nestorD

29 ms·

301.
by nestorD 6y ago
Me too! So far I have seen four actual bugs in large numerical code bases that were caused by overlooking operator precedence. I expect to see more in the years to come. I think that precedence of '*' over '+' is accepta
302.
by nestorD 6y ago
Is your camera at equal distance to the object in all photographs? NeRF has troubles with variation in the distance between the object and camera (which Mip-NeRF [0] solves). [0]: https://youtu.be/EpH175PY1A0
303.
Whole-History Rating: Bayesian Rating System for Players of TimeVarying Strength [pdf](remi-coulom.fr)
1 points by nestorD 6y ago|0 comments
304.
by nestorD 6y ago
While I do agree for MtG and Pokemon Yellow (their main value as examples is that they are fun and inexpected), Peano arithmetic is simple. It is fondamental but still probably the simplest way to implement natural integers : a natural nu
305.
by nestorD 6y ago
To further push the point that many things are Turing complete and that it should not be seen as synonymous with complex, Gwern has a nice list of Surprisingly Turing complete things [0] that includes Peano Arithmetic, Pokemon Yellow and Ma
306.
by nestorD 6y ago
The complexity of the language makes it hard to do even for state-of-the-art IDEs. I have had up to date versions of CLion suggest a dozen potential declarations in some hairy codebases...
307.
by nestorD 6y ago
It appears that the famous, often retold, story of the experiment were experts have been deceived by simple red colorants in white wine is a significant distortion of the actual experiment : https://sciencesnopes.blogspot.com&#x
308.
by nestorD 6y ago
You might want to try Zbrush (and free alternative like Sculptris). It really breached the distance between clay and 3D modeling for me.
309.
by nestorD 6y ago
Great! Don't hesitate to contact me if you have any question or roadblock. I would be happy to give a link to your implementation in the readme of the C++ reference implementation.
310.
by nestorD 6y ago
People here might also be interested in the Julia implementation of Shaman: https://gitlab.com/numerical_shaman/shaman_julia
311.
by nestorD 6y ago
I am not familiar with NSan (and impatient to read the paper) but a common problem with tools based on higher precision shadow values is that they can be easily defeated by a large cancelation (`(1e200 + 1) - 1e200` will return 0 in both 64
312.
by nestorD 6y ago
I am not familiar with NSan but other shadow value implementations will keep a higher precision copy of the numbers and follow them in the computation. They detect local errors but also take propagation (amplification and dampening of error
313.
by nestorD 6y ago
For those interested, MakeHuman [0] is the closest open-source alternative that I am aware of. [0]: http://www.makehumancommunity.org/
314.
by nestorD 6y ago
> Added language-integrated differentiable programming into the Swift language. This work continues in the official Swift compiler. While the readme tells us that the project is in archive mode, it also tells us that the differentiable p
315.
by nestorD 6y ago
Most people regret that Google did not invest in Julia but, to my mind, it misses the point. Swift for TensorFlow's unique selling point was bringing a strong type system and ML+autodiferentiation together. That, for me, would be a hug
316.
by nestorD 6y ago
I have not but I am going to file the idea, it would indeed be a good starting point.
317.
by nestorD 6y ago
Great ! It is not far from one of my dream project : numerising a maximum of old text and let historian do research on them using state of the art tools (that work across synonims and languages) with parameters to restrict by time of public
318.
by nestorD 6y ago
At the moment no [0]. My undertsanding of the arguments is that macros and the type system already get you 90% of the way at compile time (while being able to cause compile time error if you made a mistake rather than obscure run time crash
319.
by nestorD 6y ago
On that theme, the best description of love and the feelings it creates that I ever read is "A Lover's Discourse: Fragments" by Roland Barthes. By a long shot.
320.
by nestorD 6y ago
For me the main point is that JAX is slightly lower level than Pytorch and has nicer abstraction (no need to worry about tensors that might not store a gradient or wetehr you are on the GPU: it eliminates lots of newcomers bug) which makes
321.
by nestorD 6y ago
No really, it force you into pure, functional patterns. Frameworks build on top of JAX have done a lot of experiments to find a good way to deal with states while keeping the code clean and not too magicy. Their is no clear best pattern at
322.
by nestorD 6y ago
The authors gives that quote (from the JAX documentation) but does not seem to interiorize it as his conclusion says: > This is the niche that Theano (or rather, Theano-PyMC/Aesara) fills that other contemporary tensor computation l
323.
by nestorD 6y ago
> One problem, particularly in the US, is that portion sizes are so large. Portion size was the key for me. I stopped feeling heavy and lethargic after pasta meals once I started weighting my pasta before eating. This made a world of dif
324.
by nestorD 6y ago
The concept is intriguing. One thing I would do is trying to find books orthogonal to my tastes rather than opposites. I do not want to find books I know I will hate, I want to find books around axis I have not even considered.
325.
by nestorD 6y ago
The Rocket paper [0] for time serie classification works along those lines. They use a linear classifier on top of an ensemble of random convolutions. [0]: https://arxiv.org/abs/1910.13051
326.
by nestorD 6y ago
When I looked for a similar tool for a publication, this did not work for my use case as I have a fully-connected neural network with thousands of node per layer. I falled back to draw.io wich is not specialized in neural networks by great
327.
by nestorD 6y ago
This sounds like symbolic regression (where the space of possible equations is usually searched using genetic algorithm).
328.
by nestorD 6y ago
I have used Nbdev and I am not a fan. It creates friction when one wants to contribute (in my case, to fast.ai) and forces you to write your code with notebooks which is the point but also not great when you are writing code rather than pro
329.
by nestorD 6y ago
I can't speak for others but for me it is the tooling. Knowing that I can get great online doc with a single command is a very good motivation to start writing said documentation.
330.
by nestorD 6y ago
Slate star codex has a great article on gout (covering both causes and the strange correlation with fame and profit): https://slatestarcodex.com/2020/04/27/give-yourself-gout-for...

More ›