{"678746":{"#nid":"678746","#data":{"type":"news","title":"Multipurpose Model Enhances Forecasting Across Epidemics, Energy, and Economics","body":[{"value":"\u003Cp\u003EA new machine learning (ML) model from Georgia Tech could protect communities from diseases, better manage electricity consumption in cities, and promote business growth, all at the same time.\u003C\/p\u003E\u003Cp\u003EResearchers from the School of Computational Science and Engineering (CSE) created the Large Pre-Trained Time-Series Model (LPTM) framework.\u0026nbsp;\u003Ca href=\u0022https:\/\/arxiv.org\/abs\/2311.11413\u0022\u003E\u003Cstrong\u003ELPTM\u003C\/strong\u003E\u003C\/a\u003E is a single foundational model that completes forecasting tasks across a broad range of domains.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EAlong with performing as well or better than models purpose-built for their applications, LPTM requires 40% less data and 50% less training time than current baselines. In some cases, LPTM can be deployed without any training data.\u003C\/p\u003E\u003Cp\u003EThe key to LPTM is that it is pre-trained on datasets from different industries like healthcare, transportation, and energy. The Georgia Tech group created an adaptive segmentation module to make effective use of these vastly different datasets.\u003C\/p\u003E\u003Cp\u003EThe Georgia Tech researchers will present LPTM in Vancouver, British Columbia, Canada, at the 2024 Conference on Neural Information Processing Systems (\u003Ca href=\u0022https:\/\/nips.cc\/\u0022\u003E\u003Cstrong\u003ENeurIPS 2024\u003C\/strong\u003E\u003C\/a\u003E). NeurIPS is one of the world\u2019s most prestigious conferences on artificial intelligence (AI) and ML research.\u003C\/p\u003E\u003Cp\u003E\u201cThe foundational model paradigm started with text and image, but people haven\u2019t explored time-series tasks yet because those were considered too diverse across domains,\u201d said\u0026nbsp;\u003Ca href=\u0022https:\/\/faculty.cc.gatech.edu\/~badityap\/\u0022\u003E\u003Cstrong\u003EB. Aditya Prakash\u003C\/strong\u003E\u003C\/a\u003E, one of LPTM\u2019s developers.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u201cOur work is a pioneer in this new area of exploration where only few attempts have been made so far.\u201d\u003C\/p\u003E\u003Cp\u003E[\u003Ca href=\u0022https:\/\/sites.gatech.edu\/research\/neurips-2024\/\u0022\u003E\u003Cstrong\u003EMICROSITE: Georgia Tech at NeurIPS 2024\u003C\/strong\u003E\u003C\/a\u003E]\u003C\/p\u003E\u003Cp\u003EFoundational models are trained with data from different fields, making them powerful tools when assigned tasks. Foundational models drive GPT, DALL-E, and other popular generative AI platforms used today. LPTM is different though because it is geared toward time-series, not text and image generation. \u0026nbsp;\u003C\/p\u003E\u003Cp\u003EThe Georgia Tech researchers trained LPTM on data ranging from epidemics, macroeconomics, power consumption, traffic and transportation, stock markets, and human motion and behavioral datasets.\u003C\/p\u003E\u003Cp\u003EAfter training, the group pitted LPTM against 17 other models to make forecasts as close to nine real-case benchmarks. LPTM performed the best on five datasets and placed second on the other four.\u003C\/p\u003E\u003Cp\u003EThe nine benchmarks contained data from real-world collections. These included the spread of influenza in the U.S. and Japan, electricity, traffic, and taxi demand in New York, and financial markets.\u0026nbsp; \u0026nbsp;\u003C\/p\u003E\u003Cp\u003EThe competitor models were purpose-built for their fields. While each model performed well on one or two benchmarks closest to its designed purpose, the models ranked in the middle or bottom on others.\u003C\/p\u003E\u003Cp\u003EIn another experiment, the Georgia Tech group tested LPTM against seven baseline models on the same nine benchmarks in zero-shot forecasting tasks. Zero-shot means the model is used out of the box and not given any specific guidance during training. LPTM outperformed every model across all benchmarks in this trial.\u003C\/p\u003E\u003Cp\u003ELPTM performed consistently as a top-runner on all nine benchmarks, demonstrating the model\u2019s potential to achieve superior forecasting results across multiple applications with less and resources.\u003C\/p\u003E\u003Cp\u003E\u201cOur model also goes beyond forecasting and helps accomplish other tasks,\u201d said Prakash, an associate professor in the School of CSE.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u201cClassification is a useful time-series task that allows us to understand the nature of the time-series and label whether that time-series is something we understand or is new.\u201d\u003C\/p\u003E\u003Cp\u003EOne reason traditional models are custom-built to their purpose is that fields differ in reporting frequency and trends.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EFor example, epidemic data is often reported weekly and goes through seasonal peaks with occasional outbreaks. Economic data is captured quarterly and typically remains consistent and monotone over time.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003ELPTM\u2019s adaptive segmentation module allows it to overcome these timing differences across datasets. When LPTM receives a dataset, the module breaks data into segments of different sizes. Then, it scores all possible ways to segment data and chooses the easiest segment from which to learn useful patterns.\u003C\/p\u003E\u003Cp\u003ELPTM\u2019s performance, enhanced through the innovation of adaptive segmentation, earned the model acceptance to NeurIPS 2024 for presentation. NeurIPS is one of three primary international conferences on high-impact research in AI and ML. NeurIPS 2024 occurs Dec. 10-15.\u003C\/p\u003E\u003Cp\u003EPh.D. student\u0026nbsp;\u003Ca href=\u0022https:\/\/www.harsha-pk.com\/\u0022\u003E\u003Cstrong\u003EHarshavardhan Kamarthi\u003C\/strong\u003E\u003C\/a\u003E partnered with Prakash, his advisor, on LPTM. The duo are among the 162 Georgia Tech researchers presenting over 80 papers at the conference.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EPrakash is one of 46 Georgia Tech faculty with research accepted at NeurIPS 2024. Nine School of CSE faculty members, nearly one-third of the body, are authors or co-authors of 17 papers accepted at the conference.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EAlong with sharing their research at NeurIPS 2024, Prakash and Kamarthi released an\u0026nbsp;\u003Ca href=\u0022https:\/\/github.com\/AdityaLab\/Samay\u0022\u003E\u003Cstrong\u003Eopen-source library of foundational time-series modules\u003C\/strong\u003E\u003C\/a\u003E that data scientists can use in their applications.\u003C\/p\u003E\u003Cp\u003E\u201cGiven the interest in AI from all walks of life, including business, social, and research and development sectors, a lot of work has been done and thousands of strong papers are submitted to the main AI conferences,\u201d Prakash said.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u201cAcceptance of our paper speaks to the quality of the work and its potential to advance foundational methodology, and we hope to share that with a larger audience.\u201d\u003C\/p\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003EA new machine learning (ML) model from Georgia Tech could protect communities from diseases, better manage electricity consumption in cities, and promote business growth, all at the same time.\u003C\/p\u003E\u003Cp\u003EResearchers from the School of Computational Science and Engineering (CSE) created the Large Pre-Trained Time-Series Model (LPTM) framework.\u0026nbsp;\u003Ca href=\u0022https:\/\/arxiv.org\/abs\/2311.11413\u0022\u003E\u003Cstrong\u003ELPTM\u003C\/strong\u003E\u003C\/a\u003E is a single foundational model that completes forecasting tasks across a broad range of domains.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EAlong with performing as well or better than models purpose-built for their applications, LPTM requires 40% less data and 50% less training time than current baselines. In some cases, LPTM can be deployed without any training data.\u003C\/p\u003E\u003Cp\u003EThe key to LPTM is that it is pre-trained on datasets from different industries like healthcare, transportation, and energy. The Georgia Tech group created an adaptive segmentation module to make effective use of these vastly different datasets.\u003C\/p\u003E\u003Cp\u003EThe Georgia Tech researchers will present LPTM in Vancouver, British Columbia, Canada, at the 2024 Conference on Neural Information Processing Systems (\u003Ca href=\u0022https:\/\/nips.cc\/\u0022\u003E\u003Cstrong\u003ENeurIPS 2024\u003C\/strong\u003E\u003C\/a\u003E). NeurIPS is one of the world\u2019s most prestigious conferences on artificial intelligence (AI) and ML research.\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"The Large Pre-Trained Time-Series Model (LPTM) framework completes forecasting tasks across a broad range of domains, outperforms current models,  and requires 40% less data and 50% less training time than current baselines."}],"uid":"36319","created_gmt":"2024-12-04 12:32:04","changed_gmt":"2024-12-05 20:53:31","author":"Bryant Wine","boilerplate_text":"","field_publication":"","field_article_url":"","location":"Atlanta, GA","dateline":{"date":"2024-12-03T00:00:00-05:00","iso_date":"2024-12-03T00:00:00-05:00","tz":"America\/New_York"},"extras":[],"hg_media":{"675764":{"id":"675764","type":"image","title":"LPTM Head photo.jpg","body":null,"created":"1733315535","gmt_created":"2024-12-04 12:32:15","changed":"1733315535","gmt_changed":"2024-12-04 12:32:15","alt":"CSE NeurIPS 2024","file":{"fid":"259428","name":"LPTM Head photo.jpg","image_path":"\/sites\/default\/files\/2024\/12\/04\/LPTM%20Head%20photo.jpg","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2024\/12\/04\/LPTM%20Head%20photo.jpg","mime":"image\/jpeg","size":138121,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2024\/12\/04\/LPTM%20Head%20photo.jpg?itok=-_oqygAy"}},"675765":{"id":"675765","type":"image","title":"Aditya and Harsha.jpg","body":null,"created":"1733315572","gmt_created":"2024-12-04 12:32:52","changed":"1733315572","gmt_changed":"2024-12-04 12:32:52","alt":"CSE NeurIPS 2024","file":{"fid":"259429","name":"Aditya and Harsha.jpg","image_path":"\/sites\/default\/files\/2024\/12\/04\/Aditya%20and%20Harsha.jpg","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2024\/12\/04\/Aditya%20and%20Harsha.jpg","mime":"image\/jpeg","size":54358,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2024\/12\/04\/Aditya%20and%20Harsha.jpg?itok=Dv3sFphr"}}},"media_ids":["675764","675765"],"related_links":[{"url":"https:\/\/www.cc.gatech.edu\/news\/multipurpose-model-enhances-forecasting-across-epidemics-energy-and-economics","title":"Multipurpose Model Enhances Forecasting Across Epidemics, Energy, and Economics"}],"groups":[{"id":"47223","name":"College of Computing"},{"id":"1188","name":"Research Horizons"},{"id":"50877","name":"School of Computational Science and Engineering"}],"categories":[{"id":"138","name":"Biotechnology, Health, Bioengineering, Genetics"},{"id":"139","name":"Business"},{"id":"142","name":"City Planning, Transportation, and Urban Growth"},{"id":"42901","name":"Community"},{"id":"153","name":"Computer Science\/Information Technology and Security"},{"id":"131","name":"Economic Development and Policy"},{"id":"144","name":"Energy"},{"id":"146","name":"Life Sciences and Biology"},{"id":"135","name":"Research"},{"id":"134","name":"Student and Faculty"},{"id":"8862","name":"Student Research"}],"keywords":[{"id":"10199","name":"Daily Digest"},{"id":"9153","name":"Research Horizons"},{"id":"187915","name":"go-researchnews"},{"id":"192863","name":"go-ai"},{"id":"654","name":"College of Computing"},{"id":"166983","name":"School of Computational Science and Engineering"},{"id":"2556","name":"artificial intelligence"},{"id":"9167","name":"machine learning"},{"id":"191912","name":"Data Science at GT"}],"core_research_areas":[{"id":"193655","name":"Artificial Intelligence at Georgia Tech"},{"id":"39441","name":"Bioengineering and Bioscience"},{"id":"39431","name":"Data Engineering and Science"}],"news_room_topics":[],"event_categories":[],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003EBryant Wine, Communications Officer\u003Cbr\u003E\u003Ca href=\u0022mailto:bryant.wine@cc.gatech.edu\u0022\u003Ebryant.wine@cc.gatech.edu\u003C\/a\u003E\u003C\/p\u003E","format":"limited_html"}],"email":[],"slides":[],"orientation":[],"userdata":""}}}