Center for Alaska Native Health Research
2141 Koyukuk Drive
205 Arctic Health Research Building
Fairbanks, AK 99775-7000

Phone: 907-474-5528

1-888-470-5576 toll-free within Alaska

Fax: 907-474-5700

uaf-iab-canhr@alaska.edu

From Points and Lines to 'Trees', Shallow and Deep Learning

 
 

 

A Primer on Machine Learning(ML)/Artificial Intelligence (AI), concepts, and terminology

About this presentation

Machine Learning and Artificial Intelligence (ML/AI) have been used since the 1950s when Arthur Samuel of IBM created a computer program for playing checkers. Since then, ML/AI applications have evolved a good bit. Still, there are some base structures and milestone events that form the basis of good ML/AI applications and their software.

Falk Huettmann, Ph.D., MBA, Professor - EWHALE Lab- Institute of Arctic Biology, Biology & Wildlife Department, University of Alaska Fairbanks, will present in lay terms for a general audience the basic terminology, key concepts, and algorithms for how a cloud of points can be analyzed with ML/AI; an emphasize is made for uses in healthcare.

This presentation will cover linear and non-linear applications, regression and classification trees, boosting and bagging (bootstrap aggregation; Random Forest), maximum entropy, Vector Support Machines, Neural Networks, Ensemble Models, and wider 'The Cloud' workflow applications with MS Azure and ORACLE. Dr. Huettmann will address research design, significances, spatial and temporal applications, model selection, multivariate predictors, data mining, and 'inference from predictions' (Leo Breiman) as well as why 'many weak learners make for a strong learner' (J. Friedman). Some software examples in commercial tools, R and python are briefly mentioned emphasizing the need for transparent and repeatable research and associated open-access data.

 

 

 

The Presenter
Falk Huettmann is a digital naturalist. With an M.Sc. (Germany), Ph.D. (Canada), and MBA (UAF) he works as a Wildlife Ecologist worldwide, on all continents, with a strong focus on global research data networks, and polar regions connecting wider health and sustainability with landscapes, people, the tropics, oceans, and atmosphere. His three-decade-long efforts combining remote fieldwork with open source geographic information systems (GIS), computing, and Machine Learning have resulted in over 300 publications, many 'massive' open-access data sets, and 9 books, including a textbook on Machine Learning/AI and Open Access, as well as projects on human health and zoonotic diseases. Falk reviews for over 60 journals and publishers. His research and teaching was already awarded by Quality Matters (QM), National Geographic, Killam Foundation (Canada), NIH/NIAID, NSF, and various NGOs, agencies, and governments, e.g. WWF, the EU Parliament, Environment Canada, Dept of Fisheries and Oceans (Canada), U.S. Fish & Wildlife Service, and the Global Environmental Fund (GEF).