HEiMDaL: Highly Efficient Method for Detection and Localization of wake-words

TL;DR

HEiMDaL is a low-footprint CNN model designed for efficient streaming wake-word detection and localization, outperforming traditional DNN-HMM methods in detection metrics while maintaining similar localization accuracy and memory footprint.

Contribution

The paper introduces HEiMDaL, a novel CNN-based approach with alignment and offset losses for improved wake-word detection and localization in streaming conditions.

Findings

73% reduction in detection metrics

Maintains localization accuracy

Same memory footprint as existing models

Abstract

Streaming keyword spotting is a widely used solution for activating voice assistants. Deep Neural Networks with Hidden Markov Model (DNN-HMM) based methods have proven to be efficient and widely adopted in this space, primarily because of the ability to detect and identify the start and end of the wake-up word at low compute cost. However, such hybrid systems suffer from loss metric mismatch when the DNN and HMM are trained independently. Sequence discriminative training cannot fully mitigate the loss-metric mismatch due to the inherent Markovian style of the operation. We propose an low footprint CNN model, called HEiMDaL, to detect and localize keywords in streaming conditions. We introduce an alignment-based classification loss to detect the occurrence of the keyword along with an offset loss to predict the start of the keyword. HEiMDaL shows 73% reduction in detection metrics along with equivalent localization accuracy and with the same memory footprint as existing DNN-HMM style models for a given wake-word.