Linear Digressions

This episode is for all you (us) traffic nerds--we're talking about the hidden structure underlying traffic on-ramp metering systems. These systems slow down the flow of traffic onto highways so that the highways don't get overloaded with cars and clog up. If you're someone who listens to podcasts while commuting, and especially if your area has on-ramp metering, you'll never look at highway access control the same way again (yeah, we know this is super nerdy; it's also super awesome). Relevant links: http://its.berkeley.edu/sites/default/files/publications/UCB/99/PWP/UCB-ITS-PWP-99-19.pdf http://www.its.uci.edu/~lchu/ramp/Final_report_mou3013.pdf

One tricky thing about working with time series data, like the audio data in our "um" detector (remember that? because we barely do...), is that sometimes events look really similar but one is a little bit stretched and squeezed relative to the other. Besides having an amazing name, the dynamic time warp is a handy algorithm for aligning two time series sequences that are close in shape, but don't quite line up out of the box. Relevant link: http://www.aaai.org/Papers/Workshops/1994/WS-94-03/WS94-03-031.pdf

It's storytime this week--the story, from beginning to end, of how Katie designed and built the main project for Udacity's Intro to Machine Learning class, when she was developing the course. The project was to use email and financial data to hunt for signatures of fraud at Enron, one of the biggest cases of corporate fraud in history; that description makes the project sound pretty clean but getting the data into the right shape, and even doing some dataset merging (that hadn't ever been done before), made this project much more interesting to design than it might appear. Here's the story of what a data analysis like this looks like...from the inside.

Data science and is often mentioned in the same breath as big data. But how big is big data? And who has the biggest big data? CERN? Youtube? ... Something (or someone) else? Relevant link: http://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1002195

Supervised machine learning assumes that the features and labels used for building a classifier are isolated from each other--basically, that you can't cheat by peeking. Turns out this can be easier said than done. In this episode, we'll talk about the many (and diverse!) cases where label information contaminates features, ruining data science competitions along the way. Relevant links: https://www.researchgate.net/profile/Claudia_Perlich/publication/221653692_Leakage_in_data_mining_Formulation_detection_and_avoidance/links/54418bb80cf2a6a049a5a0ca.pdf

Machine learning algorithms can be black boxes--inputs go in, outputs come out, and what happens in the middle is anybody's guess. But understanding how a model arrives at an answer is critical for interpreting the model, and for knowing if it's doing something reasonable (one could even say... trustworthy). We'll talk about a new algorithm called LIME that seeks to make any model more understandable and interpretable. Relevant Links: http://arxiv.org/abs/1602.04938 https://github.com/marcotcr/lime/tree/master/lime

We've got updates for you about topics from past shows! First, the political science scandal of the year 2015 has a new chapter, we'll remind you about the original story and then dive into what has happened since. Then, we've got an update on AlphaGo, and his/her/its much-anticipated match against the human champion of the game Go. Relevant Links: https://soundcloud.com/linear-digressions/electoral-insights-part-2 https://soundcloud.com/linear-digressions/go-1 http://www.sciencemag.org/news/2016/04/talking-people-about-gay-and-transgender-issues-can-change-their-prejudices http://science.sciencemag.org/content/sci/352/6282/220.full.pdf http://qz.com/639952/googles-ai-won-the-game-go-by-defying-millennia-of-basic-human-instinct/ http://www.wired.com/2016/03/two-moves-alphago-lee-sedol-redefined-future/ http://www.wired.com/2016/03/sadness-beauty-watching-googles-ai-play-go/

Simpson's paradox is the data science equivalent of looking through one eye and seeing a very clear trend, and then looking through the other eye and seeing the very clear opposite trend. In one case, you see a trend one way in a group, but then breaking the group into subgroups gives the exact opposite trend. Confused? Scratching your head? Welcome to the tricky world of ecological inference. Relevant links: https://gking.harvard.edu/files/gking/files/part1.pdf http://blog.revolutionanalytics.com/2013/07/a-great-example-of-simpsons-paradox.html

Sometimes when we say an algorithm discriminates, we mean it can tell the difference between two types of items. But in this episode, we'll talk about another, more troublesome side to discrimination: algorithms can be... racist? Sexist? Ageist? Yes to all of the above. It's an important thing to be aware of, especially when doing people-centered data science. We'll discuss how and why this happens, and what solutions are out there (or not). Relevant Links: http://www.nytimes.com/2015/07/10/upshot/when-algorithms-discriminate.html http://techcrunch.com/2015/08/02/machine-learning-and-human-bias-an-uneasy-pair/ http://www.sciencefriday.com/segments/why-machines-discriminate-and-how-to-fix-them/ https://medium.com/@geomblog/when-an-algorithm-isn-t-2b9fe01b9bb5#.auxqi5srz

This episode started out as a discussion of recommendation engines, like Netflix uses to suggest movies. There's still a lot of that in here. But a related topic, which is both interesting and important, is how to keep data private in the era of large-scale recommendation engines--what mistakes have been made surrounding supposedly anonymized data, how data ends up de-anonymized, and why it matters for you. Relevant links: http://www.netflixprize.com/ http://bits.blogs.nytimes.com/2010/03/12/netflix-cancels-contest-plans-and-settles-suit/?_r=0 http://arxiv.org/PS_cache/cs/pdf/0610/0610105v2.pdf

One neural net is creating counterfeit bills and passing them off to a second neural net, which is trying to distinguish the real money from the fakes. Result: two neural nets that are better than either one would have been without the competition. Relevant links: http://arxiv.org/pdf/1406.2661v1.pdf http://arxiv.org/pdf/1412.6572v3.pdf http://soumith.ch/eyescream/

In this episode, we're taking many episodes' worth of insights and unpacking an extremely complex and important question--in what ways are we winning the fight against cancer, where might that fight go in the coming decade, and how do we know when we're making progress? No matter how tricky you might think this problem is to solve, the fact is, once you get in there trying to solve it, it's even trickier than you thought.

Hey, sick of the election yet? Fear not, there are algorithms that can automagically generate political-ish speech so that we never need to be without an endless supply of Congressional speeches and Donald Trump twitticisms! Relevant links: http://arxiv.org/pdf/1601.03313v2.pdf http://qz.com/631497/mit-built-a-donald-trump-ai-twitter-bot-that-sounds-scarily-like-him/ https://twitter.com/deepdrumpf

Multi-armed bandits: how to take your randomized experiment and make it harder better faster stronger. Basically, a multi-armed bandit experiment allows you to optimize for both learning and making use of your knowledge at the same time. It's what the pros (like Google Analytics) use, and it's got a great name, so... winner! Relevant link: https://support.google.com/analytics/answer/2844870?hl=en

"People with a family history of heart disease are more likely to eat healthy foods, and have a high incidence of heart attacks." Did the healthy food cause the heart attacks? Probably not. But establishing causal links is extremely tricky, and extremely important to get right if you're trying to help students, test new medicines, or just optimize a website. In this episode, we'll unpack randomized experiments, like AB tests, and maybe you'll be smarter as a result. Will you be smarter BECAUSE of this episode? Well, tough to say for sure... Relevant link: http://tylervigen.com/spurious-correlations

The reason that neural nets are taking over the world right now is because they can be efficiently trained with the backpropagation algorithm. In short, backprop allows you to adjust the weights of the neural net based on how good of a job the neural net is doing at classifying training examples, thereby getting better and better at making predictions. In this episode: we talk backpropagation, and how it makes it possible to train the neural nets we know and love.

First up in this episode: a crash course in natural language processing, and important steps if you want to use machine learning techniques on text data. Then we'll take that NLP know-how and talk about a really cool analysis of State of the Union text, which analyzes the topics and word choices of every President from Washington to Obama. Relevant link: https://civisanalytics.com/blog/data-science/2016/01/15/data-science-on-state-of-the-union-addresses/

Artificial intelligence includes a number of different strategies for how to make machines more intelligent, and often more human-like, in their ability to learn and solve problems. An ambitious group of researchers is working right now to classify all the approaches to AI, perhaps as a first step toward unifying these approaches and move closer to strong AI. In this episode, we'll touch on some of the most provocative work in many different subfields of artificial intelligence, and their strengths and weaknesses. Relevant links: https://www.technologyreview.com/s/544606/can-this-man-make-aimore-human/ https://www.youtube.com/watch?v=B8J4uefCQMc http://venturebeat.com/2013/11/29/sentient-code-an-inside-look-at-stephen-wolframs-utterly-new-insanely-ambitious-computational-paradigm/ http://www.slate.com/articles/technology/bitwise/2014/03/stephen_wolfram_s_new_programming_language_can_he_make_the_world_computable.html

Survival analysis is all about studying how long until an event occurs--it's used in marketing to study how long a customer stays with a service, in epidemiology to estimate the duration of survival of a patient with some illness, and in social science to understand how the characteristics of a war inform how long the war goes on. This episode talks about the special challenges associated with survival analysis, and the tools that (data) scientists use to answer all kinds of duration-related questions.

All aboard the gravitational waves bandwagon--with the first direct observation of gravitational waves announced this week, Katie's dusting off her physics PhD for a very special gravity-related episode. Discussed in this episode: what are gravitational waves, how are they detected, and what does this announcement mean for future studies of the universe. Relevant links: http://www.nytimes.com/2016/02/12/science/ligo-gravitational-waves-black-holes-einstein.html https://www.ligo.caltech.edu/news/ligo20160211

Let's imagine a future in which a truly intelligent computer program exists. How would it convince us (humanity) that it was intelligent? Alan Turing's answer to this question, proposed over 60 years ago, is that the program could convince a human conversational partner that it, the computer, was in fact a human. 60 years later, the Turing Test endures as a gold standard of artificial intelligence. It hasn't been beaten, either--yet. Relevant links: https://en.wikipedia.org/wiki/Turing_test http://commonsensereasoning.org/winograd.html http://consumerist.com/2015/09/29/its-not-just-you-robots-are-also-bad-at-assembling-ikea-furniture/

Psychometrics is all about measuring the psychological characteristics of people; for example, scholastic aptitude. How is this done? Tests, of course! But there's a chicken-and-egg problem here: you need to know both how hard a test is, and how smart the test-taker is, in order to get the results you want. How to solve this problem, one equation with two unknowns? Item response theory--the data science behind such tests and the GRE. Relevant links: https://en.wikipedia.org/wiki/Item_response_theory

As you may have heard, a computer beat a world-class human player in Go last week. As recently as a year ago the prediction was that it would take a decade to get to this point, yet here we are, in 2016. We'll talk about the history and strategy of game-playing computer programs, and what makes Google's AlphaGo so special. Relevant link: http://googleresearch.blogspot.com/2016/01/alphago-mastering-ancient-game-of-go.html

The Medici were one of the great ruling families of Europe during the Renaissance. How did they come to rule? Not power, or money, or armies, but through the strength of their social network. And speaking of great historical social networks, analysis of the network of letter-writing during the Enlightenment is helping humanities scholars track the dispersion of great ideas across the world during that time, from Voltaire to Benjamin Franklin and everyone in between. Relevant links: https://www2.bc.edu/~jonescq/mb851/Mar12/PadgettAnsell_AJS_1993.pdf http://republicofletters.stanford.edu/index.html

A few small encores on auction theory, and then--how can you value a piece of information before you know what it is? Decision theory has some pointers. Some highly relevant information if you are trying to figure out how much to pay a spy. Relevant links: https://tuecontheoryofnetworks.wordpress.com/2013/02/25/the-origin-of-the-dutch-auction/ http://www.nowozin.net/sebastian/blog/the-fair-price-to-pay-a-spy-an-introduction-to-the-value-of-information.html

The Google ads auction is a special kind of auction, one you might not know as well as the famous English auction (which we talked about in the last episode). But if it's what Google uses to sell billions of dollars of ad space in real time, you know it must be pretty cool. Relevant links: https://en.wikipedia.org/wiki/English_auction http://people.ischool.berkeley.edu/~hal/Papers/2006/position.pdf http://www.benedelman.org/publications/gsp-060801.pdf

The Google AdWords algorithm is (famously) an auction system for allocating a massive amount of online ad space in real time--with that fascinating use case in mind, this episode is part one in a two-part series all about auctions. We dive into the theory of auctions, and what makes a "good" auction. Relevant links: https://en.wikipedia.org/wiki/English_auction http://people.ischool.berkeley.edu/~hal/Papers/2006/position.pdf http://www.benedelman.org/publications/gsp-060801.pdf

A data visualization extravaganza in this episode, as we discuss Chernoff faces (you: "faces? huh?" us: "oh just you wait") and the greatest data visualization of all time, or at least the Napoleonic era. Relevant links: http://lya.fciencias.unam.mx/rfuentes/faces-chernoff.pdf https://en.wikipedia.org/wiki/Charles_Joseph_Minard

Ever tried to visualize a cluster of data points in 40 dimensions? Or even 4, for that matter? We prefer to stick to 2, or maybe 3 if we're feeling well-caffeinated. The t-SNE algorithm is one of the best tools on the market for doing dimensionality reduction when you have clustering in mind. Relevant links: https://www.youtube.com/watch?v=RJVL80Gg3lA

The town of [expletive deleted], England, is responsible for the clbuttic [expletive deleted] problem. This week on Linear Digressions: we try really hard not to swear too much. Related links: https://en.wikipedia.org/wiki/Scunthorpe_problem https://www.washingtonpost.com/news/worldviews/wp/2016/01/05/where-is-russia-actually-mordor-in-the-world-of-google-translate/

In order to do supervised learning, you need a labeled training dataset. Or do you...? Relevant links: http://www.cs.columbia.edu/~dplewis/candidacy/goldman00enhancing.pdf

Machine learning: it can be fooled, just like you or me. Here's one of our favorite examples, a study into hacking neural networks. Relevant links: http://arxiv.org/pdf/1412.1897v4.pdf

Zipf's law is related to the statistics of how word usage is distributed. As it turns out, this is also strikingly reminiscent of how income is distributed, and populations of cities, and bug reports in software, as well as tons of other phenomena that we all interact with every day. Relevant links: http://economix.blogs.nytimes.com/2010/04/20/a-tale-of-many-cities/ http://arxiv.org/pdf/cond-mat/0412004.pdf https://terrytao.wordpress.com/2009/07/03/benfords-law-zipfs-law-and-the-pareto-distribution/

We've gone indie! Which shouldn't change anything about the podcast that you know and love, but we're super excited to keep bringing you Linear Digressions as a fully independent podcast. Some links mentioned in the show: https://twitter.com/lindigressions https://twitter.com/benjaffe https://twitter.com/multiarmbandit https://soundcloud.com/linear-digressions http://lineardigressions.com/

It's Da Vinci meets Skynet: what makes a portrait beautiful, according to a machine learning algorithm. Snap a selfie and give us a listen.

Grab a cocktail, put on your favorite karaoke track, and let’s talk some more about disentangling audio data!

Because there are more interesting problems than there are labeled datasets, semi-supervised learning provides a framework for getting feedback from the environment as a proxy for labels of what's "correct." Of all the machine learning methodologies, it might also be the closest to how humans usually learn--we go through the world, getting (noisy) feedback on the choices we make and learn from the outcomes of our actions.

Overfitting to your training data can be avoided by evaluating your machine learning algorithm on a holdout test dataset, but what about overfitting to the test data? Turns out it can be done, easily, and you have to be very careful to avoid it. But an algorithm from the field of privacy research shows promise for keeping your test data safe from accidental overfitting

How many data scientists are there, where do they live, where do they work, what kind of tools do they use, and how do they describe themselves? RJMetrics wanted to know the answers to these questions, so they decided to find out and share their analysis with the world. In this very special interview episode, we welcome Tristan Handy, VP of Marketing at RJMetrics, who will talk about "The State of Data Science Report."

There's a good chance that great data science is going on close to you, and that it's going toward making your city, state, country, and planet a better place. Not all the data science questions being tackled out there are about finding the sleekest new algorithm or billion-dollar company idea--there's a whole world of social data science that just wants to make the world a better place to live in.

The Kalman Filter is an algorithm for taking noisy measurements of dynamic systems and using them to get a better idea of the underlying dynamics than you could get from a simple extrapolation. If you've ever run a marathon, or been a nuclear missile, you probably know all about these challenges already. By the way, we neglected to mention in the episode: Katie's marathon time was 3:54:27!

When you sleep, the neural pathways in your brain take the "white noise" of your resting brain, mix in your experiences and imagination, and the result is dreams (that is a highly unscientific explanation, but you get the idea). What happens when neural nets are put through the same process? Train a neural net to recognize pictures, and then send through an image of white noise, and it will start to see some weird (but cool!) stuff.

Sometimes numbers are... weird. Benford's Law is a favorite example of this for us--it's a law that governs the distribution of the first digit in certain types of numbers. As it turns out, if you're looking up the length of a river, the population of a country, the price of a stock... not all first digits are created equal.

Not to oversell it, but the student's t-test has got to have the most interesting history of any statistical test. Which is saying a lot, right? Add some boozy statistical trivia to your arsenal in this epsiode.

Doing some science, and want to know if you might have found something? Or maybe you've just accomplished the scientific equivalent of going fishing and reeling in an old boot? Frequentist p-values can help you distinguish between "eh" and "oooh interesting". Also, there's a lot of physics in this episode, nerds.

This machine learning algorithm beat the human champions at Jeopardy. What is... Watson?

Come get a little "out there" with us this week, as we use a meta-study of extrasensory perception (or ESP, often used in the same sentence as "psychics") to chat about Bayesian vs. frequentist statistics.

Ever found yourself wasting time reading online comments from trolls? Of course you have; we've all been there (it's 4 AM but I can't turn off the computer and go to sleep--someone on the internet is WRONG!). Now there's a way to use machine learning to automatically detect trolls, and minimize the impact when they try to derail online conversations.

Imagine a language that is mostly spoken rather than written, contains many words in other languages, and has relatively little written overlap with English. Now imagine writing a machine-learning-based translation system that can convert that language to English. That's the problem that confronted researchers when they set out to automatically translate between Yiddish and English; the tricks they used help us understand a lot about machine translation.

In a fun historical journey, Katie and Ben explore the history of the Manhattan Project, discuss the difficulties in modeling particle movement in atomic bombs with only punch-card computers and ingenuity, and eventually come to present-day uses of the Metropolis-Hastings algorithm... mentioning Solitaire along the way.