Adaptive inference through early-exit networks: Design, challenges and directions

S Laskaridis, A Kouris, ND Lane - … of the 5th International Workshop on …, 2021 - dl.acm.org
DNNs are becoming less and less over-parametrised due to recent advances in efficient
model design, through careful hand-crafted or NAS-based methods. Relying on the fact that …

Why should we add early exits to neural networks?

S Scardapane, M Scarpiniti, E Baccarelli… - Cognitive Computation, 2020 - Springer
Deep neural networks are generally designed as a stack of differentiable layers, in which a
prediction is obtained only after running the full stack. Recently, some contributions have …

Green edge AI: A contemporary survey

Y Mao, X Yu, K Huang, YJA Zhang… - Proceedings of the …, 2024 - ieeexplore.ieee.org
Artificial intelligence (AI) technologies have emerged as pivotal enablers across a multitude
of industries, including consumer electronics, healthcare, and manufacturing, largely due to …

Branchynet: Fast inference via early exiting from deep neural networks

S Teerapittayanon, B McDanel… - 2016 23rd international …, 2016 - ieeexplore.ieee.org
Deep neural networks are state of the art methods for many learning tasks due to their ability
to extract increasingly better features at each network layer. However, the improved …

Revisiting batch normalization for training low-latency deep spiking neural networks from scratch

Y Kim, P Panda - Frontiers in neuroscience, 2021 - frontiersin.org
Spiking Neural Networks (SNNs) have recently emerged as an alternative to deep learning
owing to sparse, asynchronous and binary event (or spike) driven processing, that can yield …

Energy-efficient edge based real-time healthcare support system

S Abirami, P Chitra - Advances in computers, 2020 - Elsevier
The ubiquitous usage of wearable IoT (wIoT) devices has created a formidable opportunity
for remote health monitoring system to provide paramount services such as preventive care …

Edge-AI-driven framework with efficient mobile network design for facial expression recognition

Y Wu, L Zhang, Z Gu, H Lu, S Wan - ACM Transactions on Embedded …, 2023 - dl.acm.org
Facial Expression Recognition (FER) in the wild poses significant challenges due to realistic
occlusions, illumination, scale, and head pose variations of the facial images. In this article …

Learning to weight samples for dynamic early-exiting networks

Y Han, Y Pu, Z Lai, C Wang, S Song, J Cao… - European conference on …, 2022 - Springer
Early exiting is an effective paradigm for improving the inference efficiency of deep networks.
By constructing classifiers with varying resource demands (the exits), such networks allow …

Dynamic convolutions: Exploiting spatial sparsity for faster inference

T Verelst, T Tuytelaars - … of the ieee/cvf conference on …, 2020 - openaccess.thecvf.com
Modern convolutional neural networks apply the same operations on every pixel in an
image. However, not all image regions are equally important. To address this inefficiency …

Seenn: Towards temporal spiking early exit neural networks

Y Li, T Geller, Y Kim, P Panda - Advances in Neural …, 2024 - proceedings.neurips.cc
Abstract Spiking Neural Networks (SNNs) have recently become more popular as a
biologically plausible substitute for traditional Artificial Neural Networks (ANNs). SNNs are …