Bailey Andrew (@EBaileyAndrew) profile picture

Knows some stuff, doesn’t know other stuff

Joined May 2021
Photo by Bailey Andrew (@EBaileyAndrew)Photo by Bailey Andrew (@EBaileyAndrew)Photo by Bailey Andrew (@EBaileyAndrew)Photo by Bailey Andrew (@EBaileyAndrew)Photo by Bailey Andrew (@EBaileyAndrew)Photo by Bailey Andrew (@EBaileyAndrew)Photo by Bailey Andrew (@EBaileyAndrew)Photo by Bailey Andrew (@EBaileyAndrew)Photo by Bailey Andrew (@EBaileyAndrew)Photo by Bailey Andrew (@EBaileyAndrew)
If we look at a cube from the correct angle, its outline will trace a hexagon rather than a square. Instead of building the cube out of square sides, we can build it out of these hexagons. This gives us a 4-sided, hexagon-faced, self-intersecting 'cube'... 😅 (1/3)
1
1
112
We can do a similar thing for each platonic solid - these are called "Petrials". Is this useful? Probably not, but it sure took a lot of work to make. (2/3)
1
81
Petrials don't have flat faces, so images online show wireframes... But a Petrial cube has the same wireframe as a normal cube! I used minimal surfaces to draw them. Wireframe from Wikimedia commons.wikimedia.org/wiki/F… (3/3)
1
78
Bailey Andrew retweeted
Humans can learn general-purpose and meaningful representations from perception by leveraging suitable abstractions. For example, being able to write out numbers involves the generic ability to reason about what strokes to use… [1/6]
2
27
1
158
Numpy broadcasting ‘cheat sheet’. Useful trick to know, can often give orders of magnitude speedup, especially if your old solution used for loops [yuck :P]
Bailey Andrew retweeted
Here is a game based on the impossible staircases typical to Nil geometry. The road track here acts like an impossible staircase, so our unicycle, powered only by gravity, can gain speed just by going in counterclockwise circles on it! (RT to show you want this fleshed out! 1/3)
7
41
2
129
Many feel as if their actions won't make a difference - and they won't, if they do nothing - so they offer only their thoughts and their prayers. Email your representative. What does it cost you? Less than the sin of the bystander costs others: 🇺🇦
What does the number -1 actually mean? I tried making a (1-minute) video to explain; youtu.be/J81UnII6X7w. It's of dubious quality ;)
1
Something that was not clear in the video; for an 'Action Number' X, if you start with a rubber band on your left finger at the origin and right finger at "1", then X is the action that results from you moving your right finger to where X would be on the number line.
Foolproof plan to end world hunger; just draw some lines on a graph. Since correlation clearly equals causation, the rest will work itself out.
Suppose we had a set of rules describing a type of object, such as "eats grass and has a tail". We might want to know what type of objects follow those rules. For example, the set of all cows follows ("models") these rules, and the set of all sheep also models the rules. (1/3)
1
Thus, it is impossible to *uniquely* define what it means to be a number 0, 1, 2, 3, etc in a "sufficiently nice" way! (3/3)
1
Here, "sufficiently nice" means the rules are in First-Order Logic (i.e. built by "If A, then B", "A and B", "A or B", "not A", "everything", "at least one thing"). There are more powerful types of rules which can uniquely define things, but they have some major drawbacks.
Maximum Modulus Principle: For any "holomorphic" function, the maximum value in *any* region of that function will always appear on the boundary of said region. (For regions in the complex plane). (1/2)
1
1
"Holomorphic" is a technical condition, but it basically just means "sufficiently nice"; most functions you can think of, such as polynomials, trig functions, etc are holomorphic. (2/2)
1
Slight correction; not all trig functions are holomorphic. Basically, anything that involves division is probably not holomorphic, as you could divide by zero which is not 'nice'. But they are 'meromorphic', which basically means "nice but sometimes divides by zero". (3/2)
Bell curves show up all the time when one talks about statistics. Some classic examples are height and IQ. But why should this happen? Let's look at a few examples of different real-world(ish) processes that end up bell-curvy. (1/10)
1
For example, we described the set of height genes with the Binomial distribution, and the set of eureka moments with Erlang. Of course, it is possible to create non-bell-curvy sets - but under reasonable assumptions, many sets become bell-curvy as they grow. (9/10)
1
For the curious, the fundamental underlying reason for all this is the "Central Limit Theorem". (10/10)