A Survey of Methods for Low-Power Deep Learning and Computer Vision

Goel, Abhinav; Tung, Caleb; Lu, Yung-Hsiang; Thiruvathukal, George K.

doi:10.1109/wf-iot48130.2020.9221198

Cited by 75 publications

(37 citation statements)

References 59 publications

(69 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, it is evident that to tackle this issue we need to focus on the effective construction of the application domain of IoMT. The components of this application domain can be addressed as advanced level machine learning and deep learning [ 49 ], reasoning [ 50 ], natural language processing [ 51 ], speech recognition [ 52 ] and computer vision (image object recognition) [ 53 ], human-computer interaction, and dialog and narrative generation. From a global perspective, this can be used to incorporate the new generation hardware and software systems that imitate the human brain and cognitive functionality and thus enhance the human decision-making process.…”

Section: Discussionmentioning

confidence: 99%

Federated Semi-Supervised Multi-Task Learning to Detect COVID-19 and Lungs Segmentation Marking Using Chest Radiography Images and Raspberry Pi Devices: An Internet of Medical Things Application

Alam

Rahmani

2021

Sensors

View full text Add to dashboard Cite

Internet of Medical Things (IoMT) provides an excellent opportunity to investigate better automatic medical decision support tools with the effective integration of various medical equipment and associated data. This study explores two such medical decision-making tasks, namely COVID-19 detection and lung area segmentation detection, using chest radiography images. We also explore different cutting-edge machine learning techniques, such as federated learning, semi-supervised learning, transfer learning, and multi-task learning to explore the issue. To analyze the applicability of computationally less capable edge devices in the IoMT system, we report the results using Raspberry Pi devices as accuracy, precision, recall, Fscore for COVID-19 detection, and average dice score for lung segmentation detection tasks. We also publish the results obtained through server-centric simulation for comparison. The results show that Raspberry Pi-centric devices provide better performance in lung segmentation detection, and server-centric experiments provide better results in COVID-19 detection. We also discuss the IoMT application-centric settings, utilizing medical data and decision support systems, and posit that such a system could benefit all the stakeholders in the IoMT domain.

show abstract

Section: Discussionmentioning

confidence: 99%

Federated Semi-Supervised Multi-Task Learning to Detect COVID-19 and Lungs Segmentation Marking Using Chest Radiography Images and Raspberry Pi Devices: An Internet of Medical Things Application

Alam

Rahmani

2021

Sensors

View full text Add to dashboard Cite

show abstract

“…Traditional DNN training. DNN training has long been considered as a slow and energy harvesting task [16,25,52]. On the surface, the massive energy consumption and high latency mainly come from millions of Multiply-accumulate operations (MACs) and intensive data movement between memory and processing elements (PEs).…”

Section: Challenges Of Bnn Trainingmentioning

confidence: 99%

“…Training DNN models on current hardware devices has long been considered as a slow and energy-consuming task [16,25,52]. Compared with the traditional DNN training, BNN training inefficiency is further exacerbated by the requirement of training an ensemble of sampled DNN models to ensure robustness.…”

Section: Introductionmentioning

confidence: 99%

Shift-BNN: Highly-Efficient Probabilistic Bayesian Neural Network Training via Memory-Friendly Pattern Retrieving

Wan

Xia

Zhang

et al. 2021

MICRO-54: 54th Annual IEEE/ACM International Symposium on Microarchitecture

View full text Add to dashboard Cite

Bayesian Neural Networks (BNNs) that possess a property of uncertainty estimation have been increasingly adopted in a wide range of safety-critical AI applications which demand reliable and robust decision making, e.g., self-driving, rescue robots, medical image diagnosis. The training procedure of a probabilistic BNN model involves training an ensemble of sampled DNN models, which induces orders of magnitude larger volume of data movement than training a single DNN model. In this paper, we reveal that the root cause for BNN training inefficiency originates from the massive off-chip data transfer by Gaussian Random Variables (GRVs). To tackle this challenge, we propose a novel design that eliminates all the off-chip data transfer by GRVs through the reversed shifting of Linear Feedback Shift Registers (LFSRs) without incurring any training accuracy loss. To efficiently support our LFSR reversion strategy at the hardware level, we explore the design space of the current DNN accelerators and identify the optimal computation mapping scheme to best accommodate our strategy. By leveraging this finding, we design and prototype the first highly efficient BNN training accelerator, named Shift-BNN, that is low-cost and scalable. Extensive evaluation on five representative BNN models demonstrates that Shift-BNN achieves an average of 4.9× (up to 10.8×) boost in energy efficiency and 1.6× (up to 2.8×) speedup over the baseline DNN training accelerator. CCS CONCEPTS• Computer systems organization → Neural networks; • Hardware → Hardware accelerators.

show abstract

“…Low-Power Computer Vision: Goel et al [7] survey lowpower DNNs and describe the benefits of reducing memory and operations for low-power applications. DNN quantization reduces the memory requirement [19] and DNN pruning reduces the DNN operations [20].…”

Section: B Related Workmentioning

confidence: 99%

Low-power object counting with hierarchical neural networks

Goel

Tung

Aghajanzadeh

et al. 2020

Proceedings of the ACM/IEEE International Symposium on Low Power Electronics and Design

Self Cite

View full text Add to dashboard Cite

Low-power computer vision on embedded devices has many applications. This paper describes a low-power technique for the object re-identification (reID) problem: matching a query image against a gallery of previously-seen images. State-ofthe-art techniques rely on large, computationally-intensive Deep Neural Networks (DNNs). We propose a novel hierarchical DNN architecture that uses attribute labels in the training dataset to perform efficient object reID. At each node in the hierarchy, a small DNN identifies a different attribute of the query image. The small DNN at each leaf node is specialized to re-identify a subset of the gallery-only the images with the attributes identified along the path from the root to a leaf. Thus, a query image is reidentified accurately after processing with a few small DNNs. We compare our method with state-of-the-art object reID techniques. With a ∼4% loss in accuracy, our approach realizes significant resource savings: 74% less memory, 72% fewer operations, and 67% lower query latency, yielding 65% less energy consumption.

show abstract

A Survey of Methods for Low-Power Deep Learning and Computer Vision

Cited by 75 publications

References 59 publications

Federated Semi-Supervised Multi-Task Learning to Detect COVID-19 and Lungs Segmentation Marking Using Chest Radiography Images and Raspberry Pi Devices: An Internet of Medical Things Application

Federated Semi-Supervised Multi-Task Learning to Detect COVID-19 and Lungs Segmentation Marking Using Chest Radiography Images and Raspberry Pi Devices: An Internet of Medical Things Application

Shift-BNN: Highly-Efficient Probabilistic Bayesian Neural Network Training via Memory-Friendly Pattern Retrieving

Low-power object counting with hierarchical neural networks

Contact Info

Product

Resources

About