Natural Language Processing Template Engine for Data Science and Machine Learning

Explore a presentation on completing computational templates using parameters extracted from text specifications through a question answering system. Learn about the general method and see demonstrations of various computational workflows, including classification, latent semantic analysis, quantile regression, random data generation, and recommendations. Discover how to leverage the engine by bringing your own templates. Focus on data science and machine learning computations while understanding that the described method and overall algorithm can be applied to workflows from other fields. Delve into topics such as problem formulation, interface usage, workflow classifiers, grammars, prefix trees, and querying the system. Gain insights into semantic analysis and quantile regression workflows, with opportunities for questions at the end.

Introduction
Agenda
Problem formulation
Question answering system
Interface
Wolfram
Titanic
How is this happening
What are the raw answers
What do you see
Extract
Quantile Regression
Quantitative Regression
Workflow Classifier
Default Classifier
Send Mail Classifier
Wrong Results
Grammars
Prefix Trees
Alternative
What are the templates
Querying the system
Semantic analysis workflow
Quantile regression workflow
Questions