NSF I-GUIDE

🚀 Registration for the NSF I-GUIDE Forum is Now Open! 🚀 NSF I-GUIDE is thrilled to announce that registration is officially open for the I-GUIDE Forum 2024: Convergence Science and Geospatial AI for Environmental Sustainability! Join us from October 14-16, 2024, at Snow King Mountain Resort in Jackson, Wyoming, for a unique opportunity to engage with top experts and innovators in the fields of geospatial AI, convergence science, and environmental sustainability. 🌍 Secure your spot now and take advantage of early bird pricing available until September 14, 2024. With options to attend all events, specific days, or just the tutorials, there’s something for everyone - whether you’re a student or a seasoned professional. 👉 Register Here: https://lnkd.in/eSxkk6mM 👉 Forum 2024 Information: https://lnkd.in/eGvryzPD

🌟 New conference paper by the NSF I-GUIDE team! "OntoType: Ontology-Guided and Pre-Trained Language Model Assisted Fine-Grained Entity Typing" presented at the 2024 ACM, Association for Computing Machinery Knowledge Discovery and Data Mining (SIGKDD) Conference. By Authors: Tanay Komarlu of University of Illinois Urbana-Champaign Minhao Jiang of University of Illinois Urbana-Champaign Xuan Wang of Virginia Tech Jiawei Han of University of Illinois Urbana-Champaign Read the paper here 👉 https://lnkd.in/gmPh5iav This paper introduces OntoType, a novel, annotation-free method for fine-grained entity typing (FET) that leverages an ontology-guided approach to improve accuracy. Unlike traditional methods that rely on costly human-annotated corpora, OntoType uses pre-trained language models (PLMs) to generate type candidates and refines them based on a hierarchical ontological structure, ensuring context-sensitive and precise typing. Experiments on several datasets demonstrate that OntoType outperforms state-of-the-art zero-shot FET methods, including ChatGPT, highlighting the potential for further improvement by refining existing ontologies. Read the paper here 👉 https://lnkd.in/gmPh5iav Abstract Fine-grained entity typing (FET), which assigns entities in text with context-sensitive, fine-grained semantic types, is a basic but important task for knowledge extraction from unstructured text. FET has been studied extensively in natural language processing and typically relies on human-annotated corpora for training, which is costly and difficult to scale. Recent studies explore the utilization of pre-trained language models (PLMs) as a knowledge base to generate rich and context-aware weak supervision for FET. However, a PLM still requires direction and guidance to serve as a knowledge base as they often generate a mixture of rough and fine-grained types, or tokens unsuitable for typing. In this study, we vision that an ontology provides a semantics-rich, hierarchical structure, which will help select the best results generated by multiple PLM models and head words. Specifically, we propose a novel annotation-free, ontology-guided FET method, OntoType, which follows a type ontological structure, from coarse to fine, ensembles multiple PLM prompting results to generate a set of type candidates, and refines its type resolution, under the local context with a natural language inference model. Our experiments on the Ontonotes, FIGER, and NYT datasets using their associated ontological structures demonstrate that our method outperforms the state-of-the-art zero-shot fine-grained entity typing methods as well as a typical LLM method, ChatGPT. Our error analysis shows that refinement of the existing ontology structures will further improve fine-grained entity typing.

Happy to share that I'll be speaking on “GeoAI Applications to Predict Field Scale Actual Evapotranspiration!” at the NSF I-GUIDE VCO webinar series! Join the conversation on October 23!

At the next installment of the NSF I-GUIDE Virtual Consulting Office (VCO) webinar series join participants from the 2024 I-GUIDE Summer School as they present their project work on "GeoAI Applications to Predict Field Scale Actual Evapotranspiration." Coming up Wednesday, October 23 at 11:00am CT. 🔗 Registration and Information: https://lnkd.in/gv7fKJW4 Speakers Include: Laurel DiSera of Columbia University Eshita Eva of The Ohio State University Ehsan Foroutan of Oklahoma State University Kelechi Igwe of Kansas State University Manoj Lamichhane of South Dakota State University Sushant Mehan of South Dakota State University Musab Waqar of Virginia Tech and Aleksandra W. of Kent State University 🔗 Registration and Information: https://lnkd.in/gv7fKJW4 In water-scarce regions, effective agricultural water management hinges on accurate evapotranspiration (ETa) estimation. Traditional methods and satellite products often fall short due to spatial biases and limited ground representation. In this talk, we will present our efforts to develop a data-driven model that leverages machine learning techniques to predict daily reference evapotranspiration at a finer spatial resolution. This approach aims to optimize water use in semi-arid regions, ensuring sustainable crop production. Join us for a detailed exploration of the model development process, its applications, and implications for agricultural water management.

This week join Anand P. of the University of Illinois Urbana-Champaign for a free introductory session to the I-GUIDE Platform! Coming up Wednesday, October 2 at 11:00am CT - part of the NSF I-GUIDE Virtual Consulting Office (VCO) webinar series! 🔗 Registration and Information: https://lnkd.in/gbnnW26c

I'm excited to announce the publication of our latest paper, "PRIME: A CyberGIS Platform for Disaster Resilience Assessment and Enhancement," in Computers, Environment and Urban Systems! The 50-day free access link is https://lnkd.in/gSkhnppb. Congratulations to my student and PhD candidate Debayan Mandal for leading this impactful research 🎉 🎉 🎉 This work introduces a customizable resilience inference model (RIM) for evaluating and improving community resilience to natural disasters. Advanced machine learning approaches, such as Bayesian Networks, are introduced to identify the causal relationships between socioeconomic factors and resilience capacities. Based on this research, Debayan has received several prestigious awards, including 2nd place in the 2023 Texas GIS Day Graduate Student Competition, the 2024 NSF I-GUIDE Rising Geospatial Data Scientist Award, and the 2024 Mr. and Mrs. Kenneth P. Pipes Endowed Scholarship in Geosciences at Texas A&M College of Arts & Sciences! I’m incredibly proud of Debayan's dedication and grateful for the support from the collaborators, including Dr. Rohan Singh Wilkho, Dr. Furqan Baig, Joynal Abedin 🌎 Dr. Bing Zhou, Dr. Heng Cai, Dr. Nasir Gharaibeh, and Dr. Nina Lam! Feel free to check out the full paper and let us know your thoughts! We appreciate and welcome any suggestions and feedback! #GIScience #GearLab

🌟 New conference tutorial by the NSF I-GUIDE team! "Automated Mining of Structured Knowledge from Text in the Era of Large Language Models" by Ming Zhong, Siru Ouyang, Yizhu Jiao, Sizhe Zhou, Linyi Ding, and Jiawei Han of the University of Illinois Urbana-Champaign and presented at the 2024 ACM, Association for Computing Machinery Knowledge Discovery and Data Mining (SIGKDD) Conference. Read the tutorial paper here 👉 https://lnkd.in/gZhgn9fA This tutorial explores recent advancements in using large language models (LLMs) for mining structured knowledge from massive unstructured text data with minimal supervision. It covers topics including an introduction to LLMs, automated ontology construction, and weakly-supervised techniques for text classification and information extraction. The focus is on leveraging LLMs effectively without the need for extensive human-annotated training data. Read the tutorial paper here 👉 https://lnkd.in/gZhgn9fA Abstract Massive amount of unstructured text data are generated daily, ranging from news articles to scientific papers. How to mine structured knowledge from the text data remains a crucial research question. Recently, large language models (LLMs) have shed light on the text mining field with their superior text understanding and instruction-following ability. There are typically two ways of utilizing LLMs: fine-tune the LLMs with human-annotated training data, which is labor intensive and hard to scale; prompt the LLMs in a zero-shot or few-shot way, which cannot take advantage of the useful information in the massive text data. Therefore, it remains a challenge on automated mining of structured knowledge from massive text data in the era of large language models. In this tutorial, we cover the recent advancements in mining structured knowledge using language models with very weak supervision. We will introduce the following topics in this tutorial: (1) introduction to large language models, which serves as the foundation for recent text mining tasks, (2) ontology construction, which automatically enriches an ontology from a massive corpus, (3) weakly-supervised text classification in flat and hierarchical label space, (4) weakly-supervised information extraction, which extracts entity and relation structures.

I'm looking forward to running a workshop at the NSF I-GUIDE Forum on  Convergence Science and Geospatial AI for Environmental Sustainability in October as part of the build-up to the Taylor Geospatial Institute's Food Security Challenge! I'll be talking (a little) and listening (a lot) to the community's ideas about how #geospatial technologies can contribute to improving #foodsecurity. I was lucky to work with several members of the i-guide community last year and I know they will bring their knack for #convergence thinking and their computing and analytics talent to the table. You can read about the Challenge and learn how to get involved here: https://lnkd.in/ed9X3rZk And if you're lucky enough to be headed to the i-guide forum, Mark Korver and I hope to see you at the workshop! You can still register at: https://lnkd.in/e5WAyHQY if you're not yet signed up. Anand P. and Shaowen Wang - thanks for the opportunity!

This week I represented the NSF I-GUIDE Institite in the NSF HDR Ecosystem Conference. My comments in a panel identified HDR research in AI and DS relevant to development of the geospatial and Spatial Web standards: - Expanding Space into Hyperspace:  HDR results on High-dimensional spaces, Graph networks, and  Knowledge Representation (KGML) support the IEEE Spatial Web standard for Hyperspace as a common interface across various forms of knowledge representation.  - AI agents dynamic activities: HDR results in representing dynamic entities including behavior trait labeling supports the IEEE Spatial Web standard ontology for autonomous agents. https://lnkd.in/eaFBcyKt