Applications of Resource Efficient Machine Learning in Smart Sensors
A special issue of Sensors (ISSN 1424-8220). This special issue belongs to the section "Physical Sensors".
Deadline for manuscript submissions: closed (31 August 2022) | Viewed by 2010
Special Issue Editors
Interests: Applications of machine (deep) learning; Anomaly detection; (Semi-)supervised learning strategies; Automated interpretation of time-series signals such as acoustic, radar and accelerometer signals; Real-time machine learning on resource constrained devices
Special Issue Information
Dear Colleagues,
In cloud-based machine learning (ML) applications, inference and learning occur in the cloud. However, due to a growing number of more powerful edge devices, together with certain privacy concerns and low latency requirements, there is a movement to transfer the information and part of the learning to areas near the sensors. Since sensors also increasingly include local processing capabilities, a part of the information can even shift to the sensors themselves. Such smart sensors will process all raw data locally and only transmit meaningful results. This trend is supported by the increase in offerings of commercially available ML processing cores, with different performance levels for various applications, including small devices such as for IoT and even dedicated processors designed for battery-powered operation. This paradigm change is clearly seen in the smartphone market, where local ML inference is introduced by including powerful but energy-efficient ML engines in the devices. Due to less communication occurring with a central computing unit in the cloud, the latency is significantly reduced. However, to fully support this trend, new resource-efficient ML solutions are needed that meet power, speed, and size constraints and that maximally benefit from the processing advantages ML accelerators offer.
This Special Issue aims to gather the latest results pertaining to novel, resource-efficient ML algorithms used in smart sensors for smart city, healthcare, industry 4.0, precision farming, and smart transportation applications. Thus, we welcome contributions on—but not limited to—the following topics:
- Resource-efficient machine learning (including deep learning) model inference;
- Resource-efficient training of machine learning (including deep learning) models at the edge (smart sensor);
- Federated learning or stream-based active learning methods;
- Approximation, quantization, and reduced precision computing;
- Sparse modeling (e.g., model pruning);
- Neural architecture search methods;
- Communication or computation scheduling for better performance and energy use;
- Load balancing and efficient task distribution techniques;
- Exploring the interplay between precision, performance, power, and energy;
- Security and privacy implications and preservation solutions;
- Novel applications of machine learning in smart sensors across all fields and emerging use cases;
- Discussions about real-world use cases;
- Surveys on practical experiences.
Prof. Dr. Peter Karsmakers
Guest Editor
Manuscript Submission Information
Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.
Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Sensors is an international peer-reviewed open access semimonthly journal published by MDPI.
Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.
Keywords
- Resource-efficient machine learning
- embedded machine learning
- smart sensors
- extreme edge computing
- applications of machine learning in smart sensors
