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 👇

One of the most popular activation functions used in deep learning models is ReLU.

I asked: "Is ReLU continuous and differentiable?"

Surprisingly, a lot of people were confused about this.

Let's break this down step by step: ↓

Let's start by defining ReLU:

f(x) = max(0, x)

In English: if x <= 0, the function will return 0. Otherwise, the function will return x.

If you draw this function, you'll get the attached chart.

Notice there are no discontinuities in the function.

This should be enough to answer half of the original question: the ReLU function is continuous.

Let's now think about the differentiable part.

A necessary condition for a function to be differentiable: it must be continuous.

ReLU is continuous. That's good, but not enough.

Its derivative should also exist for every individual point.

Here is where things get interesting.

We can compute the derivative of a function using the attached formula.

(I'm not going to explain where this is coming from; you can trust me on this one.)

We can use this formula to see whether ReLU is differentiable.

You are building a binary classification model to determine whether a picture shows a hot dog or not.

Which of the following activation functions is a good candidate for the output layer?

1. Rectifier Linear Unit (ReLU)
2. Leaky ReLU
3. Sigmoid
4. Softmax

Think about it.

The answer:

You could use either Sigmoid or Softmax.

I'm sure most of you picked Sigmoid. It's certainly the best way to approach this problem.

But Softmax is also an option —although unconventional for a binary classifier.

How about the other choices?

ReLU returns its input if positive, zero otherwise. Leaky ReLU is based on ReLU but with a slight slope for negative values.

ReLU and Leaky ReLU will always give us the previous layer output with no modifications if the input is always positive.

This doesn't help us.

Sigmoid has the ability to squash its input into a range that goes from 0 to 1.

This is good because we can now set a specific threshold for the output of a binary classifier.

Here is an example.

Any image with an output equal to or under 0.5 doesn't show a hot dog. Any image with an output above 0.5 shows a hot dog.

Sigmoid makes it very easy.

How about Softmax?

A surprising situation.

A machine learning model predicting 3 classes ends up with these results:

• Class A: 90% accuracy
• Class B: 80% accuracy
• Class C: 70% accuracy

Class C seems to be the worst-performing.

But you know already that I'll say it isn't. ↓

Here is the correct answer: We don't know yet which class is the worst.

This is important to understand, so keep reading.

Many people make this same mistake every day and try to fix things that ain't broken.

We are missing something, so let's go and get it.

How about if we ask a friend to look at the dataset manually?

Let's say these are the results of our friend doing the job:

• Class A: 95% accuracy
• Class B: 100% accuracy
• Class C: 71% accuracy

Do you see what just happened?

The model is actually really accurate when predicting Class C! 70% versus 71% when done by a human.

Class B, however, is a disaster: 80% versus 100% when done manually.

Class B seems to be the worse performing class!

The work that our friend did is called a "baseline."

Without having it, it's hard to determine how good are the results of a model.

I've seen tons of money down a drain just because teams didn't create a baseline beforehand.

But now, that won't happen to you 😉.

New Cannabis Based Menu for Chao Phraya Hospital patients (service being rolled out to other hospitals too)
"This cannabis special menu is bringing different parts of the cannabis plant to use with leaves to cook as ingredients in food" #cannabis #Thailand

Whilst patients in the UK are still being denied with almost no progress with #NHS Amazing to see how much progress is being made with #cannabis as #medicine in #Thailand is the same 2 year period:

A Thread: 1/10 #Thailand allowed #cannabis #medicine in Jan 2019, 2 months after the UK.
NOW:
Is approved for 160+ conditions (everyone!)
Village doctors, clinics and hospitals can supply approved products or grow and produce their own, from nearly 100 approved formulas

2/10 - Average monthly prescription cost is £20-£30.
Educational programs started for the Medical Profession within 3 months, for many others at 6 months and school kids after a year. (Google Dr Ganja - see photo)

3/10 - Shortly they will launch a new policy for any individual or collectives to grow #cannabis for #medicine for themselves and any surplus material can be sold to the government for use in centralised production.

Right, screw it. Here's a thread on *that* FBLive by Alex Cairnie about #Govanhill. Lets do a watch-through shall we? (Mind, I'm a resident of Govanhill, and a sociologist. I'm also 5'3, and a visibly queer woman so if half his claims are true I should be living in terror)

00:30: Welcomes viewers and uses points out that his videos are only getting attention b/c they're 'horrific'. Directs folk to his Youtube page. So already clout hunting.

1:08: "I've been warned to stay out of Govanhill. I'm a marked man" (proceeds to use a beanie as a disguise)

1:30 "Laughing and joking aside, here's somebodys front garden." Proceeds to show litter and claim its thrown from windows and this is 'common'. I know a handful of streets with litter issues like this, but its neither common nor unique to #govanhill and other deprived areas.

2:10 "Don't be under any illusion I'm only showing you the bad bits" - only shows you the bad bits.

3:45: He's had offers of help, but hes said no because hes not responsible for 'anybody elses safety' Good lord.

5:01: "Its not that bad. But its not all good either" -Correct

6:00: TW: child abuse - He starts claiming to have been contacted by a health professional and told about child abuse 'in the area'. Huge ethical issues here as the 'professional' who has 'been in the game a long time' is reporting to Facebook Dude and not authorities.

So many people don’t realize that they have OCD because actual obsessive compulsive disorder is rarely talked about openly. So here are the basics: 🧵(thread)

I think that a lot of people forget that OCD is a mental illness, emphasis on the MENTAL. When it’s portrayed on TV, it’s always shown as a very behavioral disorder. And while it’s true that overt behavioral compulsions are part of OCD, they don’t begin to tell the whole story

OCD stands for obsessive compulsive disorder. That’s obsessive thoughts and compulsive behaviors. The OCD cycle is the cycle of obsessions that lead to compulsions that lead to more obsessions that lead to more compulsions etc etc etc.

People with OCD experience intrusive thoughts (which become obsessions.) Intrusive thoughts are unwanted and invasive, and they’re generally REALLY disturbing. It’s a common misconception that people with OCD are exclusively concerned with tidiness and cleanliness. This is false

TW/ OCD subtypes

Intrusive thoughts can be about anything, but there are a few common “subtypes” of obsessions. These include harm OCD, self harm OCD, sexual OCD, pedophelia OCD, relationship OCD, contamination OCD, religious OCD, moral OCD, health OCD and “just right”

Me: Made a thread about how people don't understand hyperacusis and give advice that obviously I've already tried.

This person only a few tweets into the thread:

I only tweeted this because, like it or not, this is the general population's reaction to me talking about hyperacusis.

Similar replies happened when there was the whole "not clapping" article regarding Cambridge.

They're so quick to label disabilities as not real disabilities.

It's important to remember this when people may ask why we've never talked about this before, or why it's not talked about a lot.

Most people don't talk about their hyperacusis because if they do, people respond like this. So most of the time, honestly it's easier to say nothing

Yes I am petty enough to add screenshots of this toxicity to this thread:

Asking for accessibility is "pathetic." /s : )

They called it elephantiasis
A disease said to be caused by a strange mosquito
I was at the Bible College at the time
I had been sent for my Teaching Practice at a church located at Ogudu Ori-Oke
The pastor of this church was a gentleman
He and his wife were chartered accountants

and they were blessed
I was a personal assistant to another pastor and the coordinator of the Bible College had to practically beg my boss to allow me to observe the Teaching Practice at another church (He had insisted I do it with him and I would have spent the time carrying his

Bible and practicing nothing)
When I reported for duty, I was assigned the duty of follow up and evangelism
My primary assignment was to liaise with the follow-up lady and visit all the members who have not been to that church in weeks
I was also given the altar to minister one

Sunday a month (sermon) and other Sundays in various capacities like leading the church in prayer, taking testimonies, and closing prayer
We met this brother on the way to a member's house early one Sunday morning
He was sitting outside a building, wearing yellow slippers and

looking hopeless
We had walked past him but my spirit reached out to me, so I asked the sister to turn around with me so that we can minister to this brother
He was happy we came around
He said he had spent everything he had on the leg and he was desperate enough to accept any