cgreerrun

7 ms·

31.
by cgreerrun 4y ago
Collecting fog electrostatically is 2x more efficient than desalination. It works by ionizing the droplets in the air and collecting them on wire meshes: https://www.science.org/doi/10.1126/sciadv.aao5323 . The tec
32.
by cgreerrun 4y ago
Maddness is their more recent work and yields 100x speedups: https://arxiv.org/pdf/2106.10860.pdf The code for Maddness is in the same github repo if you search for "Mithral". SIMD instructions can work wonde
33.
by cgreerrun 4y ago
I think it is notable because he's an expert in that area and would know how to assess it. I'm not a fan of the tone either though... But the authors claim "Hopular surpasses Gradient Boosting (e.g. XGBoost), Random Forests,
34.
by cgreerrun 4y ago
Also more context about criticism from the paper's authors here: https://twitter.com/bschaefl/status/1532281938593173504
35.
by cgreerrun 4y ago
Notable criticism here: https://medium.com/@tunguz/trouble-with-hopular-6649f22fa2d3
36.
by cgreerrun 5y ago
With a little formatting to make it feel like a fiction book, the writing style works much better for me: https://imgur.com/a/5R3x2BL Love the article and the writing style! /* CSS changes for fiction book sty
37.
by cgreerrun 5y ago
I haven't personally, but I know people that have. The people that I know that sold shares used Equity Zen ( https://equityzen.com/ ). The company I worked for has/had a "first right of refusal" option in
38.
by cgreerrun 5y ago
+1 to sibling comment about not choosing the SOTA just cuz SOTA (if you're in the market for an ANN algo). Almost all the ones there are pretty darn fast these days and many (e.g. hnsw) are implemented out of the box by libraries like
39.
by cgreerrun 5y ago
This site keeps track of them: https://github.com/erikbern/ann-benchmarks#glove-100-angular ScaNN is currently SOTA: https://ai.googleblog.com/2020/07/announcing-scann-efficient...
40.
by cgreerrun 6y ago
> Thanks for sharing! You're welcome. > Why did you use GBDTs instead of NNs? I mostly wanted to build an implementation to see how it worked; I was more familiar with GBDTs than NNs, so I figured I'd start with that. At its
41.
by cgreerrun 6y ago
What you can do is checkout the algorithm at particular stages of development. AlphaZero&Friends start out not being very good at the game, then over time they learn and eventually become super human. You typically checkpoint the weight
42.
by cgreerrun 6y ago
I wish the code panel was collapsed by default on the repo home page, so that you could see the README section front and center. That's usually the first thing I read. Then I dive into one or two files to check things out.
43.
by cgreerrun 6y ago
> because convolutional networks provide the right inductive bias for the problem of learning to play Go. Agreed, seems like there will be a set of environments that GBDT won't work on and a CNN will. I'm only mildly familiar w
44.
by cgreerrun 6y ago
I'm trying to answer that right now, actually. For connect 4, once it's trained a bit it seems to do really well. At 800 MCTS playouts (<.4s), it goes from "I can beat it and sometimes we draw" before training to &quo
45.
by cgreerrun 6y ago
> Are there any papers/comparisons/tradeoffs on when GBDT predictive-power plateaus compared to a NN? None specifically that I know of, but I haven't searched. "Shallow learning" GBDTs can do pretty well on MNIST
46.
by cgreerrun 6y ago
A fast, powerful bayesian model seems like it would be a game-changer. PUCT (the heart of the AlphaZero MCTS that decides which action to choose) really seems setup to model the action choices as a multinomial bayesian inference problem (it
47.
by cgreerrun 6y ago
I'll be sure to let you know how it goes. I have a few self-play experiments that I'm trying out once I get the testing setup finished. Do you know if there is a discord channel for all of us AlphaZero nerds that I'm missing?
48.
by cgreerrun 6y ago
Here was the guestimation: https://www.yuzeh.com/data/agz-cost.html
49.
by cgreerrun 6y ago
I've been working on a Python implementation that uses Gradient Boosted Decision Trees (LightGBM/Treelite) instead of using a neural network for the value/policy models: https://github.com/cgreer/alpha-ze