Deep Learning Algorithms Enabling Event Detection: A Review
Cherifa Nakkach, Amira Zrelli, Tahar Ezzeddine · Advances in intelligent systems research/Advances in Intelligent Systems Research · 2022
Deep Learning has revolutionized computer vision, natural language processing, speech recognition, and information retrieval.However, as deep learning models developed, their parameter count, latency, and resource requirements rose.As a result, a model's footprint, as well as quality, should always be addressed.Academics and industry have identified neural network-based deep learning as a possible research subject.Deep learning algorithms have had tremendous results.This paper will review neural networks' deep learning methods for auditory event detection.For this reason, this paper aims to examine both highly and weakly labelled acoustic event detection systems based on deep learning.This article also discusses how deep learning might help detect events and the challenges in upcoming real-world scenarios.We briefly define the issue of model efficiency in deep learning, then cover the foundational work in the five core areas of model efficiency (modelling approaches, infrastructure, and hardware).