Podcast
Interpretable Machine Learning: SHAP, Conformal Prediction and Model Trust
Open original DataTalks.Club episode
Interpretable Machine Learning: SHAP, Conformal Prediction and Model Trust
Original Episode
Use these links for the canonical episode and media sources.
- Open the original DataTalks.Club podcast page
- Watch on YouTube
- Listen on Spotify
- Listen on Apple Podcasts
Episode Overview
How can you reliably trust a machine learning model’s predictions in real-world settings? In this episode Christoph Molnar — statistician, machine learner, and author of Interpretable Machine Learning — walks through practical approaches for building model trust. Drawing on his experience from Kaggle competitions to authoring a technical book, Christoph explains the trade-offs between interpretability and accuracy and shows how interpretability techniques help debug models.
People
Use these links to connect the episode to guest notes.
Chapter Summary
Use these checkpoints to decide whether to open the source transcript.
- 0:00 - Podcast Introduction
- 0:42 - Guest Intro: Christoph Molnar, Interpretable ML Author
- 1:32 - Career Journey: From Statistics to Tech Writing
- 3:45 - Becoming a Full-Time Technical Writer
- 6:37 - Kaggle Beginnings: Linear Models to Practical ML
- 7:50 - Origin Story: Interest in Interpretable Machine Learning
- 9:27 - Interpretability vs Accuracy: Debugging Models with SHAP
- 11:59 - Active Competition: River Flow Forecasting Project
- 13:57 - Choosing Book Topics: Audience Data and Personal Curiosity
- 15:55 - Publishing in Public: Chapter-by-Chapter Workflow
- 17:07 - Self-Publishing vs Publishers: Control, Editors, Royalties
- 18:58 - Book Overview: Interpretable ML; Modeling Mindsets; Conformal Prediction;
- 20:27 - Conformal Prediction: Calibrated Uncertainty and Prediction Sets
- 23:44 - SHAP Deep Dive: Practical Guide and Python Examples
- 26:17 - Terminology: Explainable AI vs Interpretable Machine Learning
- 30:00 - Work Style: Solo Writing, Collaboration, and Co-authoring
- 33:07 - Staying Hands-On: Competitions to Maintain Practical Skills
- 36:21 - Logbook Practice: Obsidian Notes for Experiments and Reflection
- 42:21 - Writing Expertise: Teaching to Learn vs Being a Beginner
- 44:51 - Feedback Strategy: Open Drafts, Beta Readers, and Iteration
- 48:36 - Advice for Aspiring Technical Writers: Start Small and Publish
- 50:00 - Becoming a Full-Time Author: Timeframe, Income, and Workload
- 53:49 - Publishing Logistics: Leanpub, Amazon KDP, and Print-on-Demand
- 56:16 - Where to Find Christoph: Website, Newsletter, and Socials