Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

rrtucci · November 23rd, 2018, 3:20 pm

Call for collaborators. We are a very early startup. (Meaning no investors yet. Virgins

We have developed free, open source (under BSD and MIT license) software that combines TensorFlow/Edward with quantum computing. This software is still very young. Would be a good community project.
https://github.com/artiste-qb-net/Quantum_Edward
http://www.ar-tiste.com/quantum_ed_talk ... rdTalk.pdf

We have also deployed on AWS (Amazon Clouder), a docker image that combines a huge trove of open source software for doing quantum and classical AI
www.bayesforge.com
http://www.ar-tiste.com/BayesForgeTalk.pdf

We have also written our own quantum simulator, much better than the ones by IBM, Rigetti and Google
https://github.com/artiste-qb-net/qubiter

We have also written an application that combines seamlessly classical Bayesian nets (a la Judea Pearl) and quantum computing
https://github.com/artiste-qb-net/quantum-fog

We have also written some fintech stuff for Fujitsu quantum computer, and a library for calculating quantum entanglement called 'Q Entanglement Lab'. These last two projects are still very early and not released publicly yet. But with your help, the sky is the limit

Cuchulainn · November 23rd, 2018, 4:04 pm

Do you have venture capital?

rrtucci · November 23rd, 2018, 5:32 pm

No, we don't yet. Thanks for the tip, Chuchulainn. The Dutch are very keen on quantum computing. For example,
https://qutech.nl/

ppauper · November 23rd, 2018, 6:13 pm

Call for collaborators. We are a very early startup. (Meaning no investors yet. Virgins

what exactly are you looking for?
people to code with you? or people to give you money?
the fintech stuff you could presumably monetize, but a lot of the other stuff seems to be free open source and/or very niche

Cuchulainn · November 23rd, 2018, 6:33 pm

No, we don't yet. Thanks for the tip, Chuchulainn. The Dutch are very keen on quantum computing. For example,
https://qutech.nl/

There's some research going on but no products shipping AFAIK.

As ppauper says, it feels a bit niche. What is the product and who are the leads? What your post does not seem to address is the benefits.

Speaking of: 25 years we hired a part-time young student fellow to teach C++. He founded a e-commerce company in the late 90s. Just heard last week that he is worth 800 million Euro. He had a vision of a product.

// NN is still not mature?

rrtucci · November 23rd, 2018, 7:11 pm

All of the above, coders and funding. Obviously, if somebody contributed a lot to the open source coding, we would try to reward him/her with equity and a salary commensurate with the contribution.

"the fintech stuff you could presumably monetize, but a lot of the other stuff seems to be free open source and/or very niche"

Is Open source still a Niche?
1. Red Hat bought by IBM for $34 billion
2. GitHub bought by Microsoft for $7.5 billion
3. There is no doubt that Open Source rules all the Clouds, The One Ring to rule them all. It also rules Machine Learning (TensorFlow, PyTorch...), Big Data (R, Pandas) and HPC

If you are saying that quantum computing is a niche, I agree. But that is why Bayesforge combines both classical and quantum AI. We will do consulting in both, and in hybrids. This strategy is called "Quantum Ready" in the biz. We could also charge a small subscription fee for Bayesforge if it includes extra goodies and premium services.

The beauty of software, is that it is much cheaper to develop than hardware. Due to the high cost of developing hardware, hardware manufacturers usually have to bet on just one platform, like a Josephson junction gate model, or an ion trap or whatever. Bayesforge bets on everything. We include software for IBM and Rigetti gate model platforms as well as for Dwave's annealer.

Bayesforge is a direct competitor to Google's Colab and a similar service that Rigetti has promised, except we believe we have significant advantages over them

Cuchulainn · November 23rd, 2018, 8:48 pm

The beauty of software, is that it is much cheaper to develop than hardware.

In the 70s yes, but not anymore. Who cares what h/w costs? You just buy a box. Software is ever so expensive, E.g. a productive 80K developer writes 50 lines of real code per week. That's expensive if you are paying the salary!

You seem to be a real tech guy. What is the product (consultancy won't be lucrative) and what happens is one of the big boys will make it as well.

From my limited experience, NN is a niche, many more people talking about it than doing it , when even the tax drivers know.

rrtucci · November 23rd, 2018, 9:44 pm

Okay, let's get quantitative, Cuchulainn, :)

Here is a Google Trends 5 YEAR search for 3 of the products that Bayesforge combines: quantum computing, Tensorflow, Microsoft Azure cloud.

Notice how China stands out very strongly. Remember Synergesis,

https://trends.google.com/trends/explor ... %2F04y7lrx

Cuchulainn · November 23rd, 2018, 9:48 pm

High risk. You are competing with the big boys.
But good luck (in bocca al lupo

)

rrtucci · November 23rd, 2018, 9:52 pm

I used Microsoft Azure, because I didn't want the search to get confused by a non cloud product. But any of the popular clouds would work. China obviously prefers its owns cloud providers (Alibaba, Tencent, etc.) to Azure

rrtucci · November 23rd, 2018, 9:52 pm

Yes, it's high risk, high reward

rrtucci · November 23rd, 2018, 9:56 pm

Open source was invented to bring the big boys to their knees, and it worked quite well for Linus Torvalds (Linux owns the clouds, and is the basis of Android), not that i consider myself his equal, but one can learn from the greats

rrtucci · November 23rd, 2018, 10:06 pm

Cuchulainn, if you want to talk about lupi, remember the alpha wolf. All it takes is an alpha-wolf like Linus Torvalds, and the pack will follow
https://www.livingwithwolves.org/portfo ... lpha-male/

Cuchulainn · November 23rd, 2018, 10:07 pm

Still, a net worth of 150 mln what LT has is not a lot considering.

katastrofa · November 23rd, 2018, 11:42 pm

What's the motivation for making the conditional probabilities in the nodes of Bayesian networks complex? What's the idea behind it and what advantages does it give you? (One could put there quaternions or any vectors, but what for?)

Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud

Re: Quantum Neural Networks, Tensorflow, Quantum Computer Programming on Amazon Cloud