Meta-AF: Meta-Learning for Adaptive Filters

DSP Seminar - November 18, 2022. CCRMA, Stanford
Abstract: Adaptive filtering algorithms are pervasive throughout modern society and have had a significant impact on a wide variety of domains including audio processing, telecommunications, biomedical sensing, astrophysics and cosmology, seismology, and many more. Adaptive filters typically operate via specialized online, iterative optimization methods such as least-mean squares or recursive least squares and aim to process signals in unknown or nonstationary environments. Such algorithms, however, can be slow and laborious to develop, require domain expertise to create, and necessitate mathematical insight for improvement. In this work, we seek to go beyond the limits of human-derived adaptive filter algorithms and present a comprehensive framework for learning online, adaptive signal processing algorithms or update rules directly from data. To do so, we frame the development of adaptive filters as a meta-learning problem in the context of deep learning and use a form of self-supervision to learn online iterative update rules for adaptive filters. To demonstrate our approach, we focus on audio applications and systematically develop meta-learned adaptive filters for five canonical audio problems including system identification, acoustic echo cancellation, blind equalization, multi-channel dereverberation, and beamforming. For each application, we compare against common baselines and/or current state-of-the-art methods and show we can learn high-performing adaptive filters that operate in real-time and, in most cases, significantly outperform all past specially developed methods for each task using a single general-purpose configuration of our method.
ArXiv draft: arxiv.org/abs/...
Demo: • MetaAF: Meta-learning ...
Code: github.com/ado...
Bio: Jonah Casebeer is a 4th year PhD candidate advised by Paris Smaragdis in the Computer Science department at the University of Illinois at Urbana-Champaign (UIUC). His area of expertise is machine learning for audio signal processing where he focuses on leveraging digital signal processing tools for deep learning. He completed his bachelor’s degrees in Computer Science and Statistics at UIUC where he was selected as a finalist for the Computing Research Association’s 2019 Outstanding Undergraduate Researcher Award. He has funded his PhD through the UIUC Computer Science Excellence Fellowship, the UIUC Machine Learning Excellence Fellowship, and industry collaborations. His work has been published at conferences including ICASSP and WASPAA, and he has interned with research groups at IBM, MIT Lincoln Labs, Amazon, Meta, and Adobe.
ccrma.stanford...