Matching Items (3)
Filtering by
- All Subjects: deep learning
Description
With the emergence of edge computing paradigm, many applications such as image recognition and augmented reality require to perform machine learning (ML) and artificial intelligence (AI) tasks on edge devices. Most AI and ML models are large and computational heavy, whereas edge devices are usually equipped with limited computational and storage resources. Such models can be compressed and reduced in order to be placed on edge devices, but they may loose their capability and may not generalize and perform well compared to large models. Recent works used knowledge transfer techniques to transfer information from a large network (termed teacher) to a small one (termed student) in order to improve the performance of the latter. This approach seems to be promising for learning on edge devices, but a thorough investigation on its effectiveness is lacking.
The purpose of this work is to provide an extensive study on the performance (both in terms of accuracy and convergence speed) of knowledge transfer, considering different student-teacher architectures, datasets and different techniques for transferring knowledge from teacher to student.
A good performance improvement is obtained by transferring knowledge from both the intermediate layers and last layer of the teacher to a shallower student. But other architectures and transfer techniques do not fare so well and some of them even lead to negative performance impact. For example, a smaller and shorter network, trained with knowledge transfer on Caltech 101 achieved a significant improvement of 7.36\% in the accuracy and converges 16 times faster compared to the same network trained without knowledge transfer. On the other hand, smaller network which is thinner than the teacher network performed worse with an accuracy drop of 9.48\% on Caltech 101, even with utilization of knowledge transfer.
The purpose of this work is to provide an extensive study on the performance (both in terms of accuracy and convergence speed) of knowledge transfer, considering different student-teacher architectures, datasets and different techniques for transferring knowledge from teacher to student.
A good performance improvement is obtained by transferring knowledge from both the intermediate layers and last layer of the teacher to a shallower student. But other architectures and transfer techniques do not fare so well and some of them even lead to negative performance impact. For example, a smaller and shorter network, trained with knowledge transfer on Caltech 101 achieved a significant improvement of 7.36\% in the accuracy and converges 16 times faster compared to the same network trained without knowledge transfer. On the other hand, smaller network which is thinner than the teacher network performed worse with an accuracy drop of 9.48\% on Caltech 101, even with utilization of knowledge transfer.
ContributorsSistla, Ragini (Author) / Zhao, Ming (Thesis advisor, Committee member) / Li, Baoxin (Committee member) / Tong, Hanghang (Committee member) / Arizona State University (Publisher)
Created2018
Description
Compartmentalizing access to content, be it websites accessed in a browser or documents and applications accessed outside the browser, is an established method for protecting information integrity [12, 19, 21, 60]. Compartmentalization solutions change the user experience, introduce performance overhead and provide varying degrees of security. Striking a balance between usability and security is not an easy task. If the usability aspects are neglected or sacrificed in favor of more security, the resulting solution would have a hard time being adopted by end-users. The usability is affected by factors including (1) the generality of the solution in supporting various applications, (2) the type of changes required, (3) the performance overhead introduced by the solution, and (4) how much the user experience is preserved. The security is affected by factors including (1) the attack surface of the compartmentalization mechanism, and (2) the security decisions offloaded to the user. This dissertation evaluates existing solutions based on the above factors and presents two novel compartmentalization solutions that are arguably more practical than their existing counterparts.
The first solution, called FlexICon, is an attractive alternative in the design space of compartmentalization solutions on the desktop. FlexICon allows for the creation of a large number of containers with small memory footprint and low disk overhead. This is achieved by using lightweight virtualization based on Linux namespaces. FlexICon uses two mechanisms to reduce user mistakes: 1) a trusted file dialog for selecting files for opening and launching it in the appropriate containers, and 2) a secure URL redirection mechanism that detects the user’s intent and opens the URL in the proper container. FlexICon also provides a language to specify the access constraints that should be enforced by various containers.
The second solution called Auto-FBI, deals with web-based attacks by creating multiple instances of the browser and providing mechanisms for switching between the browser instances. The prototype implementation for Firefox and Chrome uses system call interposition to control the browser’s network access. Auto-FBI can be ported to other platforms easily due to simple design and the ubiquity of system call interposition methods on all major desktop platforms.
The first solution, called FlexICon, is an attractive alternative in the design space of compartmentalization solutions on the desktop. FlexICon allows for the creation of a large number of containers with small memory footprint and low disk overhead. This is achieved by using lightweight virtualization based on Linux namespaces. FlexICon uses two mechanisms to reduce user mistakes: 1) a trusted file dialog for selecting files for opening and launching it in the appropriate containers, and 2) a secure URL redirection mechanism that detects the user’s intent and opens the URL in the proper container. FlexICon also provides a language to specify the access constraints that should be enforced by various containers.
The second solution called Auto-FBI, deals with web-based attacks by creating multiple instances of the browser and providing mechanisms for switching between the browser instances. The prototype implementation for Firefox and Chrome uses system call interposition to control the browser’s network access. Auto-FBI can be ported to other platforms easily due to simple design and the ubiquity of system call interposition methods on all major desktop platforms.
ContributorsZohrevandi, Mohsen (Author) / Bazzi, Rida A (Thesis advisor) / Ahn, Gail-Joon (Committee member) / Doupe, Adam (Committee member) / Zhao, Ming (Committee member) / Arizona State University (Publisher)
Created2018
Description
Deep learning and AI have grabbed tremendous attention in the last decade. The substantial accuracy improvement by neural networks in common tasks such as image classification and speech recognition has made deep learning as a replacement for many conventional machine learning techniques. Training Deep Neural networks require a lot of data, and therefore vast of amounts of computing resources to process the data and train the model for the neural network. The most obvious solution to solving this problem is to speed up the time it takes to train Deep Neural networks.
AI and deep learning workloads are different from the conventional cloud and mobile workloads, with respect to: (1) Computational Intensity, (2) I/O characteristics, and (3) communication pattern. While there is a considerable amount of research activity on the theoretical aspects of AI and Deep Learning algorithms that run with greater efficiency, there are only a few studies on the infrastructural impact of Deep Learning workloads on computing and storage resources in distributed systems.
It is typical to utilize a heterogeneous mixture of CPU and GPU devices to perform training on a neural network. Google Brain has a developed a reinforcement model that can place training operations across a heterogeneous cluster. Though it has only been tested with local devices in a single cluster. This study will explore the method’s capabilities and attempt to apply this method on a cluster with nodes across a network.
AI and deep learning workloads are different from the conventional cloud and mobile workloads, with respect to: (1) Computational Intensity, (2) I/O characteristics, and (3) communication pattern. While there is a considerable amount of research activity on the theoretical aspects of AI and Deep Learning algorithms that run with greater efficiency, there are only a few studies on the infrastructural impact of Deep Learning workloads on computing and storage resources in distributed systems.
It is typical to utilize a heterogeneous mixture of CPU and GPU devices to perform training on a neural network. Google Brain has a developed a reinforcement model that can place training operations across a heterogeneous cluster. Though it has only been tested with local devices in a single cluster. This study will explore the method’s capabilities and attempt to apply this method on a cluster with nodes across a network.
ContributorsNguyen, Andrew Hoang (Author) / Zhao, Ming (Thesis director) / Biookaghazadeh, Saman (Committee member) / Computer Science and Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2019-05