cross-user signal statistics, fragile reliance on high-order feature dependencies, and the absence of architectural inductive biases aligned with the bilateral nature of typing. To address these issues, we introduce three simple modifications: (i) Rolling Time Normalization, which adaptively aligns input distributions across users; (ii) Aggressive Channel Masking, which encourages reliance on low-order feature combinations more likely to generalize across users; and (iii) a Split-and-Share encoder that processes each hand independently with weight-shared streams to reflect the bilateral symmetry of the neuromuscular system. Combined with a five-fold reduction in spectral resolution ($33!
ightarrow!6$ frequency bands), these components yield a compact Split-and-Share model, SplashNet-mini, which uses only $ frac14$ the parameters and $0.6 imes$ the FLOPs of the baseline while reducing character-error rate (CER) to $36.4%$ zero-shot and $5.9%$ after fine-tuning. An upscaled variant, SplashNet ($ frac12$ the parameters, $1.15 imes$ the FLOPs of the baseline), further lowers error to $35.7%$ and $5.5%$, representing relative improvements of $31%$ and $21%$ in the zero-shot and fine-tuned settings, respectively. SplashNet therefore establishes a new state of the art without requiring additional data.
Problem
Research questions and friction points this paper is trying to address.
High character misrecognition in sEMG-based typing systems
Mismatched signal statistics across different users
Lack of architectural biases for bilateral typing nature
Innovation
Methods, ideas, or system contributions that make the work stand out.
Rolling Time Normalization aligns input distributions
Aggressive Channel Masking uses low-order features
Split-and-Share encoder processes hands independently
🔎 Similar Papers
No similar papers found.
