Bayesian Optimization of Multi-Bit Pulse Encoding in In2O3/Al2O3 Thin-film Transistors for Temporal Data Processing

2025-10-08

AI Summary

In physical reservoir computing using In₂O₃/Al₂O₃ thin-film transistors, multibit pulse encoding suffers from low precision and poor separability of output states. Method: This work proposes a systematic operating-point optimization framework based on Bayesian optimization, using normalized degree of separation (nDoS) as the objective metric to jointly tune five-dimensional pulse parameters. It introduces, for the first time, a cross-bit migration strategy—leveraging a 4-bit trained model to guide 6-bit optimization—thereby substantially reducing experimental overhead. SHAP analysis is integrated to identify dominant factors and enhance model interpretability. Contributions/Results: The method achieves high-fidelity 6-bit temporal encoding in mobile vehicle image sequence encoding and reconstruction tasks. Crucially, the optimized 4-bit parameters attain performance nearly equivalent to that of the 6-bit configuration, demonstrating the approach’s effectiveness, generalizability, and engineering practicality.

Technology Category

• Search and Optimization: Learning to Search
• Machine Learning: Optimization
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Utilizing the intrinsic history-dependence and nonlinearity of hardware, physical reservoir computing is a promising neuromorphic approach to encode time-series data for in-sensor computing. The accuracy of this encoding critically depends on the distinguishability of multi-state outputs, which is often limited by suboptimal and empirically chosen reservoir operation conditions. In this work, we demonstrate a machine learning approach, Bayesian optimization, to improve the encoding fidelity of solution-processed Al2O3/In2O3 thin-film transistors (TFTs). We show high-fidelity 6-bit temporal encoding by exploring five key pulse parameters and using the normalized degree of separation (nDoS) as the metric of output state separability. Additionally, we show that a model trained on simpler 4-bit data can effectively guide optimization of more complex 6-bit encoding tasks, reducing experimental cost. Specifically, for the encoding and reconstruction of binary-patterned images of a moving car across 6 sequential frames, we demonstrate that the encoding is more accurate when operating the TFT using optimized pulse parameters and the 4-bit optimized operating condition performs almost as well as the 6-bit optimized condition. Finally, interpretability analysis via Shapley Additive Explanations (SHAP) reveals that gate pulse amplitude and drain voltage are the most influential parameters in achieving higher state separation. This work presents the first systematic method to identify optimal operating conditions for reservoir devices, and the approach can be extended to other physical reservoir implementations across different material platforms.

Problem

• Optimizing multi-bit pulse encoding in thin-film transistors for temporal data processing
• Improving encoding fidelity using Bayesian optimization and key pulse parameters
• Identifying optimal operating conditions to enhance output state separability

Innovation

• Bayesian optimization improves multi-bit pulse encoding fidelity
• Model trained on 4-bit data guides 6-bit encoding optimization
• SHAP analysis identifies gate amplitude and drain voltage importance