[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119471-en":3,"doc-seo-119471-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},119471,962075114765,"Quinn","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","In-Memory Computing to Accelerate Machine Learning - Dissertation","This dissertation investigates in-memory computing approaches that accelerate major machine learning workloads, including hyper dimensional computing-based reinforcement learning, convolutional neural networks, and large language models. It studies data-centric computation using content addressable memories and associative processors, then develops and evaluates accelerator designs for each workload. The work includes framework reviews for mixed-precision quantization, mapping of learning operations to a simulator and hardware-backed execution, and experimental benchmarking with detailed results and comparisons.","UC Irvine  \nUC Irvine Electronic Theses and Dissertations  \nTitle  \nIn-Memory Computing to Accelerate Machine Learning  \nPermalink  \n[https://escholarship.org/uc/item/7r59192k](https://escholarship.org/uc/item/7r59192k)  \nAuthor  \nRakka, Mariam  \nPublication Date  \n2025  \nPeer reviewed|Thesis/dissertation  \n[eScholarship.org](eScholarship.org) Powered by the California Digital Library  \nUniversity of California  \nUNIVERSITY OF CALIFORNIA,  \nIRVINE  \nIn-Memory Computing to Accelerate Machine Learning  \nDISSERTATION  \nsubmitted in partial satisfaction of the requirements  \nfor the degree of  \nDOCTOR OF PHILOSOPHY  \nin Electrical and Computer Engineering  \nby  \nMariam Rakka  \nDissertation Committee: Professor Fadi J. Kurdahi, Chair Professor Nikil Dutt  \nProfessor Rainer D¨omer  \nChapters 1-5 and 7-8 are based on material that appears in publications by the author in IEEE and ACM venues. These chapters incorporate content © IEEE/ACM Publishing, reused with permission and in accordance with the respective publisher’s author rights  \npolicies  \nChapter 6 © 2025 Mariam Rakka  \nDEDICATION  \nTo the journey that chose me before I ever chose it.  \nTo the subtle yet certain signs that felt like God’s way of leading me forward. This PhD was never just an academic pursuit. It was a calling, a quiet unfolding of  \npurpose and faith.  \nTo my parents, living 7,462 miles away, whose love bridged every inch.  \nTo my entire family, none of whom pursued a graduate degree, yet supported me in the ways they best knew how, with unwavering belief and care. To you, your presence in the final stretch made the weight lighter, the road clearer, and the  \nfinish line possible.  \nThis work is as much yours as it is mine.  \nTABLE OF CONTENTS  \nPage  \nLIST OF FIGURES vi  \nLIST OF TABLES viii  \nLIST OF ALGORITHMS x  \nACKNOWLEDGMENTS xi  \nVITA xii  \nABSTRACT OF THE DISSERTATION xiv  \n1 Introduction 1  \n1.1 Motivation ..................................... 1  \n1.2 Contributions ................................... 3  \n1.2.1 Accelerating Hyper Dimensional Computing-Based Reinforcement Learning on Associative Processors ...................... 3  \n1.2.2 Surveying Mixed-Precision Neural Network Quantization Frameworks 4  \n1.2.3 Accelerating Convolutional Neural Network Inference on Associative Processors ................................. 5  \n1.2.4 Accelerating Large Language Model Inference on Associative Processors 6  \n1.3 Organization ................................... 6  \n2 In-Memory Computing 7  \n2.1 A Data-Centric Computing Paradigm ...................... 7  \n2.2 Content Addressable Memories ......................... 8  \n2.3 Associative Processors .............................. 9  \n2.4 Two Dimensional Associative Processors .................... 12  \n3 Machine Learning Workloads 15  \n3.1 Hyper Dimensional Computing-Based Reinforcement Learning ........ 15  \n3.2 Convolutional Neural Networks ......................... 17  \n3.3 Large Language Models ............................. 19  \n4 Accelerating Hyper Dimensional Computing-Based Reinforcement Learning 21  \n4.1 HDRLPIM ..................................... 22  \n4.1.1 Simulator Flow .............................. 22  \n4.1.2 Hardware Modeled ............................ 24  \n4.2 Mapping Operations to the Simulator ...................... 27  \n4.2.1 Q-value Calculation ............................ 27  \n4.2.2 Action Index Selection .......................... 28  \n4.3 Experimental Setup ................................ 30  \n4.3.1 FPGA Setup ............................... 30  \n4.3.2 HDRLPIM Setup ............................. 31  \n4.3.3 Workloads ................................. 32  \n4.4 Results ....................................... 33  \n4.5 Related Works and Comparison ......................... 36  \n4.6 Summary ..................................... 37  \n5 Mixed-Precision Neural Networks 38  \n5.1 Frameworks for MXPDNNs ........................... 40  \n5.1.1 Gradient-Based Optimization ...................... 41  \n5.1","cbCaisIqiAFr8pF8","https://ap.wps.com/l/cbCaisIqiAFr8pF8","pdf",7354607,1,170,"English","en",105,"# Introduction\n## Motivation\n## Contributions\n## Organization\n# In-Memory Computing\n## A Data-Centric Computing Paradigm\n## Content Addressable Memories\n## Associative Processors\n## Two Dimensional Associative Processors\n# Machine Learning Workloads\n## Hyper Dimensional Computing-Based Reinforcement Learning\n## Convolutional Neural Networks\n## Large Language Models\n# Accelerating Hyper Dimensional Computing-Based Reinforcement Learning\n## HDRLPIM\n## Mapping Operations to the Simulator\n## Experimental Setup\n## Results\n## Related Works and Comparison\n## Summary\n# Mixed-Precision Neural Networks\n## Frameworks for MXPDNNs\n## Comparison Among MXPDNN Frameworks\n## Discussion\n# Accelerating Convolutional Neural Networks\n## BF-IMNA\n## Hardware Setup and Benchmarks\n## Results and Discussion\n## Summary\n# Accelerating Large Language Models\n## SoftmAP\n## Experimental Setup\n## Results and Discussion\n## Summary\n# Conclusion and Future Work","[{\"question\":\"What is the main goal of this dissertation?\",\"answer\":\"The dissertation aims to accelerate key machine learning workloads by developing and evaluating in-memory computing approaches based on associative processors and related architectures.\"},{\"question\":\"Which machine learning workloads are addressed?\",\"answer\":\"It covers hyper dimensional computing-based reinforcement learning, convolutional neural networks, and large language model inference, with dedicated accelerator designs and experiments for each.\"},{\"question\":\"How does the work connect hardware acceleration with quantization and precision?\",\"answer\":\"It surveys mixed-precision neural network quantization frameworks and analyzes precision sensitivity through experiments, then evaluates how accelerator designs perform under those precision considerations.\"}]","In-Memory Computing to Accelerate Machine Learning - Dissertation | PDF",1785724486,428,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"in-memory-computing-to-accelerate-machine-learning-dissertation","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/in-memory-computing-to-accelerate-machine-learning-dissertation/119471/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is the main goal of this dissertation?","Question",{"text":75,"@type":76},"The dissertation aims to accelerate key machine learning workloads by developing and evaluating in-memory computing approaches based on associative processors and related architectures.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Which machine learning workloads are addressed?",{"text":80,"@type":76},"It covers hyper dimensional computing-based reinforcement learning, convolutional neural networks, and large language model inference, with dedicated accelerator designs and experiments for each.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the work connect hardware acceleration with quantization and precision?",{"text":84,"@type":76},"It surveys mixed-precision neural network quantization frameworks and analyzes precision sensitivity through experiments, then evaluates how accelerator designs perform under those precision considerations.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]