Matching Items (56)
Filtering by
- Genre: Masters Thesis
- Creators: Li, Baoxin
Description
Convolutional Neural Network (CNN) has achieved state-of-the-art performance in numerous applications like computer vision, natural language processing, robotics etc. The advancement of High-Performance Computing systems equipped with dedicated hardware accelerators has also paved the way towards the success of compute intensive CNNs. Graphics Processing Units (GPUs), with massive processing capability, have been of general interest for the acceleration of CNNs. Recently, Field Programmable Gate Arrays (FPGAs) have been promising in CNN acceleration since they offer high performance while also being re-configurable to support the evolution of CNNs. This work focuses on a design methodology to accelerate CNNs on FPGA with low inference latency and high-throughput which are crucial for scenarios like self-driving cars, video surveillance etc. It also includes optimizations which reduce the resource utilization by a large margin with a small degradation in performance thus making the design suitable for low-end FPGA devices as well.
FPGA accelerators often suffer due to the limited main memory bandwidth. Also, highly parallel designs with large resource utilization often end up achieving low operating frequency due to poor routing. This work employs data fetch and buffer mechanisms, designed specifically for the memory access pattern of CNNs, that overlap computation with memory access. This work proposes a novel arrangement of the systolic processing element array to achieve high frequency and consume less resources than the existing works. Also, support has been extended to more complicated CNNs to do video processing. On Intel Arria 10 GX1150, the design operates at a frequency as high as 258MHz and performs single inference of VGG-16 and C3D in 23.5ms and 45.6ms respectively. For VGG-16 and C3D the design offers a throughput of 66.1 and 23.98 inferences/s respectively. This design can outperform other FPGA 2D CNN accelerators by up to 9.7 times and 3D CNN accelerators by up to 2.7 times.
FPGA accelerators often suffer due to the limited main memory bandwidth. Also, highly parallel designs with large resource utilization often end up achieving low operating frequency due to poor routing. This work employs data fetch and buffer mechanisms, designed specifically for the memory access pattern of CNNs, that overlap computation with memory access. This work proposes a novel arrangement of the systolic processing element array to achieve high frequency and consume less resources than the existing works. Also, support has been extended to more complicated CNNs to do video processing. On Intel Arria 10 GX1150, the design operates at a frequency as high as 258MHz and performs single inference of VGG-16 and C3D in 23.5ms and 45.6ms respectively. For VGG-16 and C3D the design offers a throughput of 66.1 and 23.98 inferences/s respectively. This design can outperform other FPGA 2D CNN accelerators by up to 9.7 times and 3D CNN accelerators by up to 2.7 times.
ContributorsRavi, Pravin Kumar (Author) / Zhao, Ming (Thesis advisor) / Li, Baoxin (Committee member) / Ren, Fengbo (Committee member) / Arizona State University (Publisher)
Created2020
Description
At modern-day intersections, traffic lights and stop signs assist human drivers to cross the intersection safely. Traffic congestion in urban road networks is a costly problem that affects all major cities. Efficiently operating intersections is largely dependent on accuracy and precision of human drivers, engendering a lingering uncertainty of attaining safety and high throughput. To improve the efficiency of the existing traffic network and mitigate the effects of human error in the intersection, many studies have proposed autonomous, intelligent transportation systems. These studies often involve utilizing connected autonomous vehicles, implementing a supervisory system, or both. Implementing a supervisory system is relatively more popular due to the security concerns of vehicle-to-vehicle communication. Even though supervisory systems are a step in the right direction for security, many supervisory systems’ safe operation solely relies on the promise of connected data being correct, making system reliability difficult to achieve. To increase fault-tolerance and decrease the effects of position uncertainty, this thesis proposes the Reliable and Robust Intersection Manager, a supervisory system that uses a separate surveillance system to dependably detect vehicles present in the intersection in order to create data redundancy for more accurate scheduling of connected autonomous vehicles. Adding the Surveillance System ensures that the temporal safety buffers between arrival times of connected autonomous vehicles are maintained. This guarantees that connected autonomous vehicles can traverse the intersection safely in the event of large vehicle controller error, a single rogue car entering the intersection, or a sybil attack. To test the proposed system given these fault-models, MATLAB® was used to create simulations in order to observe the functionality of R2IM compared to the state-of-the-art supervisory system, Robust Intersection Manager. Though R2IM is less efficient than the Robust Intersection Manager, it considers more fault models. The Robust Intersection Manager failed to maintain safety in the event of large vehicle controller errors and rogue cars, however R2IM resulted in zero collisions.
ContributorsDedinsky, Rachel (Author) / Shrivastava, Aviral (Thesis advisor) / Sen, Arunabha (Committee member) / Syrotiuk, Violet (Committee member) / Arizona State University (Publisher)
Created2019
Description
Machine learning (ML) and deep neural networks (DNNs) have achieved great success in a variety of application domains, however, despite significant effort to make these networks robust, they remain vulnerable to adversarial attacks in which input that is perceptually indistinguishable from natural data can be erroneously classified with high prediction confidence. Works on defending against adversarial examples can be broadly classified as correcting or detecting, which aim, respectively at negating the effects of the attack and correctly classifying the input, or detecting and rejecting the input as adversarial. In this work, a new approach for detecting adversarial examples is proposed. The approach takes advantage of the robustness of natural images to noise. As noise is added to a natural image, the prediction probability of its true class drops, but the drop is not sudden or precipitous. The same seems to not hold for adversarial examples. In other word, the stress response profile for natural images seems different from that of adversarial examples, which could be detected by their stress response profile. An evaluation of this approach for detecting adversarial examples is performed on the MNIST, CIFAR-10 and ImageNet datasets. Experimental data shows that this approach is effective at detecting some adversarial examples on small scaled simple content images and with little sacrifice on benign accuracy.
ContributorsSun, Lin (Author) / Bazzi, Rida (Thesis advisor) / Li, Baoxin (Committee member) / Tong, Hanghang (Committee member) / Arizona State University (Publisher)
Created2019
Description
In recent years, Convolutional Neural Networks (CNNs) have been widely used in not only the computer vision community but also within the medical imaging community. Specifically, the use of pre-trained CNNs on large-scale datasets (e.g., ImageNet) via transfer learning for a variety of medical imaging applications, has become the de facto standard within both communities.
However, to fit the current paradigm, 3D imaging tasks have to be reformulated and solved in 2D, losing rich 3D contextual information. Moreover, pre-trained models on natural images never see any biomedical images and do not have knowledge about anatomical structures present in medical images. To overcome the above limitations, this thesis proposes an image out-painting self-supervised proxy task to develop pre-trained models directly from medical images without utilizing systematic annotations. The idea is to randomly mask an image and train the model to predict the missing region. It is demonstrated that by predicting missing anatomical structures when seeing only parts of the image, the model will learn generic representation yielding better performance on various medical imaging applications via transfer learning.
The extensive experiments demonstrate that the proposed proxy task outperforms training from scratch in six out of seven medical imaging applications covering 2D and 3D classification and segmentation. Moreover, image out-painting proxy task offers competitive performance to state-of-the-art models pre-trained on ImageNet and other self-supervised baselines such as in-painting. Owing to its outstanding performance, out-painting is utilized as one of the self-supervised proxy tasks to provide generic 3D pre-trained models for medical image analysis.
However, to fit the current paradigm, 3D imaging tasks have to be reformulated and solved in 2D, losing rich 3D contextual information. Moreover, pre-trained models on natural images never see any biomedical images and do not have knowledge about anatomical structures present in medical images. To overcome the above limitations, this thesis proposes an image out-painting self-supervised proxy task to develop pre-trained models directly from medical images without utilizing systematic annotations. The idea is to randomly mask an image and train the model to predict the missing region. It is demonstrated that by predicting missing anatomical structures when seeing only parts of the image, the model will learn generic representation yielding better performance on various medical imaging applications via transfer learning.
The extensive experiments demonstrate that the proposed proxy task outperforms training from scratch in six out of seven medical imaging applications covering 2D and 3D classification and segmentation. Moreover, image out-painting proxy task offers competitive performance to state-of-the-art models pre-trained on ImageNet and other self-supervised baselines such as in-painting. Owing to its outstanding performance, out-painting is utilized as one of the self-supervised proxy tasks to provide generic 3D pre-trained models for medical image analysis.
ContributorsSodha, Vatsal Arvindkumar (Author) / Liang, Jianming (Thesis advisor) / Devarakonda, Murthy (Committee member) / Li, Baoxin (Committee member) / Arizona State University (Publisher)
Created2020
Description
ABSTRACT
This study examines the ways in which employees experience moral emotions that violate employee treatment and how employees co-construct moral emotions and subsequent expressions of dissent. This qualitative study consisted of 123 full-time employees and utilized open-coding, content analysis, constant comparison analysis, and concept mapping. The analysis revealed that employees expressed dissent laterally as a series of sensemaking processes, such as validation of feelings, moral assessments, and assessing the fear of moral transgressions. Employees also expressed dissent as a series of risk assessments that overlapped with the ways in which employees made sense of the perceived infraction. Employees' lateral dissent expression manifested as a form of social support which occasionally led to co-rumination. Employees expressed dissent upwardly when seeking a desired action or change. Circumvention was utilized as a direct reflection to the type and degree of moral transgression related to the person responsible for the mistreatment. Results indicated that experiencing moral emotions that led to expressing dissent with a designated audience was determined by where employees were situated in the cyclical model of communicating moral emotions and in relation to the co-construction of both the infraction related to employee mistreatment and the experience of moral emotions. Results contribute to the existing body of literature on dissent and emotions. A discussion synthesizing the findings and analysis is presented, in addition to the implications for future research.
KEYWORDS: Emotion, Dissent, Moral Emotions, Sensemaking, Risk-Assessment, Social Support, Co-Rumination
This study examines the ways in which employees experience moral emotions that violate employee treatment and how employees co-construct moral emotions and subsequent expressions of dissent. This qualitative study consisted of 123 full-time employees and utilized open-coding, content analysis, constant comparison analysis, and concept mapping. The analysis revealed that employees expressed dissent laterally as a series of sensemaking processes, such as validation of feelings, moral assessments, and assessing the fear of moral transgressions. Employees also expressed dissent as a series of risk assessments that overlapped with the ways in which employees made sense of the perceived infraction. Employees' lateral dissent expression manifested as a form of social support which occasionally led to co-rumination. Employees expressed dissent upwardly when seeking a desired action or change. Circumvention was utilized as a direct reflection to the type and degree of moral transgression related to the person responsible for the mistreatment. Results indicated that experiencing moral emotions that led to expressing dissent with a designated audience was determined by where employees were situated in the cyclical model of communicating moral emotions and in relation to the co-construction of both the infraction related to employee mistreatment and the experience of moral emotions. Results contribute to the existing body of literature on dissent and emotions. A discussion synthesizing the findings and analysis is presented, in addition to the implications for future research.
KEYWORDS: Emotion, Dissent, Moral Emotions, Sensemaking, Risk-Assessment, Social Support, Co-Rumination
ContributorsKamrath, Jessica K (Author) / Kassing, Jeffrey W. (Thesis advisor) / Waldron, Vincent R. (Committee member) / Meân, Lindsey J. (Committee member) / Arizona State University (Publisher)
Created2015
Description
A major challenge in automated text analysis is that different words are used for related concepts. Analyzing text at the surface level would treat related concepts (i.e. actors, actions, targets, and victims) as different objects, potentially missing common narrative patterns. Generalized concepts are used to overcome this problem. Generalization may result into word sense disambiguation failing to find similarity. This is addressed by taking into account contextual synonyms. Concept discovery based on contextual synonyms reveal information about the semantic roles of the words leading to concepts. Merger engine generalize the concepts so that it can be used as features in learning algorithms.
ContributorsKedia, Nitesh (Author) / Davulcu, Hasan (Thesis advisor) / Corman, Steve R (Committee member) / Li, Baoxin (Committee member) / Arizona State University (Publisher)
Created2015