{"686127":{"#nid":"686127","#data":{"type":"event","title":"Ph.D. Proposal Oral Exam - Payman Behnam","body":[{"value":"\u003Cp\u003E\u003Cstrong\u003ETitle:\u0026nbsp; \u003C\/strong\u003E\u003Cem\u003EFrom Neurons to Nodes to Tokens: Model-System-Hardware Tri-Design Optimization for Efficient Machine Learning\u003C\/em\u003E\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003ECommittee:\u0026nbsp;\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003EDr.\u0026nbsp;Tumanov, Advisor\u0026nbsp;\u0026nbsp;\u0026nbsp;\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EDr. Krishna, Chair\u003C\/p\u003E\u003Cp\u003EDr. Gavrilovska\u003C\/p\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003EThe objective of the proposed research is to claim that the coordinated model\u2013system\u2013hardware tri-design is the right lever to achieve Efficient ML across CNNs, GNNs, and LLMs. This dissertation substantiates this claim with advances at three granularities, (i) single-layer methods that compress, quantize, and distill models to achieve lower compute and memory footprint, (ii) two-layer co-designs that translate algorithmic structure into runtime scheduling and hardware primitives to reduce latency and data movement, and raise on-chip utilization (model\u2013system, model\u2013hardware, system-hardware), and (iii) full model\u2013system\u2013hardware integration that combines innovations in all design abstraction layers in end-to-end pipelines, thereby reducing the dominant costs of data movement, numeric precision limits, sparsity, and real-time serving into measurable gains, which eventually results in improving the accuracy\u2013latency\u2013energy Pareto frontier.\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"From Neurons to Nodes to Tokens: Model-System-Hardware Tri-Design Optimization for Efficient Machine Learning"}],"uid":"28475","created_gmt":"2025-10-31 21:16:46","changed_gmt":"2025-10-31 21:18:08","author":"Daniela Staiculescu","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2025-11-10T14:00:00-05:00","event_time_end":"2025-11-10T16:00:00-05:00","event_time_end_last":"2025-11-10T16:00:00-05:00","gmt_time_start":"2025-11-10 19:00:00","gmt_time_end":"2025-11-10 21:00:00","gmt_time_end_last":"2025-11-10 21:00:00","rrule":null,"timezone":"America\/New_York"},"location":"Room 3126, Klaus ","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":""}}}