John Purcell

Founder of

[John Purcell @CaveOfProgram]

Neural Network: Squares

Fairly simple of a neural network in Python, just squaring numbers.

https://t.co/1LDkx6B9AG

[John Purcell @CaveOfProgram]

Classifying Irises with a Neural Network.

There's a lot of new ideas in one post if you've not studied this previously, but in future posts I mean to break it all down.

https://t.co/3JaFqrV4YF

[John Purcell @CaveOfProgram]

Using PCA and Logistic Regression to Predict Breast Cancer Diagnosis.

Unfortunately not useful unless you have whatever machine measures the shapes of cell nuclei though, plus a biopsy.

https://t.co/GEWkwKWWxs

[John Purcell @CaveOfProgram]

I'd love to make a course on machine learning in Java. But libraries for Java are so lacking compared to Python.

You can build neural nets in Java, but suppose you want to visualise data. In Python there's Seaborn and Pyplot. In Java we seem to be stuck with primitive stuff.

[John Purcell @CaveOfProgram]

@Thanos__Soulis Cheers Thanos!

[John Purcell @CaveOfProgram]

Well since I started posting Python stuff, I've lost two followers and gained one.

Clearly I'm going to have to get back on the Java at some point.

Maybe I can look at how to do Python-type AI stuff in Java ....

[John Purcell @CaveOfProgram]

Principal component analysis with agglomerative clustering for the Iris Flower Dataset.

https://t.co/fJy0kKMpwy

[John Purcell @CaveOfProgram]

Principal Component Analysis: an intuitive explanation. Hopefully.

https://t.co/TEsUAzcvwW

[John Purcell @CaveOfProgram]

Adding to Clusters: Nearest Neighbours

A post on what to do if you've found some clusters in your data but then want to figure out which of these clusters new data belongs to.

https://t.co/44OGaH0sQ5

[John Purcell @CaveOfProgram]

Scikit-Learn Agglomerative Clustering
https://t.co/b1LO2rSTD1

[John Purcell @CaveOfProgram]

Hierarchal Clustering: Dendrograms https://t.co/mTpfeGE89Y

[John Purcell @CaveOfProgram]

Clustering Irises with DBSCAN https://t.co/yUf9U1gwvs

[John Purcell @CaveOfProgram]

Nearest Neighbours: Finding Epsilon for DBSCAN https://t.co/TQEFVPb1mu

[John Purcell @CaveOfProgram]

If I'm writing tutorials, what would you prefer to know about?

[John Purcell @CaveOfProgram]

There are going to be a lot of horrific AI arms races between generators and detectors, and when the dust settles, things are going to be a little different.

[Santiago @svpino]

50% of StackOverflow traffic is gone!

Look at the attached chart. It tells a scary story that will not be limited to StackOverflow.

Right now, detecting AI-generated content is impossible.

Last week, OpenAI shut down the tool they created for this purpose. They launched it in January, and it's dead today, less than seven months later.

Their statement: "The AI classifier is no longer available due to its low rate of accuracy."

I'm not surprised about any of these two events.

I don't remember the last time I visited StackOverflow. Why would I when tools like Copilot and ChatGPT answer my questions faster without making me feel bad for asking?

And I'm even less surprised about OpenAI killing their tool: Many believe detecting AI-generated text is impossible. I'm one of them.

Here's what OpenAI had to say about this:

"We are (...) currently researching more effective provenance techniques for text, and have made a commitment to develop and deploy mechanisms that enable users to understand if audio or visual content is AI-generated."

Notice how they differentiate text from audio and visual content. For the latter, they seem confident they'll find a way to recognize humans from AI. For text, they are not and are word-salad'ing us with a vague "researching more effective provenance techniques."

StackOverflow famously banned any AI-generated answers from the site.

That's the wrong move.

Instead, we need to find a way where human and AI-generated content coexist and benefit from each other. There's no putting the genie back in the bottle, so how can we get the most out of it?

Do you think StackOverflow will survive? What can they do to fend off what seems to be a life-threatening event?

[John Purcell @CaveOfProgram]

@Aceeeee64815692 Wayback machine has it!

https://t.co/nUcI53qguK

[John Purcell @CaveOfProgram]

Python machine learning: today I've got a tutorial on using confusion matrices to figure out which values have been classified correctly or incorrectly by your model.

https://t.co/ESyNWlHtpg

[John Purcell @CaveOfProgram]

Today I've got an article on the "elbow method" for finding the optimum number of clusters in data, particularly with K-means clustering.

Quite pleased with my supermarket analogy here!

https://t.co/DDO4M3lKA9

[John Purcell @CaveOfProgram]

@Aceeeee64815692 Ah sorry, that link might have permanently gone AWOL. My blog got deleted because I forgot to pay the hosting costs.

There are tons of free Java tests online though if you search. Mine wasn't especially in-depth.

[John Purcell @CaveOfProgram]

Python machine learning tutorial: today I've got an article on KMeans clustering.

An interesting thing about clustering is that it's widely misunderstood by non-statisticians. People commonly argue that if clusters aren't perfect, they don't exist.

https://t.co/EYbwxrQZta

[John Purcell @CaveOfProgram]

Thanks to this thread I finally got my nuclear bombs working properly.

[Bojan Tunguz @tunguz]

After seeing #Oppenheimer a lot of people are trying to make their own nuclear bomb. However, 99% of them are doing it wrong. Here is what you need to know in order to do it correctly.

A thread👇🧵🧶 1/235