☆ 4.7 Article

On-Device Deep Multi-Task Inference via Multi-Task Zipping

IEEE TRANSACTIONS ON MOBILE COMPUTING (2023)

Related references

Note: Only part of the references are listed.

Article Computer Science, Information Systems

A Framework for Behavioral Biometric Authentication Using Deep Metric Learning on Mobile Devices

Cong Wang et al.

Summary: Mobile authentication using behavioral biometrics is an active research area. Current research relies on machine learning classifiers to recognize unique patterns but lacks discriminative power. Challenges in mobile device implementation include behavioral dynamics, data privacy, and side-channel leaks. To address these challenges, researchers propose a new framework for training on battery-powered mobile devices, ensuring data privacy and adaptability to behavioral patterns. They reformulate the classification problem as deep metric learning to improve discriminative power and design an effective countermeasure to prevent side-channel leaks. Experimental results show high authentication accuracy, robustness against attacks, and feasible training with mobile CPUs, with comparable energy consumption to watching videos and playing games.

IEEE TRANSACTIONS ON MOBILE COMPUTING (2023)

Add to Collection

Article Computer Science, Information Systems

MDLdroidLite: A Release-and-Inhibit Control Approach to Resource-Efficient Deep Neural Networks on Mobile Devices

Yu Zhang et al.

Summary: Mobile deep learning (MDL) is a privacy-preserving learning paradigm with unique features like continual learning and low-latency inference. This paper introduces MDLdroidLite, an on-device deep learning framework that optimizes model structures and parameter adaptation mechanisms to achieve resource efficiency and fast training.

IEEE TRANSACTIONS ON MOBILE COMPUTING (2022)

Add to Collection

Article Automation & Control Systems

Magnum: A Distributed Framework for Enabling Transfer Learning in B5G-Enabled Industrial IoT

Pallav Kumar Deb et al.

Summary: This article introduces a lightweight blockchain-inspired framework called Magnum, which stores transfer learning models in blocks. By storing these blocks on nearby fog nodes, industrial plants can easily access pretrained base models for tuning before deployment. Magnum is designed for B5G scenarios to reduce block transfer time, and implements a demand-centric distribution scheme to further reduce search and access time. Through experiments and comparisons, Magnum demonstrates accuracy retention of models with maximum CPU and memory usage of 80% and 6% respectively. Additionally, Magnum can write models as large as 17 Mb on blocks in 10 seconds and fetch them in 16 ms.

IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS (2021)

Add to Collection

Article Computer Science, Information Systems

AdaDeep: A Usage-Driven, Automated Deep Model Compression Framework for Enabling Ubiquitous Intelligent Mobiles

Sicong Liu et al.

Summary: AdaDeep is a usage-driven, automated DNN compression framework that systematically explores the desired trade-off between performance and resource constraints. By automatically selecting the most suitable combination of compression techniques and hyperparameters for a given DNN in a layer-wise manner, AdaDeep can achieve significant latency reduction, energy-efficiency improvement, and storage reduction while incurring negligible accuracy loss. Additionally, AdaDeep uncovers multiple novel combinations of compression techniques.

IEEE TRANSACTIONS ON MOBILE COMPUTING (2021)

Add to Collection

Proceedings Paper Computer Science, Hardware & Architecture

EdgeDuet: Tiling Small Object Detection for Edge Assisted Autonomous Mobile Vision

Xu Wang et al.

Summary: The paper introduces a new method for accurate and real-time object detection on resource-constrained devices, utilizing an edge-device collaborative framework to enhance small object detection accuracy and reduce latency. EdgeDuet outperforms local object detection as well as state-of-the-art offloading schemes in terms of detection accuracy and latency improvement.

IEEE CONFERENCE ON COMPUTER COMMUNICATIONS (IEEE INFOCOM 2021) (2021)

Add to Collection

Article Computer Science, Information Systems

Deep CNN-Based Real-Time Traffic Light Detector for Self-Driving Vehicles

Zhenchao Ouyang et al.

IEEE TRANSACTIONS ON MOBILE COMPUTING (2020)

Add to Collection

Article Engineering, Electrical & Electronic

Model Compression and Hardware Acceleration for Neural Networks: A Comprehensive Survey

Lei Deng et al.

PROCEEDINGS OF THE IEEE (2020)

Add to Collection

Proceedings Paper Computer Science, Hardware & Architecture

EagleEye: Wearable Camera-based Person Identification in Crowded Urban Spaces

Juheon Yi et al.

MOBICOM '20: PROCEEDINGS OF THE 26TH ANNUAL INTERNATIONAL CONFERENCE ON MOBILE COMPUTING AND NETWORKING (MOBICOM 2020) (2020)

Add to Collection

Proceedings Paper Computer Science, Information Systems

AMVP: Adaptive CNN-based Multitask Video Processing on Mobile Stream Processing Platforms

Mengyuan Chao et al.

2020 IEEE/ACM SYMPOSIUM ON EDGE COMPUTING (SEC 2020) (2020)

Add to Collection

Article Engineering, Electrical & Electronic

Edge Intelligence: Paving the Last Mile of Artificial Intelligence With Edge Computing

Zhi Zhou et al.

PROCEEDINGS OF THE IEEE (2019)

Add to Collection

Article Computer Science, Artificial Intelligence

Deep Expectation of Real and Apparent Age from a Single Image Without Facial Landmarks

Rasmus Rothe et al.

INTERNATIONAL JOURNAL OF COMPUTER VISION (2018)

Add to Collection

Review Multidisciplinary Sciences

An overview of multi-task learning

Yu Zhang et al.

NATIONAL SCIENCE REVIEW (2018)

Add to Collection

Proceedings Paper Computer Science, Theory & Methods

Multimodal Deep Convolutional Neural Network for Audio-Visual Emotion Recognition

Shiqing Zhang et al.

ICMR'16: PROCEEDINGS OF THE 2016 ACM INTERNATIONAL CONFERENCE ON MULTIMEDIA RETRIEVAL (2016)

Add to Collection

Article Computer Science, Artificial Intelligence

ImageNet Large Scale Visual Recognition Challenge

Olga Russakovsky et al.

INTERNATIONAL JOURNAL OF COMPUTER VISION (2015)

Add to Collection

Article Multidisciplinary Sciences

Human-level concept learning through probabilistic program induction

Brenden M. Lake et al.

SCIENCE (2015)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Describing Textures in the Wild

Mircea Cimpoi et al.

2014 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) (2014)

Add to Collection

Article Computer Science, Artificial Intelligence

Man vs. computer: Benchmarking machine learning algorithms for traffic sign recognition

J. Stallkamp et al.

NEURAL NETWORKS (2012)

Add to Collection

Correction Computer Science, Information Systems

Recognizing human emotional state from audiovisual signals (vol 10, pg 659, 2008)

Yongjin Wang et al.

IEEE TRANSACTIONS ON MULTIMEDIA (2008)

Add to Collection

Article Computer Science, Artificial Intelligence

An experimental study on pedestrian classification

S. Munder et al.

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE (2006)

Add to Collection

© Peeref 2019-2024. All rights reserved.