Search
Search Results
-
Rethinking Online Knowledge Distillation with Multi-exits
Online knowledge distillation is a method to train multiple networks simultaneously by distilling the knowledge among each other from scratch. An... -
Rapid Fire Detection with Early Exiting
Efficient and effective fire detection has proven critical and if not achieved it can pose significant ecological and economic challenges. By... -
MMExit: Enabling Fast and Efficient Multi-modal DNN Inference with Adaptive Network Exits
Multi-modal DNNs have been demonstrated to outperform the best uni-modal DNNs by fusing information from different modalities. However, the... -
Why Should We Add Early Exits to Neural Networks?
Deep neural networks are generally designed as a stack of differentiable layers, in which a prediction is obtained only after running the full stack....
-
Learning to Weight Samples for Dynamic Early-Exiting Networks
Early exiting is an effective paradigm for improving the inference efficiency of deep networks. By constructing classifiers with varying resource... -
Multi-exit self-distillation with appropriate teachers
Multi-exit architecture allows early-stop inference to reduce computational cost, which can be used in resource-constrained circumstances. Recent...
-
FastDet: Detecting Encrypted Malicious Traffic Faster via Early Exit
Encrypted malicious traffic detection, which aims to identify encrypted malicious traffic from vast amounts of network traffic, is critical to... -
AdaInNet: an adaptive inference engine for distributed deep neural networks offloading in IoT-FOG applications based on reinforcement learning
The increasing expansion of Internet-of-Things (IoT) in the world requires Big Data analytic infrastructures to produce valuable knowledge in IoT...
-
Dynamics in Entry and Exit Registrations in a 14-Year Follow-Up of Nationwide Electronic Prescription and Patient Data Repository Services in Finland
There exist a need to carry out further research in order to describe implementation and adoption of nationwide healthcare information systems. This... -
Entropy-Based Early-Exit in a FPGA-Based Low-Precision Neural Network
In this paper, we investigate the application of early-exit strategies to fully quantized neural networks, mapped to low-complexity FPGA SoC devices.... -
Map** Gamification Elements to Heuristics and Behavior Change in Early Phase Inclusive Design: A Case Study
This paper identifies the connection between heuristics, gamification, and system visualization in the early design phase of a mobile application in... -
EB-FedAvg: Personalized and Training Efficient Federated Learning with Early-Bird Tickets
Federated learning is a well-known way to improve privacy in distributed machine learning. Its major goal is to learn a global model that provides... -
I/O Efficient Early Bursting Cohesive Subgraph Discovery in Massive Temporal Networks
Temporal networks are an effective way to encode temporal information into graph data losslessly. Finding the bursting cohesive subgraph (BCS), which...
-
Distillation-Based Multi-exit Fully Convolutional Network for Visual Tracking
Obtaining a trade-off between accuracy and efficiency for a convolutional neural network is highly desired in the deep classification-based trackers.... -
Meta-GF: Training Dynamic-Depth Neural Networks Harmoniously
Most state-of-the-art deep neural networks use static inference graphs, which makes it impossible for such networks to dynamically adjust the depth... -
Modelling and simulation of assisted hospital evacuation using fuzzy-reinforcement learning based modelling approach
Available hospital evacuation simulation models usually focus on the movement of the evacuees while ignoring the crucial behavioural factors of the...
-
Training a Lightweight ViT Network for Image Retrieval
Recently, Vision Transformer (ViT) networks have achieved promising advancements on many computer vision tasks. However, a ViT network has a large... -
Multi-Exit Semantic Segmentation Networks
Semantic segmentation arises as the backbone of many vision systems, spanning from self-driving cars and robot navigation to augmented reality and... -
FastNER: Speeding up Inferences for Named Entity Recognition Tasks
BERT and its variants are the most performing models for named entity recognition (NER), a fundamental information extraction task. We must apply... -
Privacy and safety improvement of VANET data via a safety-related privacy scheme
Vehicular Ad-hoc NETwork (VANET) safety applications allow vehicles to exchange messages with surrounding vehicles periodically to improve the...