We review current solutions and technical challenges for automatic speech recognition, keyword spotting, device arbitration, speech enhancement, and source localization in multidevice home environments to provide context for the INTERSPEECH 2022 special session, "Challenges and opportunities for signal processing and machine learning for multiple smart devices".
The first method, Simple CNAPS, employs a hierarchically regularized Mahalanobis-distance based classifier combined with a state of the art neural adaptive feature extractor to achieve strong performance on Meta-Dataset, mini-ImageNet and tiered-ImageNet benchmarks.
We propose a transductive meta-learning method that uses unlabelled instances to improve few-shot image classification performance.
We develop a transductive meta-learning method that uses unlabelled instances to improve few-shot image classification performance.
Ranked #1 on Few-Shot Image Classification on Tiered ImageNet 10-way (1-shot) (using extra training data)
Gaussian processes are powerful models for probabilistic machine learning, but are limited in application by their $O(N^3)$ inference complexity.