You gotta think about this one carefully!

Imagine you go to the doctor and get tested for a rare disease (only 1 in 10,000 people get it.)

The test is 99% effective in detecting both sick and healthy people.

Your test comes back positive.

Are you really sick? Explain below 👇

10 machine learning questions you should be able to answer.

I started experimenting with a questionnaire before interviewing candidates. Applicants submit a video or written answers to apply for the position.

Thread: Let's look at 10 of these questions.

The questionnaire solves a couple of problems:

1. Weeds out candidates that aren't ready for the job.

2. It significantly cuts down the time we spend talking synchronously.

One problem: It requires extra time from the candidate before the phone interview.

Here are some numbers on a job posting before using the questionnaire:

• 249 applicants
• 13 good applications

That's only 5%.

This sucks.

Numbers after using the questionnaire:

• 44 applicants
• 21 good applications

That's 47%.

This is much better.

I look at it this way:

Sooner or later, I'll ask these questions, and the candidate will have to answer them.

I'd rather give them space and the opportunity to think about their answers without any pressure.

Those who can't or don't want to answer won't apply. That's fine.

A topic that comes up in every interview.

Bias, variance, and their relationship with machine learning algorithms. One of the most basic concepts that you have to know by heart.

Here is a simple summary that you will easily remember.

↓

Every machine learning algorithm deals with 3 types of errors:

1. Bias error
2. Variance error
3. Irreducible error

There's nothing we can do about #3.

Let's focus on the other two.

↓ 1/5

"Bias" refers to the assumptions the model makes to simplify the process of finding answers.

The more assumptions it makes, the more biased the model is.

"Variance" refers to how much the answers given by the model will change if we use different training data.

If the answers stay the same regardless of the data, the model has low variance.

Often, linear models are high-bias, and nonlinear models are low-bias.

Example low-bias algorithms:
• Decision Trees
• SVN
• kNN

Example high-bias algorithms:
• Linear Regression
• Logistic Regression

20 machine learning questions that will make you think.

(Cool questions. Not the regular, introductory stuff that you find everywhere.)

🧵👇

1. Why is it important to introduce non-linearities in a neural network?

2. What are the differences between a multi-class classification problem and a multi-label classification problem?

3. Why does the use of Dropout work as a regularizer?

👇

4. Why you shouldn't use a softmax output activation function in a multi-label classification problem when using a one-hot-encoded target?

5. Does the use of Dropout in your model slow down or speed up the training process? Why?

👇

6. In a Linear or Logistic Regression problem, do all Gradient Descent algorithms lead to the same model, provided you let them run long enough?

7. Explain the difference between Batch Gradient Descent, Stochastic Gradient Descent, and Mini-batch Gradient Descent.

👇

8. What are the advantages of Convolution Neural Networks (CNN) over a fully connected network for image classification?

9. What are the advantages of Recurrent Neural Networks (RNN) over a fully connected network when working with text data?

👇

Machine learning education is broken.

If you are preparing for a research position, you are good. If you are looking to get out there and start solving problems, not even close.

Here are some thoughts so you can get ahead.

↓

Most classes, courses, and books cover the same road.

They start with a dataset. They finish with a working model. The focus is always on everything that happens in between.

Dataset → Model.

This is great, but not enough.

Real-life situations rarely start with a dataset, and they never end after you finish building your model.

Applying machine learning successfully is hard.

Here are a few examples that you should keep in mind.

First challenge: Properly framing up the problem.

If you don't understand the problem, you can't determine what data you need. If you don't understand the data, you can't build a good model.

I've never seen a company that had their data ready to go.

In fact, most of them don't even have data at all and need you to determine what exactly they should start collecting.

You usually have to go Problem → Potential Solution → Data.

On December 29 we were dillucidating here if the drops meant "surprise, relatively neutral expression" or "a negative expression, horror" (Remember: https://t.co/J2gjQjAENw ). Well, December 22 drops seem to clarify it. First, the December 29 drop that generated the confusion 1/*

didn't appear in today drops, while several old ones have been posted again. I would say that all December 22 drops regarding facial expression mean and clarify "surprise". 2/*

More 3/*

And more. So, they didn't point to a negative event, just to one that generates surprise. 4/*

All these drops were completed with a Christmas gift one, confirming the surprise neutral meaning and adding a related positive one connected to the surprise. 5/*

A week ago I voluntarily admitted myself to the hospital to get help bc I am losing weight as a result of increased food intolerances, secondary to my very severe #MECFS. A psychiatrist here decided that im mentally ill, 1/

purely on the basis that the medical workup was unable to explain my food intolerance. My psych eval only found mild anxiety and depression. 2/

They have put me on a 72hr psych hold while they work to get the county courts to commit me to a psych ward. This is a nightmare. They have taken away most of my belongings, put me in a new uniform, and are keeping me under 24/7 video surveillance. 3/

In the middle of the night last night when I sat up to use the urinal loud alarms sounded, as though I was trying to escape. I can’t even walk and hadn’t even left my bed. 4/

If they succeed in having me fully committed I will lose ability to communicate with the outside world except for an occasional phone call. I will lose all agency, and will very likely have treatments forced upon me. 5/

12 things you should look out for before buying health insurance

A thread 🧵

It’s a good idea to make sure your health is protected under an insurance policy so that you don’t shell out large sums of money for hospitalization and treatment. But buying an individual health insurance policy can be daunting in itself.

Websites and policy documents can be overwhelming. There’s a lot of loaded terms, there’s a whole bunch of things that your policy may or may not cover. It’s important to know and understand the details though, so you’re not surprised later.

In this thread, we'll talk about the things you absolutely need to watch out for when buying an individual health insurance policy for yourself or your family members :)

If you take an individual health insurance policy and the insurance company offers to pay up to ₹15 lakhs during the year whenever you’re diagnosed with a disease or meet an accident, that amount is the 'sum insured'.

A staple of anti-lockdown conspiracism in yoga/wellness is that public health measures are oppressive, authoritarian, and stifling what it means to be human. Masks suppress identity and speech. Vaccines assault the sovereign body. Influencers must rise up against tyranny! /1

Ironically, pandemic management in economies where yoga/wellness is lucrative has been a neoliberal, choose-your-own-adventure mess in which centre-right governments barter with public health to enact a patchwork of half-measures that punish the poor and marginalized. /2

The goal is not universal health rights, or protecting medical systems, or saving as many lives as possible. The goal is to return to business-as-usual, using public health as a pit stop on a productivity race off a cliff. The ice caps aren’t melting fast enough, I guess. /3

Right-leaning governments, like influencers, have assessed the COVID threat with the magical bias of limitless optimism. Then they work to normalize COVID casualties as a new cost of doing business. *This* is what alt-healthers and Qamoms call “socialist”, or “communist”. /4

Centre-right governments and wellness influencers share a goal: get back to the optimism of making productivity heroic and consumption virtuous, without any reflection on structural issues of injustice, or what universal health care funding and standards would look like. /5

The importance of T-cell immunity!
A recent study has found that after one dose of Pfizer’s vaccine, patients had a strong T-cell response, and a boosted antibody response after their second dose. This is good news for many reasons so let’s read up.

This study found that people’s antibodies were moderately effective against the original strain after their first dose of Pfizer’s vaccine, less effective against the B.1.1.7 variant, and were unable to neutralize the B.1.351 variant. Read what I said, that’s only after ONE dose.

However, they had strong T-cell responses against ALL known variants after the first dose. So why is this important? To recap, T-cells stimulate B-cells to make antibodies. Antibodies are just your first line of defense which is what is initiated when you get a vaccine.

It’s our T-cells that are responsible for long-term immunity. When antibodies diminish after your initial inoculation, your T-cells-will tell your B-cells it’s time to produce more antibodies. As long as your T-cells still recognize this virus and inform your B-cells they need

to produce antibodies, the vaccine is still doing its job. Your antibodies will get built up over and over again. Just FYI your TCRs and your BCRs target different parts of the Coronavirus’ spike protein. Antibodies can prevent infection, your T cell-responses ensure that those