scholarly journals Emotion Recognition on Edge Devices: Training and Deployment

Sensors ◽  
2021 ◽  
Vol 21 (13) ◽  
pp. 4496
Author(s):  
Vlad Pandelea ◽  
Edoardo Ragusa ◽  
Tommaso Apicella ◽  
Paolo Gastaldo ◽  
Erik Cambria
Keyword(s):  
Emotion Recognition ◽  
Language Processing ◽  
Computational Cost ◽  
Sequential Learning ◽  
High Quality ◽  
Fast Training ◽  
Online Sequential Learning ◽  
And Performance ◽  
Resource Constrained Devices ◽  
Constrained Devices

Emotion recognition, among other natural language processing tasks, has greatly benefited from the use of large transformer models. Deploying these models on resource-constrained devices, however, is a major challenge due to their computational cost. In this paper, we show that the combination of large transformers, as high-quality feature extractors, and simple hardware-friendly classifiers based on linear separators can achieve competitive performance while allowing real-time inference and fast training. Various solutions including batch and Online Sequential Learning are analyzed. Additionally, our experiments show that latency and performance can be further improved via dimensionality reduction and pre-training, respectively. The resulting system is implemented on two types of edge device, namely an edge accelerator and two smartphones.

scholarly journals R-OO-KASE: Revocable Online/Offline Key Aggregate Searchable Encryption

2020 ◽  
Vol 5 (4) ◽  
pp. 391-418
Author(s):  
Mukti Padhya ◽  
Devesh C. Jinwala
Keyword(s):  
Radio Frequency Identification ◽  
Keyword Search ◽  
Electrical Power ◽  
Computational Cost ◽  
Security Analysis ◽  
Searchable Encryption ◽  
Resource Constrained ◽  
Encryption Decryption ◽  
Resource Constrained Devices ◽  
Constrained Devices

Abstract The existing Key Aggregate Searchable Encryption (KASE) schemes allow searches on the encrypted dataset using a single query trapdoor, with a feature to delegate the search rights of multiple files using a constant size key. However, the operations required to generate the ciphertext and decrypt it in these schemes incur higher computational costs, due to the computationally expensive pairing operations in encryption/decryption. This makes the use of such schemes in resource-constrained devices, such as Radio Frequency Identification Devices, Wireless Sensor Network nodes, Internet of Things nodes, infeasible. Motivated with the goal to reduce the computational cost, in this paper, we propose a Revocable Online/Offline KASE (R-OO-KASE) scheme, based on the idea of splitting the encryption/decryption operations into two distinct phases: online and offline. The offline phase computes the majority of costly operations when the device is on an electrical power source. The online phase generates final output with the minimal computational cost when the message (or ciphertext) and keywords become known. In addition, the proposed scheme R-OO-KASE also offers multi-keyword search capability and allows the data owners to revoke the delegated rights at any point in time, the two features are not supported in the existing schemes. The security analysis and empirical evaluations show that the proposed scheme is efficient to use in resource-constrained devices and provably secure as compared to the existing KASE schemes.

scholarly journals FPGA Implementation of Some Second Round NIST Lightweight Cryptography Candidates

Electronics ◽  
2020 ◽  
Vol 9 (11) ◽  
pp. 1940
Author(s):  
Brisbane Ovilla-Martínez ◽  
Cuauhtemoc Mancillas-López ◽  
Alberto F. Martínez-Herrera ◽  
José A. Bernal-Gutiérrez
Keyword(s):  
Hardware Implementation ◽  
Academic Community ◽  
Implementation Cost ◽  
Lightweight Cryptography ◽  
Hardware Implementations ◽  
Open Process ◽  
And Performance ◽  
Resource Constrained Devices ◽  
The U.S ◽  
Constrained Devices

For almost one decade, the academic community has been working in the design and analysis of new lightweight primitives. This cryptography development aims to provide solutions tailored for resource-constrained devices. The U.S. National Institute of Standards and Technology (NIST) started an open process to create a Lightweight Cryptography Standardization portfolio. As a part of the process, the candidates must demonstrate their suitability for hardware implementation. Cost and performance are two of the criteria to be evaluated. In this work, we present the analysis of costs and performance in hardware implementations over five NIST LWC Round 2 candidates, COMET, ESTATE-AES/Gift, LOCUS, LOTUS, and Oribatida. Each candidate’s implementation was adapted to the Hardware API for Lightweight Cryptography for fair benchmarking of hardware cores. The results were generated for Xilinx Artix-7 xc7a12tcsg325-3. The results indicate that it is feasible to achieve the reduction of each solution below 2000 LUTs and 2000 slices where some of them (the variants of ESTATE-AES/Gift) are below 850 LUTs and 600 FF when they are included in the LWC CryptoCore.

SIBSC: Separable Identity-Based Signcryption for Resource-Constrained Devices

Informatica ◽  
2017 ◽  
Vol 28 (1) ◽  
pp. 193-214 ◽  
Author(s):  
Tung-Tso Tsai ◽  
Sen-Shan Huang ◽  
Yuh-Min Tseng
Keyword(s):  
Resource Constrained ◽  
Identity Based ◽  
Resource Constrained Devices ◽  
Constrained Devices

Sky Segmentation for Enhanced Depth Reconstruction and Bokeh Rendering with Efficient Architectures

2020 ◽  
Vol 2020 (14) ◽  
pp. 378-1-378-7
Author(s):  
Tyler Nuanes ◽  
Matt Elsey ◽  
Radek Grzeszczuk ◽  
John Paul Shen
Keyword(s):  
Real Time ◽  
Mobile Device ◽  
Computational Cost ◽  
False Positives ◽  
Compact Model ◽  
High Quality ◽  
False Negatives ◽  
Trade Off ◽  
Depth Reconstruction ◽  
Binary Classifiers

We present a high-quality sky segmentation model for depth refinement and investigate residual architecture performance to inform optimally shrinking the network. We describe a model that runs in near real-time on mobile device, present a new, highquality dataset, and detail a unique weighing to trade off false positives and false negatives in binary classifiers. We show how the optimizations improve bokeh rendering by correcting stereo depth misprediction in sky regions. We detail techniques used to preserve edges, reject false positives, and ensure generalization to the diversity of sky scenes. Finally, we present a compact model and compare performance of four popular residual architectures (ShuffleNet, MobileNetV2, Resnet-101, and Resnet-34-like) at constant computational cost.

Non-iterative online sequential learning strategy for autoencoder and classifier

2021 ◽  
Author(s):  
Adhri Nandini Paul ◽  
Peizhi Yan ◽  
Yimin Yang ◽  
Hui Zhang ◽  
Shan Du ◽  
...  
Keyword(s):  
Learning Strategy ◽  
Sequential Learning ◽  
Online Sequential Learning

scholarly journals Efficient Embedded Decoding of Neural Network Language Models in a Machine Translation System

2018 ◽  
Vol 28 (09) ◽  
pp. 1850007
Author(s):  
Francisco Zamora-Martinez ◽  
Maria Jose Castro-Bleda
Keyword(s):  
Neural Network ◽  
Machine Translation ◽  
Language Processing ◽  
Traditional Approach ◽  
Computational Cost ◽  
Integrated Approach ◽  
Language Models ◽  
Translation System ◽  
Neural Net ◽  
Network Language

Neural Network Language Models (NNLMs) are a successful approach to Natural Language Processing tasks, such as Machine Translation. We introduce in this work a Statistical Machine Translation (SMT) system which fully integrates NNLMs in the decoding stage, breaking the traditional approach based on [Formula: see text]-best list rescoring. The neural net models (both language models (LMs) and translation models) are fully coupled in the decoding stage, allowing to more strongly influence the translation quality. Computational issues were solved by using a novel idea based on memorization and smoothing of the softmax constants to avoid their computation, which introduces a trade-off between LM quality and computational cost. These ideas were studied in a machine translation task with different combinations of neural networks used both as translation models and as target LMs, comparing phrase-based and [Formula: see text]-gram-based systems, showing that the integrated approach seems more promising for [Formula: see text]-gram-based systems, even with nonfull-quality NNLMs.

Sign in / Sign up

Export Citation Format

Share Document