Focused Concept Miner (FCM) : an Interpretable Deep Learning for Text Exploration
Participate
Information Systems and Operations Management
Speaker : Dokyun LEE
Assistant Professor of Business Analytics
Tepper School of Business, Carnegie Mellon University
HEC Campus - Jouy-En-Josas - Buil. V - Room Bernard Ramanantsoa
Abstract :
We introduce the Focused Concept Miner (FCM), an interpretable deep learning text mining algorithm to (1) automatically extract interpretable high-level “concepts” from text data, (2) “focus” the mined concepts to explain any existing user-specified business outcomes, such as conversion (linked to reviews read) or crowdfunding success (linked to project descriptions), and (3) quantify the correlational relative importance of each mined concept for business outcomes, along with their relative importance to other user-specified explanatory variables. Compared to existing methods that partially achieve FCM’s goals, FCM achieves higher interpretability as measured by a variety of metrics (e.g., automated, human-judged) and competitive predictive performance.
Relative importance of discovered concepts provide managers easy ways to gauge potential impact and to inform hypotheses development. We present FCM as a complimentary technique to explore and understand unstructured textual data before applying standard causal inference techniques.
Applications can be found in any settings with text and structured data tied to a business outcome. We evaluate FCM’s performance on two business data in the context of review reading and crowdfunding. Furthermore, we run a series of experiments to investigate the accuracy- interpretability trade-off to provide empirical observations for interpretable machine learning literature. Paper concludes with ideas for future development, potential application scenarios, and managerial implications.
Bio :
Dokyun Lee is an Assistant Professor of Business Analytics at the Tepper School of Business at Carnegie Mellon University. He studies the {application, development, impact} of AI in business and society to extract consumer, market, and managerial insights. His research streams include 1) Interpretable machine learning for business: definition of interpretability and measurement & deep learning based methods for {text, user-behavior} data, 2) Measuring the economic impact of unstructured data (text and images), 3) Unintended consequence of machine learning in business. His research won several academic awards as well as research grants from companies and institutions such as Adobe, NVidia, Marketing Science Institute, NET Institute, and Bosch Institute. He will bring his expertise on deep learning techniques and interpretability.