{"652519":{"#nid":"652519","#data":{"type":"event","title":"Ph.D. Proposal Oral Exam - Foroozan Karimzadeh","body":[{"value":"\u003Cp\u003E\u003Cstrong\u003ETitle:\u0026nbsp; \u003C\/strong\u003E\u003Cem\u003EHardware-Friendly Model Compression for DNN Accelerators\u003C\/em\u003E\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u003Cstrong\u003ECommittee:\u0026nbsp; \u003C\/strong\u003E\u003C\/p\u003E\r\n\r\n\u003Cp\u003EDr. Raychowdhury, Advisor\u0026nbsp;\u0026nbsp;\u0026nbsp;\u0026nbsp;\u003C\/p\u003E\r\n\r\n\u003Cp\u003EDr. Yu, Chair\u003C\/p\u003E\r\n\r\n\u003Cp\u003EDr. Romberg\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u003Cstrong\u003EAbstract: \u003C\/strong\u003EThe objective of the proposed research is to introduce solutions to make powerful Deep Neural Network, DNN, algorithms to be deployable on edge devices through developing hardware-aware DNN compression methods. The rising popularity of intelligent mobile devices and the computational cost of deep learning-based models call for efficient and accurate on-device inference schemes. In particular, we proposed two compression techniques. In the first method, LGPS, we present a hardware-aware pruning method where the locations of non-zero weights are derived in real-time from a LFSR. Using the pro-posed method, we demonstrate a total saving of energy and area up to 63.96% and 64.23%for VGG-16 network on down-sampled ImageNet, respectively for iso-compression-rate and iso-accuracy. Secondly, we propose a novel model compression scheme that allows inference to be carried out using bit-level sparsity, which can be efficiently implemented using in-memory computing macros. We introduce a method called BitS-Net to leverage the benefits of bit-sparsity (where the number of zeros is more than number of ones in binary representation of weight\/activation values) when applied to Compute-In-Memory(CIM) with Resistive Random-Access Memory (RRAM) to develop energy efficient DNN accelerators operating in the inference mode. We demonstrate that BitS-Net improves the energy efficiency by up to 5x for ResNet models on the ImageNet dataset.\u003C\/p\u003E\r\n","summary":null,"format":"limited_html"}],"field_subtitle":"","field_summary":"","field_summary_sentence":[{"value":"Hardware-Friendly Model Compression for DNN Accelerators"}],"uid":"28475","created_gmt":"2021-11-05 21:18:03","changed_gmt":"2021-11-05 21:18:03","author":"Daniela Staiculescu","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2021-11-09T10:00:00-05:00","event_time_end":"2021-11-09T12:00:00-05:00","event_time_end_last":"2021-11-09T12:00:00-05:00","gmt_time_start":"2021-11-09 15:00:00","gmt_time_end":"2021-11-09 17:00:00","gmt_time_end_last":"2021-11-09 17:00:00","rrule":null,"timezone":"America\/New_York"},"extras":[],"groups":[{"id":"434371","name":"ECE Ph.D. Proposal Oral Exams"}],"categories":[],"keywords":[{"id":"102851","name":"Phd proposal"},{"id":"1808","name":"graduate students"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1788","name":"Other\/Miscellaneous"}],"invited_audience":[{"id":"78771","name":"Public"}],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[],"email":[],"slides":[],"orientation":[],"userdata":""}}}