Zephyr314

14 ms·

91.
by Zephyr314 11y ago
Great points. This was just a cool demonstration of the underlying technology hoping to inspire people to use it on their models in domains other than prop betting. Potentially it could help those sharks already swimming in these pools, but
92.
by Zephyr314 11y ago
We're also really passionate about building optimization platforms vs betting schemes. We believe in the long term we can have a better positive impact on the world by helping every expert cut the trial and error out of their workflow
93.
by Zephyr314 11y ago
Great idea! You could reuse most of the code [1] to give this a shot. In future posts we may extend this to other sports as well. [1]: https://github.com/sigopt/sigopt-examples
94.
by Zephyr314 11y ago
We give a high level overview in footnote #4 of the post, but more detail can be found in the code [1]. We tried to pick a relatively small set of features (and kept those picks constant) in order to isolate the model tuning gains from pure
95.
by Zephyr314 11y ago
Vegas always gets their cut as you point out. You need to reliably win 52.5% of the time to beat the bet $110 to win $100 edge. In general, betting schemes tend to stop working in the long term, but can have large gains in the short term [1
96.
by Zephyr314 11y ago
Post author and co-founder of SigOpt (YC W15) here. Thanks for all the great questions and comments, I'll be around all day answering questions. Feel free to ask anything about the post or what we do at SigOpt (or how). All of the code
97.
by Zephyr314 11y ago
If you decide to grab the code and have any questions please feel free to reach out.
98.
by Zephyr314 11y ago
Thanks! Great question. There is a huge variance in the cumulative gains as you pointed out. Because of the 100 win/110 loss edge that Vegas has you need to win over 52.5% of the time to register a profit. Even with a model that can wi
99.
by Zephyr314 11y ago
Thanks! Let me know if you have any questions.
100.
by Zephyr314 11y ago
Hello, I'm the author of this post and co-founder of SigOpt (YC W15). Let me know if you have any questions about the post or what we do at SigOpt. All of the code used in this post can be found at https://github.com/si
101.
Using Model Tuning to Beat Vegas(blog.sigopt.com)
3 points by Zephyr314 11y ago|7 comments
102.
by Zephyr314 11y ago
One of the co-founders of SigOpt (YC W15) here, I'm also happy to answer any questions about what we do and how.
103.
Profile Likelihood vs. Kriging Variance(blog.sigopt.com)
3 points by Zephyr314 11y ago|1 comments
104.
Likelihood for Gaussian Processes(blog.sigopt.com)
40 points by Zephyr314 11y ago|26 comments
105.
Automatically tuning text classifiers (with Bayesian Optimization)(blog.sigopt.com)
3 points by Zephyr314 11y ago|0 comments
106.
Approximation of Data (and why it matters for optimization)(blog.sigopt.com)
1 points by Zephyr314 11y ago|0 comments
107.
by Zephyr314 11y ago
Yes, if you could expose the parameters to MOE from within TensorFlow. We're going to work on an example of showing how this can be done with SigOpt, which has a similar interface.
108.
by Zephyr314 11y ago
Depending on how long it takes to "evaluate the goodness of a solution" techniques like multi-start gradient decent can rapidly become intractable though, especially in higher dimensions. There are a handful of open source librari
109.
by Zephyr314 11y ago
For those who missed it the first time around this is the post referenced: https://news.ycombinator.com/item?id=10486476
110.
by Zephyr314 11y ago
I'm one of the founders of SigOpt (YC W15) and will be checking this post all day and am happy to answer any questions about this, our tech, or anything.
111.
Intuition behind Covariance Kernels(blog.sigopt.com)
12 points by Zephyr314 11y ago|1 comments
112.
by Zephyr314 11y ago
Author here (and co-founder of SigOpt). I'm happy to answer any questions about this post, our methods, or anything.
113.
Making a Better Airplane Using SigOpt (YC W15) and Rescale (YC W12)(blog.sigopt.com)
2 points by Zephyr314 11y ago|2 comments
114.
by Zephyr314 11y ago
Hi, I'm one of the founders of SigOpt, I'm happy to answer any questions about this post, GPs in general, or SigOpt. I'll be checking this post all day.
115.
Hats Off to Web Advertising. No, Really(google.com)
1 points by Zephyr314 11y ago|0 comments
116.
SigOpt (YC W15) raises 2M led by a16z and DCVC to optimize everything(blog.sigopt.com)
1 points by Zephyr314 11y ago|0 comments
117.
State of Hyperparameter Selection(startup.ml)
14 points by Zephyr314 11y ago|2 comments
118.
by Zephyr314 12y ago
Thanks for the feedback! We're constantly trying to iterate on our landing pages, trying to make the trade-off between being technical and accessible and we'll use this to iterate again. There are two ways to use us as a mobile ga
119.
by Zephyr314 12y ago
One of the founders of SigOpt here, I'm happy to answer any questions about this post, what we do, or anything about SigOpt. I'll be around all day!
120.
Picking the Right Metric(blog.sigopt.com)
10 points by Zephyr314 12y ago|3 comments

More ›