{"683212":{"#nid":"683212","#data":{"type":"event","title":"PhD Defense | Generalizable, Calibrated and Scalable Time-Series Forecasting in the age of Large-scale Neural Models","body":[{"value":"\u003Cp\u003E\u003Cstrong\u003ETitle\u003C\/strong\u003E:\u0026nbsp;Generalizable, Calibrated and Scalable Time-Series Forecasting in the age of Large-scale Neural Models\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EDate\u003C\/strong\u003E:\u0026nbsp;July 29th, 2025\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003ETime\u003C\/strong\u003E:\u0026nbsp;9:00 AM EST\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003ELocation\u003C\/strong\u003E: Coda C1315\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EZoom Link\u003C\/strong\u003E: \u003Ca href=\u0022https:\/\/gatech.zoom.us\/j\/99980991336?pwd=AFmePvjXJbsn8TCMg41ivtF9QV2tax.1\u0026amp;from=addon\u0022\u003Ehttps:\/\/gatech.zoom.us\/j\/99980991336?pwd=AFmePvjXJbsn8TCMg41ivtF9QV2tax.1\u0026amp;from=addon\u003C\/a\u003E\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EHarshavardhan Kamarthi\u003C\/p\u003E\u003Cp\u003EMachine Learning PhD Student\u003C\/p\u003E\u003Cp\u003ESchool of Computational Science and Engineering\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EGeorgia Institute of Technology\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003ECommittee\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003E1. Dr. B Aditya Prakash (Advisor)\u003C\/p\u003E\u003Cp\u003E2. Dr. Chao Zhang\u003C\/p\u003E\u003Cp\u003E3. Dr. Yao Xie\u003C\/p\u003E\u003Cp\u003E4. Dr. Thomas Ploetz\u003C\/p\u003E\u003Cp\u003E5. Dr. Albert Gu\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EAbstract\u003C\/strong\u003E:\u003C\/p\u003E\u003Cp\u003EDeep learning has significantly improved time\u2011series analytics, but existing models often fail to remain reliable and generalizable across domains, offering poor calibration and struggling to scale to high\u2011dimensional data.\u0026nbsp; This thesis introduces a suite of frameworks that make large time\u2011series models more generalizable, reliable, and scalable.\u0026nbsp; Practitioners will be able to deploy these models in real\u2011time settings, obtain robust calibrated forecasts despite distribution shifts, and reuse the same data\u2011efficient models across diverse domains, even in zero or few\u2011shot scenarios.\u0026nbsp; The thesis first develops non\u2011parametric probabilistic methods that capture and propagate uncertainty across multiple time-series and modalities to deliver dependable prediction intervals.\u0026nbsp;\u0026nbsp;\u003C\/p\u003E\u003Cp\u003ENext, we construct foundational time\u2011series models pre\u2011trained on heterogeneous datasets using novel supervised pre\u2011training and semantic tokenization strategies, providing reliable performance across domains, and minimizing the need for task\u2011specific fine\u2011tuning.\u0026nbsp; Building on these advances, this thesis scales solutions to applications involving thousands of interrelated series, supporting hierarchical and multivariate forecasting while retaining performance and reliability.\u0026nbsp; Extensive benchmark tests and industrial deployments validate our approach, pushing the field toward more trustworthy, scalable, and broadly applicable time\u2011series forecasting systems.\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003EGeneralizable, Calibrated and Scalable Time-Series Forecasting in the age of Large-scale Neural Models\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"Harshavardhan Kamarthi - Machine Learning PhD Student - School of Computational Science and Engineering "}],"uid":"36518","created_gmt":"2025-07-22 14:09:36","changed_gmt":"2025-07-22 14:09:36","author":"shatcher8","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2025-07-29T09:00:00-04:00","event_time_end":"2025-07-29T11:00:00-04:00","event_time_end_last":"2025-07-29T11:00:00-04:00","gmt_time_start":"2025-07-29 13:00:00","gmt_time_end":"2025-07-29 15:00:00","gmt_time_end_last":"2025-07-29 15:00:00","rrule":null,"timezone":"America\/New_York"},"location":"Coda C1315","extras":[],"groups":[{"id":"576481","name":"ML@GT"}],"categories":[],"keywords":[],"core_research_areas":[],"news_room_topics":[],"event_categories":[],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[],"email":[],"slides":[],"orientation":[],"userdata":""}}}