Linear Digressions

LinearDegresion

Deep neural nets have a deserved reputation as the best-in-breed solution for computer vision problems. But there are many aspects of human vision that we take for granted but where neural nets struggle—this episode covers an eye-opening paper that summarizes some of the interesting weak spots of deep neural nets.

Relevant links: https://arxiv.org/abs/1805.04025

Limitations of Deep Nets for Computer Vision

Bringing you another old classic this week, as we gear up for 2019! See you next week with new content.

Word2Vec is probably the go-to algorithm for vectorizing text data these days.  Which makes sense, because it is wicked cool.  Word2Vec has it all: neural networks, skip-grams and bag-of-words implementations, a multiclass classifier that gets swapped out for a binary classifier, made-up dummy words, and a model that isn't actually used to predict anything (usually).  And all that's before we get to the part about how Word2Vec allows you to do algebra with text.  Seriously, this stuff is cool.

Re-release: Word2Vec

When you build a model for natural language processing (NLP), such as a recurrent neural network, it helps a ton if you’re not starting from zero. In other words, if you can draw upon other datasets for building your understanding of word meanings, and then use your training dataset just for subject-specific refinements, you’ll get farther than just using your training dataset for everything. This idea of starting with some pre-trained resources has an analogue in computer vision, where initializations from ImageNet used for the first few layers of a CNN have become the new standard. There’s a similar progression under way in NLP, where simple(r) embeddings like word2vec are giving way to more advanced pre-processing methods that aim to capture more sophisticated understanding of word meanings, contexts, language structure, and more.

Relevant links:
https://thegradient.pub/nlp-imagenet/

Pre-training language models for natural language processing problems

When data scientists use a linear regression to look for causal relationships between a treatment and an outcome, what they’re usually finding is the so-called average treatment effect. In other words, on average, here’s what the treatment does in terms of making a certain outcome more or less likely to happen. But there’s more to life than averages: sometimes the relationship works one way in some cases, and another way in other cases, such that the average isn’t giving you the whole story. In that case, you want to start thinking about heterogeneous treatment effects, and this is the podcast episode for you.

Relevant links:
https://eng.uber.com/analyzing-experiment-outcomes/
https://multithreaded.stitchfix.com/blog/2018/11/08/bandits/
https://www.locallyoptimistic.com/post/against-ab-tests/

Heterogeneous Treatment Effects

Linear regression is a great tool if you want to make predictions about the mean value that an outcome will have given certain values for the inputs. But what if you want to predict the median? Or the 10th percentile? Or the 90th percentile. You need quantile regression, which is similar to ordinary least squares regression in some ways but with some really interesting twists that make it unique. This week, we’ll go over the concept of quantile regression, and also a bit about how it works and when you might use it.

Relevant links:
https://www.aeaweb.org/articles?id=10.1257/jep.15.4.143
https://eng.uber.com/analyzing-experiment-outcomes/

Quantile Regression

Ordinary least squares (OLS) is often used synonymously with linear regression. If you’re a data scientist, machine learner, or statistician, you bump into it daily. If you haven’t had the opportunity to build up your understanding from the foundations, though, listen up: there are a number of assumptions underlying OLS that you should know and love. They’re interesting, force you to think about data and statistics, and help you know when you’re out of “good” OLS territory and into places where you could run into trouble.

The Assumptions of Ordinary Least Squares

Open source software is ubiquitous throughout data science, and enables the work of nearly every data scientist in some way or another. Open source projects, however, are disproportionately maintained by a small number of individuals, some of whom are institutionally supported, but many of whom do this maintenance on a purely volunteer basis. The health of the data science ecosystem depends on the support of open source projects, on an individual and institutional level.

https://hdsr.mitpress.mit.edu/pub/xsrt4zs2/release/2

The Data Science Open Source Ecosystem

This is a re-release of an episode that originally ran in October 2019.

If you’re trying to manage a project that serves up analytics data for a few very distinct uses, you’d be wise to consider having custom solutions for each use case that are optimized for the needs and constraints of that use cases. You also wouldn’t be YouTube, which found themselves with this problem (gigantic data needs and several very different use cases of what they needed to do with that data) and went a different way: they built one analytics data system to serve them all. Procella, the system they built, is the topic of our episode today: by deconstructing the system, we dig into the four motivating uses of this system, the complexity they had to introduce to service all four uses simultaneously, and the impressive engineering that has to go into building something that “just works.”

Procella: YouTube's super-system for analytics data storage

A few weeks ago, we put out a call for data scientists interested in issues of race and racism, or people studying how those topics can be studied with data science methods, should get in touch to come talk to our audience about their work. This week we’re excited to bring on Todd Hendricks, Bay Area data scientist and a volunteer who reached out to tell us about his studies with the Stanford Open Policing dataset.

A Data Science Take on Open Policing Data

The data science and artificial intelligence community has made amazing strides in the past few years to algorithmically automate portions of the healthcare process. This episode looks at two computer vision algorithms, one that diagnoses diabetic retinopathy and another that classifies liver cancer, and asks the question—are patients now getting better care, and achieving better outcomes, with these algorithms in the mix? The answer isn’t no, exactly, but it’s not a resounding yes, because these algorithms interact with a very complex system (the healthcare system) and other shortcomings of that system are proving hard to automate away. Getting a faster diagnosis from an image might not be an improvement if the image is now harder to capture (because of strict data quality requirements associated with the algorithm that wouldn’t stop a human doing the same job). Likewise, an algorithm getting a prediction mostly correct might not be an overall benefit if it introduces more dramatic failures when the prediction happens to be wrong. For every data scientist whose work is deployed into some kind of product, and is being used to solve real-world problems, these papers underscore how important and difficult it is to consider all the context around those problems.

A Reality Check on AI-Driven Medical Assistants

All good things must come to an end, including this podcast. This is the last episode we plan to release, and it doesn’t cover data science—it’s mostly reminiscing, thanking our wonderful audience (that’s you!), and marveling at how this thing that started out as a side project grew into a huge part of our lives for over 5 years.

It’s been a ride, and a real pleasure and privilege to talk to you each week. Thanks, best wishes, and good night!

—Katie and Ben