Thread by @hima_lakkaraju

If you have less than 3 hours to spare & want to learn (almost) everything about state-of-the-art explainable ML, this thread is for you! Below, I am sharing info about 4 of our recent tutorials on explainability presented at NeurIPS, AAAI, FAccT, and CHIL conferences. [1/n]

NeurIPS 2020: Our longest tutorial (2 hours 46 mins) discusses various types of explanation methods, their limitations, evaluation frameworks, applications to domains such as decision making/nlp/vision, and open problems explainml-tutorial.github.io/neurips20 @sameer_ @julius_adebayo [2/n]

AAAI 2021: Can't spend 2 hours 46 mins on this topic? No problem! Our tutorial at AAAI 2021 is right here (1 hour 32 mins): explainml-tutorial.github.io/aaai21. This one discusses different explanation methods, their limitations, evaluation, and open problems. @sameer_ @julius_adebayo [3/n]

FAccT 2021: Want to know more about the ethical/practical implications of explainability along with a gentle intro to the topic? Our tutorial on "Explainable ML in the Wild" (1 hr 31 mins): slides: bit.ly/2REdAhe @shalmali_joshi_ @_cagarwal [4/n]

CHIL 2021: Alright, you can't even spare 1 hr 30 mins you say, no worries! Our shortest tutorial (just 1 hour) on this topic gives a quick overview of various state-of-the-art methods, their limitations, open problems: chilconference.org/tutorial_T04.h… slides bit.ly/3vQQwua [5/n]

If you think that's all we have got, you are wrong! :) This semester I taught a full fledged seminar course on explainability in ML @Harvard. All the readings and slides will be posted very soon. Stay tuned! Meanwhile, last year's version of the course at interpretable-ml-class.github.io

@trustworthy_ml @ai4life_harvard @black_in_ai @_LXAI
@DiverseInAI @harvard_data @fiddlerlabs Please RT above thread if it is of interest

Also tagging @DorsaSadigh (please see above thread) :) It was so great to meet you today, Dorsa.