Mark Zangari, CEO, Quantellia @ MLconf SEA
summary
Presents Agency Theory — a math framework that uses entropy-style measures to find which levers and datapoints actually change outcomes, claiming organizations can find the 20% of data that yields 80% of value and avoid billions wasted on irrelevant data. Combines simple system models and embedded dynamics with standard machine learning (training clouds, probability-of-default surfaces) to decompose decisions (example: small‑business lending), and reports practical success where blind ML fails.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 5:45Why is ML underutilized
- 11:35Four Pillars
- 13:27Organizational Trust
- 15:15Consumer Credit
- 16:30Credit History
- 17:34Commercial Credit
- 18:47Small Medium Businesses
- 21:22Application Process
- 23:40Outcomes
- 24:54Probability of Default
- 26:51Surface
- 27:33Solution
related talks
18mLorien Pratt, Co-Founder & Chief Scientist, Quantellia
32mMLDS 2019: Pradeep Gulipalli of Tiger Analytics at Machine Learning Developers Summit 2019
33mQuantamental Investing - the future of finance
04mJustin Basilico interview
32mMaking Models Work: How an optimized ML lifecycle can drive better ROI
31mMLDS 2019: Indranath Mukherjee of AXA XL at Machine Learning Developers Summit 2019
45mMixed Model Arts
17mKurt Muehmel (Dataiku) interviewed at the O'Reilly Artificial Intelligence Conference NY 2019