A holiday reading list condensed into 30 quotes
For those who like brainfood on your vacation, here's a handy index of all my articles from 2018 boiled down to 30 (occasionally cheeky) punchlines to help you avoid/cause awkward silences at family events and holiday parties.
Sections: Data Science and Analytics, ML/AI Concepts, How Not To Fail At ML/AI, Data Science Leadership, Technology, Statistics.
Bonus: Videos, podcasts, foreign language translations for your non-English-speaking friends and family to enjoy, and an end-to-end deep learning tutorial for the Pythonistas among you.
Data Science and Analytics
Data science is the discipline of making data useful.
Each of the three data science disciplines has its own excellence. Statisticians bring rigor, ML engineers bring performance, and analysts bring speed.
Secret Paragraphs from HBR's Analytics
A collection of musings omitted from the article above. Let's talk about hybrid roles, the nature of research, Bat Signals, data charlatans, and awesome analysts!
Buyer beware: there are many data charlatans out there posing as data scientists. There's no magic that makes certainty out of uncertainty.
If a researcher is your first hire, you probably won't have the right environment to make good use of them.
Machine learning is a new programming paradigm, a new way of communicating your wishes to a computer. It's exciting because it allows you to automate the ineffable.
Are you using the term ‚??AI' incorrectly?
With poorly defined terms, there's not really such a thing as using them correctly. We can all be winners, but here's a quick guide to the alphabet soup of AI, ML, DL, RL, and HLI.
If you're worried that there's a human-like intelligence lurking in every cupboard, breathe easy. All those industry AI applications are too busy solving real business problems.
Don't be intimidated by jargon. For example, a model is just a fancy word for "recipe."
Don't hate machine learning for being simple. Levers are simple too, but they can move the world.
Unsupervised learning demystified
. Unsupervised learning helps you find inspiration in data by grouping similar things together for you. The results are a Rorschach card to help you dream.
Think of unsupervised learning as a mathematical version of making "birds of a feather flock together."
- Explainable AI won't deliver. Here's why. Many people are drawn to XAI because they think it's a good basis for trust. It isn't, and getting caught up in the trust hype might mean you'll miss out on something XAI is great for: inspiration.
If you refuse to trust decision-making to something whose process you don't understand, then you should fire all your human workers, because no one knows how the brain (with its hundred billion neurons!) makes decisions.
How Not To Fail At ML/AI
Imagine trying to start a restaurant by hiring folks who've been building microwave parts their whole lives but have never cooked a thing‚?¶ what could possibly go wrong?
A common mistake businesses make is to assume machine learning is magic, so it's okay to skip thinking about what it means to do the task well.
The first step in AI might surprise you
. What's the right way to start an AI project? Get an AI degree? No. Hire an AI wizard? Nope. Pick an awesome algorithm? Not that either. Dive into the data? Wrong again! Here's how to do it better.
Never ask a team of PhDs to "Go sprinkle machine learning over the top of the business so‚?¶ good things happen."
Don't waste your time on AI for AI's sake. Be motivated by what it will do for you, not by how sci-fi it sounds.
Just because you can do something, doesn't mean it's a good use of anyone's time. We humans fall in love with what we have poured effort into‚?¶ even if it is a pile of poisonous rubbish.
If you use a tool where it hasn't been verified safe, any mess you make is your fault. AI is a tool like any other.
Data Science Leadership
Data-Driven? Think again.
For a decision to be data-driven, it has to be the data‚??-‚??as opposed to something else entirely‚??-‚??that drive it. Seems so straightforward, and yet it's so rare in practice because decision-makers lack a key psychological habit.
The more ways there are to slice the data, the more your analysis is a breeding ground for confirmation bias. The antidote is setting your decision criteria in advance.
Is data science a bubble?
Learn more about the people calling themselves "data scientists" and why the industry is playing a dangerous game.
"I think you might be hiring data scientists the way a drug lord buys a tiger for his backyard," I told him. "You don't know what you want with the tiger, but all the other drug lords have one."
‚?¶a pro-math subculture where it's fashionable to display disdain for anything that smells like "soft" skills. It's all chest-thumping about how hardcore you are for staying up all night proving some theorem or coding in your sixth language.
Inspiration is cheap, but rigor is expensive.
Interview: Advice for data scientists
. Candid answers to a fellow data scientist's questions. Topics include: favorite resources, careers, statistics education, and data science leadership.
Useful is worth more than complicated. Data quality is worth more than method quality. Communication skills are worth more than yet another programming language.
9 Things You Should Know About TensorFlow
. TensorFlow might be your new best friend if you have a lot of data and/or you're after the state-of-the-art in AI. It's not a data science Swiss Army Knife, it's the industrial lathe. Here's what's new with it.
With TensorFlow Hub, you can engage in a more efficient version of the time-honored tradition of helping yourself to someone else's code and calling it your own (otherwise known as professional software engineering).
Congratulations on waiting it out long enough to have the infrastructure taken care of for you, kind of like you don't need to build your own computer anymore.
AI spent over half a century being more hype than happening. So, why now? Many people don't realize that the story of today's applied AI is actually a story about The Cloud.
Statistics is the science of changing your mind.
Never start with a hypothesis.
Starting with hypotheses instead of actions is a common mistake among those who learn the math without absorbing any of the philosophy. Let's look at how to do use statistics for decision-making.
Hypotheses are like cockroaches. When you see one, it's never just the one. There's always more hiding somewhere nearby.
Statistics for people in a hurry
. Ever wished someone would just tell you what the point of statistics is and what the jargon means in plain English? Let me try to grant that wish for you in 8 minutes!
The math is all about building a toy model of the null hypothesis universe. That's how you get the p-value.
Populations‚??-‚??You're doing it wrong.
A statistical approach only makes sense when there's a mismatch between the information you want (population) and the information you have (sample). What happens if the project's leader doesn't know what information they want?
In the Icarus-like leap from sample to population, expect a big splat if you don't know where you're aiming.
Statistics Savvy Self-Test
. Will you pass this small quiz that checks your statistical expertise? You might not if you believed what they told you in STAT101‚?¶
If you had facts, you wouldn't need statistics.
Incompetence, delegation, and population.
If the decision-maker doesn't have the right skills, your whole statistical project is doomed. When is it appropriate for the statistician to make a fuss and when should they meekly follow orders?
If your goal is to persuade people using data, you may as well throw rigor out the window (since that's where it belongs) and make pretty graphs instead.
Hands on Deep Learning Tutorial
My end-to-end deep learning tutorial screenshot walkthrough
and (mostly Python) code from last year's Supercomputing conference.