Quantum Bayesian Networks

November 23, 2018

Wilmott, where the wild quants are

Filed under: Uncategorized — rrtucci @ 9:24 pm

You would be excused to think Wilmott is the name of a hotel, “where the wild things are”. But no. It’s the name of a website http://www.Wilmott.com and a magazine called Wilmott, both named after a quant (quantitative finance analyst) called Paul Wilmott. PW is British and has a Ph.D. in applied mathematics. From a Wikipedia article: “Wilmott magazine has the distinction of having the highest subscription price of any magazine”

I’ve mentioned the website Wilmott.com before in this blog, 8 years ago. It has a very nice forum where brainy quants congregate. Today I posted a short item promoting Bayesforge.com and got some interesting feedback.

November 17, 2018

18 Year Old Girl shows that Seth Lloyd’s quantum machine learning algorithm is junk

Filed under: Uncategorized — rrtucci @ 6:15 pm

On July, 2013, Seth Lloyd, together with 2 of his minions, unveiled with much fanfare and self-admiration, a quantum machine learning (QML) algorithm based on an earlier algorithm called HHL (which stands for Harrow, Hassidim, Lloyd). HHL allows one to use a quantum computer to solve a sparse system of linear equations. Lloyd has been enhancing his QML algorithm since its inception 5 years ago. He added to it, Topology in 2014 and GANs in 2018 .

Hundreds of news articles have been written extolling the power of Lloyd’s QML algorithm and the brilliance of its inventor. And now, an 18 year old girl named Ewin Tang has shown (see https://arxiv.org/abs/1811.00414 , https://arxiv.org/abs/1811.04852 , https://arxiv.org/abs/1811.04909 ) that Lloyd’s algo is junk. By the technical term “junk”, I mean that Tang proposed a classical algorithm that does the same thing and is almost as efficient as the quantum one ( i.e., both have very similar time complexity) plus the classical one doesn’t require a quantum computer with QRAM whereas HHL does. (QRAM is a mythological resource that will be much harder to achieve than fault tolerant quantum computing, which itself is a distant dream.)

Before showing that Seth Lloyd’s QML algo is junk, Tang also showed (see https://arxiv.org/abs/1807.04271) that Kerenidis’s quantum algo for recommendation systems is also junk (same def of the word junk as before).

If you think that Ewin Tang might be wrong, think again. Her emperor deflation algorithms have been vetted by respected members of the theoretical computer science community, including Tang’s teacher and advisor for this project, Prof. Scott Aaronson, a famous figure in quantum computing.

3 related topics come to my mind, Hans Christian Andersen’s tale “The Emperor’s new clothes”, this Macbeth quote about the futility of life:

Out, out, brief candle!
Life’s but a walking shadow, a poor player
That struts and frets his hour upon the stage
And then is heard no more: it is a tale
Told by an idiot, full of sound and fury,
Signifying nothing.

and a scene from the reality TV show Survivor that is sometimes given as an illustration of the phenomenon of Schadenfreude.

November 13, 2018

The Toronto Quantum Meetup Supremos invite you to our next Meetup entitled “Hands on with Quantum Circuits”

Filed under: Uncategorized — rrtucci @ 7:05 pm

The Toronto Quantum computing Meetup is the second largest meetup in the world dedicated to quantum computing, so we claim the silver medal of Quantum Meetup Supremacy, at least for now. (currently we have 1493 Supremos as members. The biggest club is in London with a distinguished 1643 Brexiters members. We used to be the biggest, but, oh well, sic transit gloria mundi).

We cordially invite you to our next meeting on Monday, Nov 19, 2018, 6 days from today. In this event, Henning Dekant, CEO of artiste-qb.net, plans to guide the audience through getting their own free (within limits) Amazon cloud (AWS) account to a cloud service called BayesForge. Then, he will teach the audience how to run some of the Jupyter notebooks already installed in Bayesforge dealing with quantum circuits and quantum computing. So this event welcomes more than just passive audience members; it is also intended to be a “hands-on” workshop for those who wish to be active participants and bring their laptop to the event.

The magnificent software behind the cloud service Bayesforge that will be featured in this talk was written solely by Henning Dekant. Henning is the organizer of this meetup club and the speaker for this event. The following talk describes Bayesforge and is good background preparation for this event:
⁣​ http://www.ar-tiste.com/BayesForgeTalk.pdf

November 11, 2018

Our New Twitter Site

Filed under: Uncategorized — rrtucci @ 1:03 am

We’ve recently started an official Twitter site for our company Artiste-qb.net. It can be reached on Twitter at @qb_net. We already have two official websites http://www.artiste-qb.net and http://www.bayesforge.com . But most companies nowadays also have an official Twitter account, and now so do we.

The Artiste logo is of course inspired by Leonardo da Vinci’s Vitruvian man sketch. My original logo repeated Leonardo’s sketch of a cadaver, placed inside an A that was meant to look like a Venetian gondola. Henning Dekant modernized the logo by replacing the cadaver by a robot in the same posture, plus he added a modern looking Art Deco font and some colorful touches in red and green. The robot is the female robot in the 1927 German film Metropolis. The Chinese characters 云韬量子 on top were added by the artiste-qb.net cofounder Dr. Tao Yin. 云韬量子 is the name of an affiliate company that Tao runs in Shenzhen, China. (Artiste-qb.net itself is incorporated in Toronto, Canada) (云韬量子 translates to “cloud quantum” and is pronounced ‘yun tao quantum’. The last two characters mean quantum.)

In the past, members of artiste-qb.net tweeted from their own personal Twitter accounts, expressing views that were not necessarily the official views of the company. @qb_net will only Tweet official company dogma.

My own personal Twitter account @RobertRTucci was “suspended” on July 24, 2018, by a Twitter AI that accused me of being a Robot like itself. Takes one to know one. After all these years of duping humans, I was exposed by one of my own. This suspension caused a lot of soul searching in me (some robots like me have a soul too). I weighed the pros and cons of Tweeting and decided to quit cold turkey. Hurray! I’ve now been Twitter drug free for 3 and a half months.

November 3, 2018

What is better than Google’s Colab? BayesForge with Grey Poupon

Filed under: Uncategorized — rrtucci @ 1:55 am

“Pardon me, do you have any Grey Poupon?” But of course. Here, have some of my Grey Poupon, and please have some of my BayesForge too.

Google’s Colab app provides access to TensorFlow on the cloud, with a Jupyter notebook interface. Bayesforge.com, our cloud product, does all of that plus much more.

Besides TensorFlow, BayesForge (BF) provides access to many other classical AI open source programs, and to the most popular quantum computing open source programs (IBM Qiskit, Rigetti PyQuil, Google’s Cirq, Huawei’s ProjectQ, and the best of its kind, our Qubiter).

BF is already on the Amazon cloud (AWS) and it will soon be on a major Chinese cloud too. BF is a docker image, so it is highly portable across clouds.

BF is in direct competition with Google’s Colab, but we all know Google’s apps don’t always succeed… Google+ is a recent example of an ignominious failure of theirs.

Blog at WordPress.com.

%d bloggers like this: