{"id":{"repo_id":"missouri","oai_identifier":"oai:mospace.umsystem.edu:10355/105046"},"canonical_url":"https://search.dev.ndltd.org/etd/missouri/oai:mospace.umsystem.edu:10355/105046","repository":{"repo_id":"missouri","name":"University of Missouri","base_url":"https://mospace.umsystem.edu/oai/request"},"display":{"title":"Reliable, secure and energy-efficient AI hardware","abstract":"The embedded AI hardware chips are being widely used in consumer devices and enterprise markets, such as high-end smartphones, tablets, smart speakers, wearables, autonomous vehicles, cameras, sensors, and other IoT (internet of things) devices. According to a recent report, the embedded AI hardware market is projected to grow from 920 million units in 2021 to 2,080 million units by 2026; it is expected to grow at a CAGR of 17.7 percent from 2021 to 2026. Unfortunately, these embedded AI devices are not only resource and power-constrained but, also vulnerable to reliability and security threats. To date, many energy-aware solutions such as approximate computing have been proposed to address the energy constraints of AI devices. Approximate computing-based deep learning algorithms relax the abstraction with near-perfect accuracy for energy efficiency in errorresilient applications. However, similar to traditional deep neural networks (DNNs), approximate deep neural networks (AxDNNs) and approximate spiking neural networks (AxSNNs) are vulnerable to many reliability threats, such as permanent and transient faults, and security threats, such as adversarial attacks. Considering that approximate computing is energy efficient technique but has an error-inducing nature, there is a pent-up need to exploit the vulnerabilities of AxDNNs against reliability and security threats. In this thesis, we aim to develop sustainable and dependable AI hardware that can analyze and mitigate reliability and security threats in energy-constrained AI hardware. We exploit emerging computing paradigms such as explainable artificial intelligence, neural architecture search, and moving target defense to guarantee reliability, security, and energy efficiency. My Ph.D. thesis is the first effort toward developing energy, reliability, and robustness-aware AI hardware for safety-critical applications.","abstract_html":"The embedded AI hardware chips are being widely used in consumer devices and enterprise markets, such as high-end smartphones, tablets, smart speakers, wearables, autonomous vehicles, cameras, sensors, and other IoT (internet of things) devices. According to a recent report, the embedded AI hardware market is projected to grow from 920 million units in 2021 to 2,080 million units by 2026; it is expected to grow at a CAGR of 17.7 percent from 2021 to 2026. Unfortunately, these embedded AI devices are not only resource and power-constrained but, also vulnerable to reliability and security threats. To date, many energy-aware solutions such as approximate computing have been proposed to address the energy constraints of AI devices. Approximate computing-based deep learning algorithms relax the abstraction with near-perfect accuracy for energy efficiency in errorresilient applications. However, similar to traditional deep neural networks (DNNs), approximate deep neural networks (AxDNNs) and approximate spiking neural networks (AxSNNs) are vulnerable to many reliability threats, such as permanent and transient faults, and security threats, such as adversarial attacks. Considering that approximate computing is energy efficient technique but has an error-inducing nature, there is a pent-up need to exploit the vulnerabilities of AxDNNs against reliability and security threats. In this thesis, we aim to develop sustainable and dependable AI hardware that can analyze and mitigate reliability and security threats in energy-constrained AI hardware. We exploit emerging computing paradigms such as explainable artificial intelligence, neural architecture search, and moving target defense to guarantee reliability, security, and energy efficiency. My Ph.D. thesis is the first effort toward developing energy, reliability, and robustness-aware AI hardware for safety-critical applications.","abstract_has_math":false,"creators":["Siddique, Ayesha"],"institution":"University of Missouri--Columbia","degree_name":"Ph. D.","degree_level":"Doctoral","degree_discipline":"Electrical and computer engineering (MU)","degree_department":null,"school":null,"contributors":[],"advisors":["Hoque, Khaza Anuarul"],"committee_chairs":[],"committee_members":[],"year":2024,"date_issued":"2024","date_published":"2024","updated_at":"2026-07-24T03:07:35Z","subjects":[],"languages":["eng","English"],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier.doi","label":"DOI","values":["https://doi.org/10.32469/10355/105046"],"render_values":[{"text":"https://doi.org/10.32469/10355/105046","href":"https://doi.org/10.32469/10355/105046","code":true}]}]},"links":{"outbound_url":"https://hdl.handle.net/10355/105046","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Hoque, Khaza Anuarul"]},{"key":"dc:creator","label":"Author","values":["Siddique, Ayesha"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2024-09-10T20:42:12Z"]},{"key":"dc:date.issued","label":"Date","values":["2024"]},{"key":"dc:publisher","label":"Institution","values":["University of Missouri--Columbia"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Electrical and computer engineering (MU)"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Doctoral"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph. D."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Missouri--Columbia"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["English"]},{"key":"dc:language.iso","label":"Language (ISO)","values":["eng"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.doi","label":"DOI","values":["https://doi.org/10.32469/10355/105046"]},{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/10355/105046"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["The embedded AI hardware chips are being widely used in consumer devices and enterprise markets, such as high-end smartphones, tablets, smart speakers, wearables, autonomous vehicles, cameras, sensors, and other IoT (internet of things) devices. According to a recent report, the embedded AI hardware market is projected to grow from 920 million units in 2021 to 2,080 million units by 2026; it is expected to grow at a CAGR of 17.7 percent from 2021 to 2026. Unfortunately, these embedded AI devices are not only resource and power-constrained but, also vulnerable to reliability and security threats. To date, many energy-aware solutions such as approximate computing have been proposed to address the energy constraints of AI devices. Approximate computing-based deep learning algorithms relax the abstraction with near-perfect accuracy for energy efficiency in errorresilient applications. However, similar to traditional deep neural networks (DNNs), approximate deep neural networks (AxDNNs) and approximate spiking neural networks (AxSNNs) are vulnerable to many reliability threats, such as permanent and transient faults, and security threats, such as adversarial attacks. Considering that approximate computing is energy efficient technique but has an error-inducing nature, there is a pent-up need to exploit the vulnerabilities of AxDNNs against reliability and security threats. In this thesis, we aim to develop sustainable and dependable AI hardware that can analyze and mitigate reliability and security threats in energy-constrained AI hardware. We exploit emerging computing paradigms such as explainable artificial intelligence, neural architecture search, and moving target defense to guarantee reliability, security, and energy efficiency. My Ph.D. thesis is the first effort toward developing energy, reliability, and robustness-aware AI hardware for safety-critical applications."]},{"key":"dc:title","label":"Title","values":["Reliable, secure and energy-efficient AI hardware"]}]}],"canonical_facts":{"dc:contributor.advisor":["Hoque, Khaza Anuarul"],"dc:creator":["Siddique, Ayesha"],"dc:date.accessioned":["2024-09-10T20:42:12Z"],"dc:date.issued":["2024"],"dc:description.abstract":["The embedded AI hardware chips are being widely used in consumer devices and enterprise markets, such as high-end smartphones, tablets, smart speakers, wearables, autonomous vehicles, cameras, sensors, and other IoT (internet of things) devices. According to a recent report, the embedded AI hardware market is projected to grow from 920 million units in 2021 to 2,080 million units by 2026; it is expected to grow at a CAGR of 17.7 percent from 2021 to 2026. Unfortunately, these embedded AI devices are not only resource and power-constrained but, also vulnerable to reliability and security threats. To date, many energy-aware solutions such as approximate computing have been proposed to address the energy constraints of AI devices. Approximate computing-based deep learning algorithms relax the abstraction with near-perfect accuracy for energy efficiency in errorresilient applications. However, similar to traditional deep neural networks (DNNs), approximate deep neural networks (AxDNNs) and approximate spiking neural networks (AxSNNs) are vulnerable to many reliability threats, such as permanent and transient faults, and security threats, such as adversarial attacks. Considering that approximate computing is energy efficient technique but has an error-inducing nature, there is a pent-up need to exploit the vulnerabilities of AxDNNs against reliability and security threats. In this thesis, we aim to develop sustainable and dependable AI hardware that can analyze and mitigate reliability and security threats in energy-constrained AI hardware. We exploit emerging computing paradigms such as explainable artificial intelligence, neural architecture search, and moving target defense to guarantee reliability, security, and energy efficiency. My Ph.D. thesis is the first effort toward developing energy, reliability, and robustness-aware AI hardware for safety-critical applications."],"dc:identifier.doi":["https://doi.org/10.32469/10355/105046"],"dc:identifier.uri":["https://hdl.handle.net/10355/105046"],"dc:language":["English"],"dc:language.iso":["eng"],"dc:publisher":["University of Missouri--Columbia"],"dc:title":["Reliable, secure and energy-efficient AI hardware"],"dc:type":["Thesis"],"thesis:degree_discipline":["Electrical and computer engineering (MU)"],"thesis:degree_level":["Doctoral"],"thesis:degree_name":["Ph. D."],"thesis:institution_name":["University of Missouri--Columbia"]},"updated_at":"2026-07-24T03:07:35Z"}