A Nod to Nothing: January 2021 Reading List

Dark Data: Why What You Don't Know Matters by David J. Hand
Mack Lecture at the Walker: Conversation with Annie Dorsen [artist], Zachary Chase Lipton [ML], and Simon Adler [reporter]. Dorsen merges machine learning with art in projects such as Hello, Hi There and Yesterday, Tomorrow [60:00] - talked a lot about Eliza and how the bots were trained and where to find gaps in ML that are worth exploring using art.

Ring Shout by P. Djeli Clark
3 facts about time series forecasting that surprise experienced machine learning practitioners - by Skander Hannachi, PhD

ARIMA and seasonality...found out this ties into the AWS Forecast service.
You need to retrain your model every time you want to generate a new prediction:
Sometimes, you have to do away with train/test splits:
The uncertainty of the forecast is just as important as, or even more so, than the forecast itself:

This is cool: https://airtable.com/shrndwzEx01al2jHM/tblYMAiGeDLXe35jC
Prove that you’re solving a real problem that people are struggling with
Prove that you’re solving it at least ten times better than others in the market
Prove that you can deliver your services at scale
Have a roadmap for profitability, where the cost of acquiring a customer is lower than the average revenue per customer
Have a business model that can’t be copied by competitors
You must have a solid infrastructure that consolidates the data needed to train your AI models
You must have the means to collect fresh quality data to continuously learn from users’ interactions with your product and fine-tune your algorithms

MN Hardcore: The Binge Episode by TPT [1:46:58]
Half Life [45:00] - TPT / KTCA made for TV movie from 1985 about the dangers of city women and being in a hardcore/punk band downtown. Farm boys beware.
Ring Shout by P. Djeli Clark

1/22/2021: Near and Far: Games Modes [12:32]
1/21/2021: Ring Shout by P. Djeli Clark - he wrote The Secret Lives of the Nine Negro Teeth of George Washington which was excellent.
1/20/2021: Near and Far: How to Play [38:36]
1/19/2021: The Churn by James S.A. Corey [audio via Hoopa, 118:00]
1/18/2021: The World is Full of Monsters by Jeff Vandermeer [audio via Hoopla, 60:00]
1/17/2021: Why Sleep Matters, Google Talk by Matthew Walker [57:00, but watched twice, once with my team]
1/16/2021: Episode 1: A premeditated lie lit the fire - Axios, by Jonathan Swan and Zachary Basu
1/15/2021: La Jatee', roughly 30 minutes - influenced 12 Monkeys.
1/14/2021: At This Banana Farm, the Bunches Grow in 430 Shapes and Sizes - Atlas Obscura
Republicans Still Don’t Get It: Even after almost dying, they are screaming about their right to blather while in the act of blathering.

"We have come to a moment in which one half of the country is fighting to be free of crippling, life-ending acts of stochastic terror, while another half of the same country is chillingly preoccupied with their right to just talk shit. These harms are not the same."

1/13/2021: I Was Raised On the Lie of the Reluctant, Rural Insurrection by Emily Alford
1/12/2021: Microservices [humor, 3:09]
1/11/2021: A Plague of Madness: Twenty-five years ago, Terry Gilliam and the other outsider : creators of ‘12 Monkeys’ gave humanity a warning about our pandemic-filled future, whether they meant to or not
1/10/2021:

The state of America’s division produces little hope that ‘we can work it out’: By Mary Stanik
Engaging With Trump’s Die-Hard Supporters Isn’t Productive: The loyalists who still cling to conspiracy theories should be deprived of the attention they seek. By Tom Nicols
The Conservative Cult of Victimhood: Trump was a perpetrator who thought himself a victim, and American society has indulged that same illusion among Trump supporters. By David Frum

"Unfortunately, just because something may pique interest, it is not always relevant to the conversation."
I think this focuses more on the specific data points and doesn't account for trend.

1/8/2021: Outlier Detection and Treatment: A Beginner's Guide by Swetha Lakshmanan on Medium.com

Impacts mean, variance, and correlation.
Identify: Extreme value analysis, z-score method, k means clustering, visualize.
Pre-process: Mean/median or random imputation, trimming, top/bottom/zero coding, discretization [intervals or binning].
"There is no rigid mathematical definition of what constitutes an outlier"
+/- 3x STDev >> there are "levels" of outliers [upper limit and extreme upper limit]
Z-score uses +/-3. Has a specific formula.
Kmeans for groups of data?