Efficient Deep Structure Learning for Resource-Limited IoT Devices

AbstractNowadays, deep neural networks (DNNs) have been rapidly deployed to realize a number of functionalities like sensing, imaging, classification, recognition, etc. However, the computational-intensive requirement of DNNs makes it difficult to be applicable for resource-limited Internet of Things (IoT) devices. In this paper, we propose a novel pruning-based paradigm that aims to reduce the computational cost of DNNs, by uncovering a more compact structure and learning the effective weights therein, on the basis of not compromising the expressive capability of DNNs. In particular, our algorithm can achieve efficient end-to-end training that transfers a redundant neural network to a compact one with a specifically targeted compression rate directly. We comprehensively evaluate our approach on various representative benchmark datasets and compared with typical advanced convolutional neural network (CNN) architectures. The experimental results verify the superior performance and robust effectiveness of our scheme. For example, when pruning VGG on CIFAR-10, our proposed scheme is able to significantly reduce its FLOPs (floating-point operations) and number of parameters with a proportion of 76.2% and 94.1%, respectively, while still maintaining a satisfactory accuracy. To sum up, our scheme could facilitate the integration of DNNs into the common machine-learning-based IoT framework and establish distributed training of neural networks in both cloud and edge.

Download Full-text

IoT Notary : Attestable Sensor Data Capture in IoT Environments

ACM Transactions on Internet of Things ◽

10.1145/3478290 ◽

2022 ◽

Vol 3 (1) ◽

pp. 1-30

Author(s):

Nisha Panwar ◽

Shantanu Sharma ◽

Guoxi Wang ◽

Sharad Mehrotra ◽

Nalini Venkatasubramanian ◽

...

Keyword(s):

Real Time ◽

Location Based Services ◽

Third Party ◽

Sensor Data ◽

Published Data ◽

Resource Limited ◽

User Data ◽

Data Capturing ◽

Iot Devices ◽

The University

Contemporary IoT environments, such as smart buildings, require end-users to trust data-capturing rules published by the systems. There are several reasons why such a trust is misplaced—IoT systems may violate the rules deliberately or IoT devices may transfer user data to a malicious third-party due to cyberattacks, leading to the loss of individuals’ privacy or service integrity. To address such concerns, we propose IoT Notary , a framework to ensure trust in IoT systems and applications. IoT Notary provides secure log sealing on live sensor data to produce a verifiable “proof-of-integrity,” based on which a verifier can attest that captured sensor data adhere to the published data-capturing rules. IoT Notary is an integral part of TIPPERS, a smart space system that has been deployed at the University of California, Irvine to provide various real-time location-based services on the campus. We present extensive experiments over real-time WiFi connectivity data to evaluate IoT Notary , and the results show that IoT Notary imposes nominal overheads. The secure logs only take 21% more storage, while users can verify their one day’s data in less than 2 s even using a resource-limited device.

Download Full-text

Deep Structure Learning for Fraud Detection

2018 IEEE International Conference on Data Mining (ICDM) ◽

10.1109/icdm.2018.00072 ◽

2018 ◽

Cited By ~ 6

Author(s):

Haibo Wang ◽

Chuan Zhou ◽

Jia Wu ◽

Weizhen Dang ◽

Xingquan Zhu ◽

...

Keyword(s):

Structure Learning ◽

Deep Structure ◽

Fraud Detection

Download Full-text

Deep Structure Learning: Beyond Connectionist Approaches

2012 11th International Conference on Machine Learning and Applications ◽

10.1109/icmla.2012.34 ◽

2012 ◽

Cited By ~ 3

Author(s):

Ben Mitchell ◽

John Sheppard

Keyword(s):

Structure Learning ◽

Deep Structure

Download Full-text

Bryofolios

E-Portfolios and Global Diffusion ◽

10.4018/978-1-4666-0143-7.ch003 ◽

2012 ◽

pp. 29-40

Author(s):

Joanne Nakonechny ◽

Shona Ellis

Keyword(s):

Science Education ◽

Learning Theory ◽

Critical Reflection ◽

Structure Learning ◽

Teaching Method ◽

Deep Structure ◽

Authentic Science ◽

Learning Context ◽

Constructivist Learning Theory ◽

Implementation Support

Throughout this chapter, the authors trace how the theoretical and practical understanding, interpretation, and interactions with e-portfolios and their implementation support, both individually and through group work, students’ abilities to engage in deeper structure learning, and their resulting growth as authentic science scholars. The bryofolio, an individual and group course e-portfolio, begins this online journey to facilitate deeper structure learning for 31 students in Biology 321, Bryophytes: Mosses, Hornworts and Liverworts. (“Bryfolio” is a contraction of “bryophytes” and “e-portfolio.”) Initially, the authors give a short introduction to science education and how constructivist learning theory can include the use of e-portfolios as a teaching method. Following this, e-portfolios are situated within the learning context by providing a definition, a condition, and discussion on the key e-portfolio element,of critical reflection. The authors continue by introducing the bryofolio, its major components, and our analysis of how the bryofolio encourages deep structure learning at both individual and group levels.

Download Full-text

7. Lie group deep structure learning

Lie Group Machine Learning ◽

10.1515/9783110499506-007 ◽

2018 ◽

pp. 207-234

Keyword(s):

Lie Group ◽

Structure Learning ◽

Deep Structure

Download Full-text

Bitcoin Payment-Channels for Resource Limited IoT Devices

Proceedings of the International Conference on Omni-Layer Intelligent Systems - COINS '19 ◽

10.1145/3312614.3312629 ◽

2019 ◽

Cited By ~ 1

Author(s):

Christopher Hannon ◽

Dong Jin

Keyword(s):

Resource Limited ◽

Iot Devices

Download Full-text

LDAF: Low-Bandwidth Distributed Applications Framework in a Use Case of Blockchain-Enabled IoT Devices

Sensors ◽

10.3390/s19102337 ◽

2019 ◽

Vol 19 (10) ◽

pp. 2337 ◽

Cited By ~ 4

Author(s):

Matevž Pustišek ◽

Dejan Dolenc ◽

Andrej Kos

Keyword(s):

Fog Computing ◽

Distributed Applications ◽

Use Case ◽

Smart Meter ◽

Resource Limited ◽

Application Logic ◽

Periodic Traffic ◽

Communication Traffic ◽

Iot Devices ◽

Application Aware

In this paper, we present Low-Bandwidth Distributed Applications Framework (LDAF)—an application-aware gateway for communication-constrained Internet of things (IoT) devices. A modular approach facilitates connecting to existing cloud backend servers and managing message formats and APIs’ native application logic to meet the communication constraints of resource-limited end devices. We investigated options for positioning the LDAF server in fog computing architectures. We demonstrated the approach in three use cases: (i) a simple domain name system (DNS) query from the device to a DNS server, (ii) a complex interaction of a blockchain—based IoT device with a blockchain network, and (iii) difference based patching of binary (system) files at the IoT end devices. In a blockchain smart meter use case we effectively enabled decentralized applications (DApp) for devices that without our solution could not participate in a blockchain network. Employing the more efficient binary content encoding, we reduced the periodic traffic from 16 kB/s to ~1.1 kB/s, i.e., 7% of the initial traffic. With additional optimization of the application protocol in the gateway and message filtering, the periodic traffic was reduced to ~1% of the initial traffic, without any tradeoffs in the application’s functionality or security. Using a function of binary difference we managed to reduce the size of the communication traffic to the end device, at least when the binary patch was smaller than the patching file.

Download Full-text

Survey on Revocation in Ciphertext-Policy Attribute-Based Encryption

Sensors ◽

10.3390/s19071695 ◽

2019 ◽

Vol 19 (7) ◽

pp. 1695 ◽

Cited By ~ 9

Author(s):

Al-Dahhan ◽

Shi ◽

Lee ◽

Kifayat

Keyword(s):

Computational Cost ◽

Sensitive Data ◽

Resource Limited ◽

Attribute Based Encryption ◽

Wide Range ◽

Secure Cloud Storage ◽

Iot Devices ◽

Ciphertext Policy ◽

Cryptographic Techniques ◽

High Computational Cost

Recently, using advanced cryptographic techniques to process, store, and share datasecurely in an untrusted cloud environment has drawn widespread attention from academicresearchers. In particular, Ciphertext‐Policy Attribute‐Based Encryption (CP‐ABE) is a promising,advanced type of encryption technique that resolves an open challenge to regulate fine‐grainedaccess control of sensitive data according to attributes, particularly for Internet of Things (IoT)applications. However, although this technique provides several critical functions such as dataconfidentiality and expressiveness, it faces some hurdles including revocation issues and lack ofmanaging a wide range of attributes. These two issues have been highlighted by many existingstudies due to their complexity which is hard to address without high computational cost affectingthe resource‐limited IoT devices. In this paper, unlike other survey papers, existing single andmultiauthority CP‐ABE schemes are reviewed with the main focus on their ability to address therevocation issues, the techniques used to manage the revocation, and comparisons among themaccording to a number of secure cloud storage criteria. Therefore, this is the first review paperanalysing the major issues of CP‐ABE in the IoT paradigm and explaining the existing approachesto addressing these issues.

Download Full-text