BONSAI: A Democratized, Low-Barrier Machine Learning / Artificial Intelligence Application

in machine-learning •  8 years ago  (edited)

In an exchange with @lemouth yesterday in one of my posts, he expressed that he's looking forward to using machine learning for developing new techniques for looking for new phenomena at particle colliders like the Large Hadron Collider.

On a personal capacity - while I do have a background in machine learning and artificial intelligence, it was based on a raw, non-modular implementation using Matlab back in 2007 - I have been out of the loop for quite some time now. Plus, I had nothing to do with deep-learning AIs. But just like anything to do with initially esoteric technologies, it's anyone's guess that it would eventually be democratized at one point where any casual researcher can do it easily as well.

So after doing some searches yesterday, I landed on this article by Steven Levy of Backchannel in Medium.com (my favourite channel ever since being able to bypass the Medium restriction in my country!): https://backchannel.com/you-too-can-become-a-machine-learning-rock-star-no-phd-necessary-107a1624d96b#.qldvmez3m

It's called BONSAI, and it's supposed to be an easy-to-use product for anyone wanting to implement ML / AI into their projects. Not sure if it'd be a good fit for LHC experiments though, as I'm still barely scratching the surface of these new ML / AI platforms!

Anyway, check it out here and sign-up for the beta: https://bons.ai/

Do you know of any other applications? Please recommend it down in the comments :)


Image is a screenshot of the bons.ai website.


Follow me @kevinwong

Authors get paid when people like you upvote their post.
If you enjoyed what you read here, create your account today and start earning FREE STEEM!
Sort Order:  
  ·  8 years ago (edited)

Interesting. I don't spend a lot of time reading about it these days, but the area of machine learning is a fascinating one.

It is the heralding of a new era, where the expertise of machines will begin to match and surpass the expertise of humans.

Perhaps we could use it to improve the bots on Steemit. :-)

Yup, although it's also important to recognise another dimension :)

Found here: http://www.futuristgerd.com/2015/11/10/every-great-algorithm-needs-a-great-humarithm/

Basically left / right, masculine / feminine intelligence, for those who like to put it that way!

That's a great infographic. It's important in that it defines the niche that still remains for us humans to inhabit. :-)

And trading bots!

I used to make trading bots for forex on certain pairs. Everytime I thought I've hit the holy grail of bots, quality testing data just isn't available to make it good enough. And considering the spread of trades, it actually quite less than 50% to hit take profit compared to stop loss. And not using stop loss for bots is a recipe for disaster.. although there's a chance to it may make one rich as well. Never worked that well for me, so far. Crypto is the best atm, since the upside is conceivably higher than forex pairs.

Forex always scared the shit out of me, crypto seems far more "democratic" and transparent.

Suffice to say, i spent years on it and never earned much :p

Apart from the knowledge!

Thanks for the suggestion (and for keeping the momentum with respect to our exchange of yesterday!).

I have bookmarked all of this. I am unfortunately busy with intense committee meetings until Mid March so that I am afraid I will not be able to start digging into that now, but I will definitely have a look when time will allow me to do so ^^

I'll be your virtual assistant on this research :).. since I need to learn it too for my own stuff! Anyway, just sign up first :p

I will let you know when I will start. I think the first thing I will do is to attend this course.

Looks like a good one!

This post has been ranked within the top 10 most undervalued posts in the first half of Jan 20. We estimate that this post is undervalued by $15.74 as compared to a scenario in which every voter had an equal say.

See the full rankings and details in The Daily Tribune: Jan 20 - Part I. You can also read about some of our methodology, data analysis and technical details in our initial post.

If you are the author and would prefer not to receive these comments, simply reply "Stop" to this comment.