LearnAFE: Circuit-Algorithm Co-design Framework for Learnable Audio Analog Front-End

2025-07-01

AI Summary

Conventional audio analog front-ends (AFEs) and digital classifiers are typically designed in isolation, limiting overall system performance. To address this, this work proposes a circuit–algorithm co-optimization framework that jointly models a learnable bandpass filter (BPF) bank with a lightweight neural network classifier. Leveraging SNR-aware end-to-end training, we introduce a dedicated BPF co-loss function $mathcal{L}_{ ext{BPF}}$, enabling backpropagation-driven optimization of analog filter parameters. The hardware prototype is implemented in SKY130 130-nm CMOS technology. Under SNR conditions ranging from 5 to 20 dB, the system achieves 90.5%–94.2% accuracy on a 10-class audio classification task using only 22k parameters, while reducing power consumption and capacitor area by 8.7% and 12.9%, respectively. This work presents the first hardware-deployable, jointly optimized paradigm for learnable AFEs and digital classifiers tailored to audio classification.

Technology Category

• Machine Learning: Hardware-aware ML
• Natural Language Processing: Learning & Optimization for NLP
• Computer Vision: Learning & Optimization for CV

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

This paper presents a circuit-algorithm co-design framework for learnable analog front-end (AFE) in audio signal classification. Designing AFE and backend classifiers separately is a common practice but non-ideal, as shown in this paper. Instead, this paper proposes a joint optimization of the backend classifier with the AFE's transfer function to achieve system-level optimum. More specifically, the transfer function parameters of an analog bandpass filter (BPF) bank are tuned in a signal-to-noise ratio (SNR)-aware training loop for the classifier. Using a co-design loss function LBPF, this work shows superior optimization of both the filter bank and the classifier. Implemented in open-source SKY130 130nm CMOS process, the optimized design achieved 90.5%-94.2% accuracy for 10-keyword classification task across a wide range of input signal SNR from 5 dB to 20 dB, with only 22k classifier parameters. Compared to conventional approach, the proposed audio AFE achieves 8.7% and 12.9% reduction in power and capacitor area respectively.

Problem

• Joint optimization of AFE and classifier for audio signal classification
• SNR-aware tuning of analog bandpass filter bank parameters
• Reducing power and area in audio AFE design

Innovation

• Circuit-algorithm co-design for learnable AFE
• SNR-aware joint optimization of BPF and classifier
• Implemented in SKY130 CMOS with reduced power/area