Episode 32: Economics, Data For Good and AI Research with Sara Hooker

Show Notes:

• (2:20) Sara shared her childhood growing up in Africa.
• (4:05) Sara talked about her undergraduate experience at Carleton College studying Economics and International Relations.
• (9:07) Sara discussed her first job working as an Economics Analyst at Compass Lexecon in the Bay Area.
• (12:20) Sara then joined Udemy as a data analyst, then transitioned to the engineering team to work on spam detection and recommendation algorithms.
• (14:58) Sara dig deep into the “hustling period” of her career and how she brute-forced her way to grow as an engineer.
• (17:24) Sara founded Delta Analytics - a local Bay Area non-profit community of data scientists, engineers, and economists in 2014 that believes in using data for good.
• (20:53) Sara shared Delta’s collaboration with Eneza Education to empower students to access quizzes by mobile texting in Kenya (check out her presentation at the ODSC West 2016).
• (25:16) Sara shared Delta’s partnership with Rainforest Connection to identify illegal de-forestation using steamed audio from the rainforest (check out her presentation at MLconf Seattle 2017).
• (28:22) Sara unpacked her blog post Why “data for good” lacks precision, in which she described 4 key criteria frequently used to qualify an initiative as “data for good” and discussed some open challenges associated with each.
• (36:34) Sara unpacked her blog post Slow learning, in which she revealed her journey to get accepted into the AI Residency program at Google AI.
• (41:03) Sara discussed her initial research interest on model interpretability for deep neural networks and her work done at Google called The (Un)reliability of Saliency Methods - which argues that saliency methods are not reliable enough to explain model prediction.
• (45:55) Sara pushed the research above further with A Benchmark for Interpretability Methods in Deep Neural Networks, which proposes an empirical measure of the approximate accuracy of feature importance estimates in deep neural networks called RemOve And Retrain.
• (48:46) Sara explained why model interpretability is not always required (check out her talks at PyBay 2018, REWORK Toronto 2018, and REWORK San Francisco 2019).
• (52:10) Sara explained the typical measurements of model reliability and the limitations of them, such as localization methods and points of failure.
• (59:04) Sara explained why model compression is an interesting research direction and her work The State of Sparsity in Deep Neural Networks - which highlights the need for large-scale benchmarks in the field of model compression.
• (01:02:49) Sara discussed her paper Selective Brain Damage: Measuring the Disparate Impact of Model Pruning - which explores the impact of pruning techniques for neural networks trained for computer vision tasks. Check out the paper website!
• (01:05:08) Sara shared her future research directions on efficient pruning, sparse network training, and local gradient updates.
• (01:06:56) Sara explained the premise behind her talk Gradual Learning at the Future of Finance Summit in 2019, in which she shared the three fundamental approaches to machine learning impact.
• (01:12:20) Sara described the AI community in Africa as well as the issues the community is currently facing: both from the investment landscape and the infrastructure ecosystem.
• (01:18:00) Sara and her brother recently started a podcast called Underrated ML which pitches the underrated ideas in machine learning.
• (01:20:15) Sara reflected how her background in economics influences her career outlook in machine learning.
• (01:25:42) Sara reflected on the differences between applied ML a