Interpretable Machine Learning - Fairness, Accountability and Transparency in ML systems



The good news is building fair, accountable, and transparent machine learning systems is possible. The bad news is it’s harder than many blogs and software package docs would have you believe. The truth is nearly all interpretable machine learning techniques generate approximate explanations, that the fields of eXplainable AI (XAI) and Fairness, Accountability, and Transparency in Machine Learning (FAT/ML) are very new, and that few best practices have been widely agreed upon. This combination can lead to some ugly outcomes!

This talk aims to make your interpretable machine learning project a success by describing fundamental technical challenges you will face in building an interpretable machine learning system, defining the real-world value proposition of approximate explanations for exact models, and then outlining the following viable techniques for debugging, explaining, and testing machine learning models:

  • Model visualizations including decision tree surrogate models, individual conditional expectation (ICE) plots, partial dependence plots, and residual analysis.
  • Reason code generation techniques like LIME, Shapley explanations, and Tree-interpreter.
  • Sensitivity Analysis.
  • Plenty of guidance on when, and when not, to use these techniques will also be shared.

The talk will conclude by providing guidelines for testing generated explanations themselves for accuracy and stability.


Basic familiarity with machine learning concepts.

Content URLs:

Demo deck is available here (not yet complete, though). A demo repository can be found here:

This proposal is also accepted at ODSC West this year and a short blog piece can be found here.

Outline of the talk:

  • What is Machine Learning Interpretability? [3 Mins]
  • Why should you care about Machine Learning Interpretability? [3 Mins]
  • Why is Machine Learning Interpretability difficult? [3 Mins]
  • What is the Value Proposition of Machine Learning Interpretability? [10 Mins]
  • How can Machine Learning Interpretability be practiced? (several examples) [10Mins]
  • Can Machine Learning Interpretability be tested? (General recommendations Tool-based observations) [10 Mins]
  • Q & A [6 Mins]

Preview video:

And this is the learning outcome from this talk: [HTML_REMOVED] By the end of the session, the attendees will have a clear idea of the importance of fairness, accountability and transparency in machine learning and how it stands up in real-world scenarios. They will also get to see some real examples justifying the importance of interpretability of ML systems. They will get know about some of the tools that are used in this regard (such as LIME, Shapley etc.).

My previous slide decks can be checked here:

I blog on a daily basis. All of my blogs can be found here:

Speaker Info:

I am currently with PyImageSearch where I write code for our articles, books and courses. I am also responsible for providing Q&A support to PyImageSearch readers.

Previously at DataCamp, I developed projects for DataCamp Project. My DataCamp project Predicting Credit Card Approvals is now launched and so is my DataCamp practice pool Advanced Deep Learning with Keras in Python (I created exercises for DataCamp Practice too). I also write technical tutorials for DataCamp Community on a daily basis. Prior to DataCamp, I have worked at TCS Research and Innovation (TRDDC) as a developer where the domain of work was Cyber Security (specifically Data Privacy). There, I was a part of TCS's critically acclaimed GDPR solution called Crystal Ball. Prior to that, I have worked as a Web Services Developer at TCS (Kolkata area).

I am also working with Dr. Anupam Ghosh and my beloved college juniors for Applied Machine Learning research/tinkering. Currently, we are working on the application of machine learning in Phonocardiogram classification.

Two accolades I take pride in:

My subject of interest broadly lies in areas like Machine Learning Interpretability, Full-Stack Data Science.

Id: 1048
Section: Data Science, Machine Learning and AI
Type: Talks
Target Audience: Intermediate
Last Updated: