markisus

15 ms·

181.
by markisus 3y ago
Minimizing Minkowski difference between convex bodies can also be formalized as a quadratic program. I wonder how the GJK performs against the QP formulation.
182.
by markisus 3y ago
Nit: The most likely outcome is zero return, but expected return is positive (if the investor is rational) due to tail probability.
183.
by markisus 3y ago
Do you think 1D could work? Maybe topic-space is some sort of tree-shaped structure where documents live in the thin strands.
184.
by markisus 3y ago
I don’t think this is realtime reconstruction, meaning that changes in the scene will not be represented in the reconstructed map immediately.
185.
by markisus 3y ago
They are only waves until you “measure” their position whatever that means. Then they behave like particles.
186.
by markisus 3y ago
The author's workload is sort of different than the usual ML workload since the author's expression tree is large (10k nodes), while a modern neural net has a relatively smaller expression tree, maybe fewer than 100 for the larger
187.
by markisus 3y ago
It seems like the author is claiming to show P != NP using a new form of information theory? Edit: The claims in the abstract about P and NP were conjectures that the authors plan to investigate in a later paper.
188.
by markisus 3y ago
I think they took a picture of a vase, and created two derivative images of the vase, one which was more cat-like and one which was more truck-like. In both cases, participants did better than average at guessing which one was the derivativ
189.
by markisus 3y ago
This is a relatively recent paper you might find interesting. https://arxiv.org/abs/2204.03771 A random forest is also a universal function approximator so anything a neural net can do, so can a random forest (in theor
190.
by markisus 3y ago
It would be interesting to see these classical methods compared to a simple RL method that optimizes a tiny neural net or decision tree which is given access to the map, player, and monster positions. I think it would cost a tiny bit more c
191.
by markisus 3y ago
In fact, a neural net is “just” a stack of logistic regressions if you only use sigmoid activations.
192.
by markisus 3y ago
Familiarity is part of the equation. Writing an efficient web server requires being familiar with the imperative model since the underlying specs, protocol reference implementations, and libraries are probably given in imperative style. Fam
193.
by markisus 3y ago
> Like dieters falling for magic 10-minute-miracle exercise gadgets, developers seem ready to fall for easy solutions to the latest crises in their field. > Just like "mostly secure," "mostly pure" is wishful think
194.
by markisus 3y ago
The conclusion seems to contradict the thesis. > The most enjoyable time is when you feel yourself growing, but the real test is whether you can persevere when you hit that wall or if you’ll lose interest and give up. Why does it matter
195.
by markisus 3y ago
The referenced paper [1] invents a version of evolutionary algorithm where each gene is associated with a target "population ratio" which is modulated upwards or downwards by that gene's performance, and ultimately decides ho
196.
by markisus 3y ago
This is a case where a picture is worth a thousand words. The slicing intuition is quite useful and I imagine quite common for those who have taken calculus. It is the very same mental picture as you would use for computing area using two n
197.
by markisus 3y ago
I consider exp(At) to be an exponential expression in A. Depending on A, you may get a matrix whose entries are quadratic in t, as in this case. For another case, you may have trigonometric entries in t, as in the case of A = [0, 1; -1, 0].
198.
by markisus 3y ago
Solutions to linear differential equations are exponentials, and this may be what the author meant. In a state space model, we can write Newton's law under gravity if we let y(t) = [x(t), x'(t)]^T, let A = [ 0, 1; 0 0], let b = [0
199.
by markisus 3y ago
Where can you get a robot arm like that for $4k? On the other hand, I found this ($34k) https://www.digikey.com/en/products/detail/kuka-robotics-cor...
200.
by markisus 3y ago
Maybe Miso will ultimately foot the bill for the hardware. As I mentioned, the current system probably isn't capable enough to replace the equivalent of one full time employee. And employees must be paid for in discrete units, so you w
201.
by markisus 3y ago
The embedded video has a little bit more information than the text article. I have a couple of thoughts here. 1. They mention the fry tending robot (Flippy) costs $4,000 which seems quite off to me. It seems to be an off the shelf robot arm
202.
by markisus 3y ago
I tried to find more information about this and I only saw a location where the drive-thru was automated. I think it would be very difficult for McDonalds to automate the food preparation and assembly steps, given the current state of the a
203.
by markisus 3y ago
They discovered a quantized version of Newton’s law because their processor had no floating point unit. Pretty creative. > S(0)=0 (reference position) S(1)=v₀-½g (initial velocity) S(t+2)=2S(t+1)-S(t)-g Thus, the value of the previous sc
204.
by markisus 3y ago
Racism is a societal disease. But it doesn’t follow that racial quotas is the cure. That is the Politician’s Fallacy. https://en.wikipedia.org/wiki/Politician%27s_syllogism
205.
by markisus 3y ago
For completing online orders there is a privacy issue but is there anything actually wrong with the drive thru processing startup? It seems smart to start with outsourcing to prove the business model and then replace with AI once the produc
206.
by markisus 3y ago
> It's possible but it doesn't happen by accident, you have to show that they are related, and no one does (because the people who work on LLMs aren't actually interested in this question, despite the fact that argue it to
207.
by markisus 3y ago
Even if you only care about fish, it is possible that the pursuit of a swimming machine will yield insights into how fish do it. In terms of scientific understanding in general, I'm not sure why you would say that building a statistica
208.
by markisus 3y ago
Chomsky gets it wrong here. A fairer analogy can be made by first supposing that submarines do not exist yet, and people are intensely interested if one could actually be built. Add in the situation that many people are claiming that such a
209.
by markisus 3y ago
But many people did believe that. From the article, > I personally remember very heated arguments on the MIT campus between Chomsky followers and computer scientists working on statistical models in the early 2000’s. The arguments came u
210.
by markisus 3y ago
Along these lines, Gary Marcus said that "you can't build a ladder to the moon" [1]. In other words, neural nets do well on simple things, but general intelligence is such a high bar that neural nets (and statistical methods

More ›