Musings on fascinating things
I know you’ll be back: interpretable new user clustering and churn prediction on a mobile social application
I know you’ll be back: interpretable new user clustering and churn prediction on a mobile social application
— Read on blog.acolyer.org/2018/10/05/i-know-youll-be-back-interpretable-new-user-clustering-and-churn-prediction-on-a-mobile-social-application/
Interesting paper on mobile user churn prediction at Snapchat
Want to make sure your vote counts in the Georgia November election? Casting an absentee ballot may be the way to go!
Federal judge Amy Totenberg (sister of Nina Totenberg if you are an NPR listener) ruled Monday that Georgia could go ahead using insecure paperless voting machines.
Although Judge Totenberg concurred with many cybersecurity experts that the voting machines pose a credible threat of alteration of ballot counts, she decided that the last minute switch would impose a burden on voters and a logistical challenge to the state’s election commission. If you’re not up on the current events, the head of the Georgia election commission is running for governor in a highly contested election and there have been some irregularities in his management (or lack thereof) of voting records. Hmm.
There have been charges of voter suppression, a 21st century step back from a hard won right to vote — ponder the 19th century Harper’s magazine image above celebrating the democratic participation of newly emancipated peoples post Civil War.
The main concern with electronic voting machines — especially the ones used in Georgia — is that they cannot be easily audited. The National Academy of Sciences in their report recommended in fact that “Voting machines that do not provide the capacity for independent auditing (e.g., machines that do not produce a voter-verifiable paper audit trail) should be removed from service as soon as possible.” This applies to the machines used in Georgia, and Judge Totenberg effectively ruled that this should be the last election in which such machines are use. Further, a malicious and technically sophisticated insider could alter voting records — that is “hack” the election.
Now if you cast an absentee ballot, the paper copy is preserved. Voter advocate sites like vote.org can help you get your ballot in a matter of seconds. There is an online tool that the Georgia Secretary of State’s office provides. I am not whether the online tool provides the functions that the National Academy of Sciences recommends — tracking of absentee ballot delivery and receipt.
In any case, absentee is a good way to be very present in the democratic process in Georgia.
Martin Luther King feared fully expected that there would be a backlash as basic human rights for African Americans expanded. There were historic precedents for this such as the reversal of post-Civil War Reconstruction-era freedoms during the last decades of the 19th century and the rise of segregationist laws throughout the southern US during the 1950s. Ta-Nahesi Coates speaks eloquently on the last incarnation of this in his Atlantic piece The First White President.
I am witnessing the evolution of this fourth backlash wave play out where I live, just east of Atlanta. Yesterday, my family and I witnessed the incident which is the featured image of this post. A White police officer handcuffs, and humiliates a young Black man about 15 feet from us through the window of a shop were we frequently have a relaxed brunch. The shop, I should mention is owned by a Black woman, also a resident of Avondale Estates. I should have stood with my young Brother and recorded the incident — sadly I know that life will give me other opportunities. To witness is a powerful comfort and statement. The young man showed calm and grace, as many of us have learned to do in such situations (I’ve been there).
Avondale Estates, a little east of Atlanta until recently had de-facto housing segregation through special real estate covenants. The history of housing segregation in Georgia, and the Atlanta area is both fascinating and frightening. If you zoom in on the Avondale Estates area, you’ll notice interesting racial disparities. There’s more information in this 50-year look back at the fair housing act.
Avondale Estates and the small towns to it’s north and east have been notorious for the disproportionate amount of revenue garnered from African Americans being stopped for minor traffic violations.
Georgia cities — especially Stone Mountain and Clarkston lead the nation in the fining of Black people
I wish we knew more, but like it’s voting transparency, information on racism in policing in Georgia is hard to come by.
Although one of my close friends (also African American) purchased a home in Avondale Estates in the ’90s, anecdotally (just from informal conversation), the number racialized policing and other incidents around exclusion have continued to grow.
I’ll detail one particular issue I’ve encountered. In Avondale, several times when walking while Black, my family and I have encountered on several occasions the question “Do you live around here?” Word of caution and advice for any (non-Black) person with the bad home training to ask this question. The 14th amendment of the Constitution of the United States is generally understood as guaranteeing access to the public roads to any US citizen, “green card” holder, or basically any human being regardless of what they look like. Dear White People, it’s not your militia/Klan duty to keep American or Avondale Estates or any other space in this country White. There’s my peculiar rant.
What can done? A few suggestions:
Let’s end the police state together.
A few days ago I attended the talk “Sparsity, oracles and inference in high-dimensional statistics” by Sara van der Geer who is visiting Georgia Tech. The talk is described here.
But I didn’t record the talk! I had a working iPhone! I only have an after thought photo of the white board that remained after the lecture
Just focus on lambda!
Phones are ubiquitous and there’s nothing like a short clip that can distill some of the essence of an idea, a lecture. Maybe it’s all those “No recording devices, please!” announcements at concerts, or that my videography skills are in need of serious help.
PSA: If you think that someone is bring across some important knowledge, record it — give them their attribution, don’t steal their stuff — but you are sharing knowledge with the world!
So what was the talk about? If you do machine learning, the idea of regularization is probably familiar. L1 regularization a.k.a Least Absolute Shrinkage and Selection Operator ak.a. lasso in particular assigns a penalty on the absolute value of the predictor weights. It’s an technique that reduces the tendency to overfit to the training data. There’s a whole book on it called Statistical Learning with Sparsity that you can download for free!
The amazing thing about lasso is that it also drives the less extraneous parameters close to zero: it can reduce the number of parameters you need in your model, or it results in a model that is more sparse (that is, just remove the close-to-zero parameters from the model). This can make the model faster to compute.
The main things I picked up were that there are some bounds on the error for lasso regularization that can be expressed in terms of the number of parameters and the number of observations you have in your training set. The error should be within a constant of 
She introduced something called the compatibility constant that’s discussed further in a couple of papers [Belloni, et. a. 2014, Dalalyan 2017]. She also discussed how lasso behaves when you assume that you have noisy observations. The final lecture is September 6th at Georgia Tech on applications to inference.
Wouldn’t it have been better if I’d just recorded it though??
The Foundations of Data Science Boot Camp given last week (August 27 – 31) at the Simons Institute in Berkeley explored how pure mathematics and theoretical computer science are providing actionable insights that the working data scientist can use — or at least ponder.
I found the talk below by Ravi Kannan useful in pointing out how dimensionality reduction techniques like SVD can be used to set clustering up for success. When dealing with immense data sets, this can be the difference between useful or garbage clusters.
I also thought that David P. Woodruff‘s lecture on a dimensionality reduction technique called sketching was impressive for its clarity. As a data scientist or analysis, you’re often in a dilemma when your Impala cluster runs out of memory for that critical model build — you may just have to sample from that terabyte pile of web pages. It is good to know that you have some math magic behind you when the time comes.
Santosh Vempala thinks the seminar was a better value than Netflix. I’m not sure about that, but those were some good lectures.
My colleague Boris Gorelik continues to deliver the insights of a data science sage. You should study carefully his recent EuroSciPy 2018 talk “Data visualization: from default to efficient”
Then check out his notes on how to continuously improve a talk.
My partner Dr Gayatri Sethi is interviewing celebrated writer and illustrator Yuyi Morales today at the Decatur Book Festival today, Saturday, September 1 from 1:00 to 1:30. They’ll be discussing Morales’ new book Dreamers.
Isn’t it time you talked to your children about what is going on at the border? Dreamers talks about Morales’s own 1994 journey from Xalapa, Mexico to the US. with her child. It is such a beautifully illustrated book, and a profound story with so many layers.
If you’re in looking to add a highlight to your long weekend, please come through!
THIS MACHINE MOCKS FASCISTS
Musings on fascinating things
Musings on fascinating things
Musings on fascinating things
Reviews and commentary on computational biology by Lior Pachter
Tales of a single mother traveling with her boy.
cha guan 茶館
Held together by caffeine & chocolate
Stuff from James Huff
Musings and learnings on WordPress and development
I can program and like games
Personal blog of Wattanit H.
Blogger Extraordinaire
Let's make distributed companies more inclusive
