Cryospheric and Polar Processes Seminar
Applications of Machine Learning and Natural Language Processing in the Polar Deep Insights Project by Dr. Siri Jodha Khalsa, NSIDC research scientist.
ABSTRACT: Polar Deep Insights, an NSF-Funded EarthCube project that is building an end-to-end system to collect, analyze and make interactive the wealth of polar-related textual and scientific data mined from the Web. Researchers apply their domain knowledge to train machine-learning (ML) models that power an intuitive search engine to be utilized for polar research. The system includes an interface to analyze and visualize results using Banana and D3.js, giving researchers a better understanding of the relationship of the discovered content within the Polar data ecosystem. This talk will provide a brief introduction to the tools and techniques of information retrieval and data science that we are being developed and applied: Apache Sparkler, Tika, and Solr; ML models (SVM, MLP ANN, Random Forest and Naive Bayes classifiers); Gensim and spaCy libraries for Natural Language Processing; and a concept editor that allows users to build an 'ontology-of-interest' to derive insights using the Insight Visualizer. https://nsidc.org/research/bios/khalsa.html
TO JOIN BY ZOOM: