Thank you for sharing. One thing I found confusing was the introduction of l_{i}'(x). I kept thinking this was short for \frac{d}{dx} l_{i}(x), but it's not the derivative!
I've always thought the use of "Tensor" in the "TensorFlow" library is a misnomer. I'm not too familiar with ML/theory, is there a deeper geometric meaning to the multi-dimensional array of numbers we are multiplying or is "MatrixFlow" a more appropriate name?
I was thinking it would be cool to add another slider for day of the year. That way you can see how the differences change as local timezones roll through their own daylight savings time or other adjustments. For example, right now Asuncion Paraguay is +3 hours ahead of Chicago U.S. but in April it will be only one hour ahead.
I believe they want to signal to the rest of the world that they are still a stable alternative for western nuclear fuel buyers. Uranium is the spice, and the spice must flow. Of course they also export coal, gas and oil too.
Thank you so much for creating this! Under the Curve61 point addition example, I was trying to follow the formula for adding two points: P:(x1, y1) + Q(x2, y2) = R(x3=l^2-x1-x2, y3=l(x1-x3)-y1) where l=(y2-y1)/(x2-x1). I tried to use the example P:(5, 7) + 23P:(2, 24) = (226/9, 2888/7) != 24P:(59, 55) and was wondering where I've gone wrong? Appreciate your response!
Much respect to the ScummVM team as they allowed me to play the original "Secret of Monkey Island" after it left my local library! Ironically, I no longer needed the "Dial a Pirate" code-wheel.
I actually read the same thing just a short while ago: https://valueinvesting.substack.com/p/why-ant-financials-ipo.... The author states Ant would have operated by repackaging highly leveraged micro-loans in such a way that none of the risk would appear on anyone's balance sheet.
I know very little about CNNs. But, I noticed the ReLu Activation step is Max(0,x) where x is the sum of the pixel intensities from each channel. In this example, it appears x > 0 (for all x) and so the activation step isn't really doing much?
EDIT: I'm wrong. x < 0 for some of the pixels. Specifically for the more red-ish channels.
The gentle gradients and transformations from black & white to color are charming. It made me think of the scene in Wizard of Oz where Dorothy steps into Technicolor.
Logistically speaking, I wonder what's the best way to create a tree map/database for my own town? Google maps will go a long way towards getting location. Is the rest just driving/walking around collecting data (species etc..) manually?
My father owned a tree service in the town I grew up in. We spent countless hours driving around town, looking at different trees and giving estimates for trimming and removal. I could imagine leveraging this data as almost a Zillow for trees.