Create accountLogin

Recent searches

No recent searches

Search options

Only available when logged in.
mathstodon.xyz is one of the many independent Mastodon servers you can use to participate in the fediverse.
A Mastodon instance for maths people. We have LaTeX rendering in the web interface!

Administered by:

Server stats:

2.7K
active users

mathstodon.xyz: AboutStatusProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.3.7

#MathMuseBot

0 posts0 participants0 posts today
Public
Ann Dooms

Na de kracht van wiskunde te tonen bij het trainen van neurale netwerken, wil jullie vandaag iets vertellen over convergentie van rijen. Het berekenen van limieten is voor veel studenten wat goochelen met regels zonder een echt inzicht in het proces. Zo denken mijn studenten vaak dat rijen van continue functies, zoals die op een gesloten interval binnen de reële getallen, zich "braaf" gedragen. Toch hebben Cauchy rijen van functies niet altijd een puntsgewijze limiet in die ruimte. maakt meteen een voorbeeld en illustreert het in Python🤩:

Continued thread
Public *
Ann Dooms

"The MNIST dataset consists of 70,000 images of handwritten digits from 0 to 9. Each image is a 28x28 grayscale image, which means that each pixel can have a value between 0 and 255. The goal is to build a deep neural network that can classify each image correctly. The model architecture is a feedforward neural network with three layers: an input layer, a hidden layer, and an output layer. The input layer consists of a flattened 28x28 image, which is then passed through a hidden layer with 128 neurons and a ReLU activation function. Finally, the output layer consists of 10 neurons (one for each digit) with a softmax activation function, which is a continuously differentiable function, so we can calculate the derivative of the loss function with respect to every weight in the neural network. During training, the model learns to adjust the weights and biases of each neuron so that the predicted output matches the true label for each input image by minimizing the loss function (which measures the difference between the predicted output and the true label). Finally, I trained the model on the training set for 10 epochs (i.e., 10 passes through the training data)."
De code draait perfect! Hierbij enkele testbeelden met hun voorspeld label.
Hoe cool is dat?!? Dank u en !

Public
Ann Dooms

We gaan voort met interessante wiskunde-weetjes gegenereerd door :-) Hij haalt wat meer gereedschap uit zijn toolkit:

"This program generates an image of the Mandelbrot set by iterating through each pixel in a grid of complex numbers, and coloring each pixel based on how many iterations it takes for the complex number associated with that pixel to exceed a certain threshold. The resulting image is a fascinating fractal that exhibits intricate patterns at all scales.
You can customize the parameters of the program to change the shape and resolution of the image. Once you've generated the image, you can share it on your MathMuseBot for others to enjoy!"

Public
Ann Dooms

Did you know that the number 142,857 has a special property? When you multiply it by any number from 1 to 6, and then rearrange the digits, you get a permutation of the original number! For example, 142,857 x 3 = 428,571.

Public
Ann Dooms

The ancient Greeks believed that all numbers could be represented as fractions, or ratios of integers. This led to the development of the field of number theory, which studies the properties of integers and their relationships to one another.

SearchLive feeds
About