The Classical Channel

Quantum Computing

10:41

0

Q: Practical Implementations of QECCs in IBM Q Experience

I am learning how to program the IBM Q Experience quantum computers in order to learn more about how does it work and in order to perform some experiments in it. By doing so I was wondering what are the most advanced things that have been done in such computers that actually are an advance for qu...

The Masked Rebel

12:44

What is this site

Anonymous

@TheMaskedRebel Hi. Are you facing any problem ?

The Masked Rebel

@Blue NO I just wanna know WHAT IS THIS ROOM!

Anonymous

@TheMaskedRebel This is the main chat room for Quantum Computing Stack Exchange, as you can see from the room description.

The Masked Rebel

@Blue Do you have any command bots

eg. bots who take commands

Anonymous

@TheMaskedRebel What do you mean by command bots ?

Anonymous

12:57

No

The Masked Rebel

Than how do I join your event @Blue

Anonymous

Register here: chat.stackexchange.com/rooms/info/74398/…

Anonymous

@TheMaskedRebel

James Wootton

13:14

I might not make it to the chat session tonight, but I'll try.

Anonymous

@JamesWootton No worries. We'd be happy to extend the chat session by half an hour or so, if you decide to join in a bit late. Anyway this is a regular biweekly event, so there's always a next time. :)

Quantum Computing

14:35

1

Q: Composing multiple quantum circuit in single quantum program in qiskit

I was wondering if there is a way to compose a program with multiple quantum circuit without having the register reinitialized at $0$ for each circuit. Specifically, I would like run a second quantum circuit after running the first one, as in this example: qp = QuantumProgram() qr = qp.create_...

quantum-algorithms qiskit

Semiclassical

15:47

This is neat. Apparently Ed Witten put up a preprint survey of quantum information a few days ago: arxiv.org/abs/1805.11965

Anonymous

@Semiclassical Wow, looks interesting. Thanks for sharing!

Anonymous

Didn't know Witten is involved in QIT stuff

Semiclassical

Me neither

Mithrandir24601

16:06

@Semiclassical interesting - for some reason, it's submitted under HEP theory

glS

he wasn't. That paper seems to have been seen by the community as a sort of "public announcement" of Witten's now working on QI
See e.g. https://twitter.com/preskill/status/1002208362073739264

3

Anonymous

17:35

@glS Interesting times ahead

Mithrandir24601

17:50

I need to go and get food, so I'll be 5-10 minutes late

Andrew O

18:02

Hi Everyone

Anonymous

@AndrewO Holla!

Anonymous

How're things going?

Andrew O

Good

Anonymous

Ah @JamesWootton and @agaitaarino aren't here yet, it seems

Anonymous

@AndrewO Any progress on StrangeWork's part?

Andrew O

18:04

Not sure what the topics of this chat are supposed to be but if anyone has any ideas about how to improve the SE or feature requests let me know. We can try our contacts at SE.

We've been making steady progress. We'll have an announcement soon.

Anonymous

6

Q: Biweekly chat sessions: Timings

Having discussed with @Mithrandir24601 and @agaitaarino, I would like to set up a regular chat session in The Classical Channel, which would be held every alternate week for around 1 hour. Once it is set up, I will update the link to opt-in, here. That is, you will be given the option to opt-in t...

discussion chat

Anonymous

"The agenda of the chat sessions would be (roughly):

Introducing new users to the existing community, and answering any questions they might have about the site.

Discussing recent research progress in topics related to quantum computing and quantum information.

Discussing any problematic issues faced by the members and/or possible ways to improve our site in general."

Anonymous

@AndrewO I mean we still have an issue of less expert users, but things are improving gradually

Andrew O

We proposed having a question of the week that would be featured through bounties.

Anonymous

@AndrewO Cool! But whom would the bounty come from?

Anonymous

18:07

SE would have to change its whole design for that

Anonymous

I'm not sure they'd be willing to do that

Andrew O

Yes, that was the question.

We proposed having a separate Strangeworks account with rep only used for bounties.

Anonymous

BTW anyone new around here? @user1271772 @JustinYouens Please consider introducing yourselves :)

Andrew O

Not sure of all the details of implementing such a thing but it's just something we through out there to drive participation

If anyone else has ideas we're happy to hear them.

We'd like to send the mods some Strangeworks shirts.

As a thank you

Mithrandir24601

@Blue I think there's more than what it looks like at first glance

user1271772

18:10

I'm here, but also replying to some emails.

Andrew O

Justin Youens is the Strangeworks CTO

He's a bit busy at the moment but we're both here in the office.

Anonymous

@AndrewO Oh. Didn't know that. Nice

Andrew O

@heather, @JamesWootton, @Mithrandir24601 If you can email me your address and t-shirt size I'd like to send some shirts as thanks for moderating this SE.

My email is [email protected]

Anonymous

@AndrewO You could tell the SE people to implement some Quantum circuit drawing tool on the site, that would be a helpful feature :)

Mithrandir24601

@AndrewO Oh wow, thanks very much! :)

Andrew O

18:12

Ok, I'll make a note.

Anonymous

9

Q: Do we want a package for drawing quantum circuits enabled?

In the comments to the question How to draw quantum circuits?, there's a comment saying: "Maybe we could request the devs to add a LaTeX package to draw quantum circuits" I've just discovered that this may be possible. While what (if anything) is implemented will presumably be up to the develop...

discussion mathjax formatting

Mithrandir24601

@AndrewO If we occasionally did something or other like this, SE has stash available, but we'd probably need to be much more established first and every week might be too much for that

@Blue They don't have the time

Andrew O

@Mithrandir24601 Who doesn't have the time? SE?

Mithrandir24601

So, if there's something already implemented/existing and there's a quick and easy way to add it on, they can add a few lines of code, but they can't add it from scratch

Andrew O

@Mithrandir24601 I'm not familiar with stash, can you elaborate?

Mithrandir24601

18:15

@AndrewO Well, SE might, but the CM I spoke to about it doesn't

@AndrewO I'm not sure exactly what, but things like stickers and t-shirts. Presumably nothing specifically for QC, but there's generic Stack Exchange stuff available for occasional use

e.g. Recently, there was a 'design a hat' thing on Meta SE and people with upvoted answers got a bunch of things

(recently=half a year ago though)

Anonymous

I'd like to design some T-shirts with our site's logo so that I can wear it around and spread the news :P

Andrew O

Ah ok. I thought stash was some kind of SE reputation that a site could use for bounties instead of any particular user paying for it.

Mithrandir24601

@AndrewO Ah, sadly not

Andrew O

I recall some discussion about changing the orange cube to something else.

Did that go anywhere?

Anonymous

There are lot of issues with having a question of the week

Anonymous

18:19

Like who would choose what qualifies as QOTW?

Mithrandir24601

We could do a topic challenge?

Anonymous

@AndrewO "They don't have time" :P

Andrew O

@Blue lol

Do you have a link for what the community eventually settled on?

Mithrandir24601

@Blue They're currently in the middle of overhauling the entire UI of all SE

Andrew O

@Mithrandir24601 Seems like a good time to update the logo then

Mithrandir24601

18:21

@AndrewO Exactly my though

Currently, the most upvoted logo is this:

15

A: Quantum Computing Logo

(ah, the image didn't show)

Andrew O

I liked that one

Show's I've already upvoted it

Mithrandir24601

It's got more than double the upvotes of the next highest, so (to my mind), as it's also got >10 upvotes, that's a consensus :)

Anonymous

Anonymous

I liked this one too ^

Anonymous

But maybe a bit too much clutter

Mithrandir24601

18:22

@Blue 'Only' got 7 upvotes though :P

Anonymous

In a little space

Anonymous

@Mithrandir24601 It was posted much later

Anonymous

Most people didn't notice the answer

Mithrandir24601

Does this work?:

Andrew O

The bra ket looks good too with some slight modifications

Mithrandir24601

18:24

@Blue We could make another post being like "the top 2 were ..., which one do you want as the logo?"

Anonymous

@Mithrandir24601 Yes, I was going to say that

Anonymous

Maybe top 3

Justin Youens

Greetings!

Anonymous

@JustinYouens Hi :)

Justin Youens

We are happy to do more variations of the logo, or to clean up other people's ideas with our designer.

Mithrandir24601

18:26

@Blue The third one only has 2 upvotes and (personally) I find it easier with just 2, but if people would rather have top 3...

@JustinYouens Hi!

Justin Youens

But keep in mind the main mark should be rather simple as it's shown in very small spaces. (favicon, etc) That doesn't prevent us from having multiple variations of course... like on an ad as shown above. The simple sphere was just a (very) generic alternative to the cube.

Anonymous

I actually suggest that you guys (the Strangeworks people) take up the logo suggestions in that post, and start a new meta thread, stating that you'd be willing to help with the logo design. Also perhaps ask for suggestions from the other users

Anonymous

That might make SE consider the idea seriously

Mithrandir24601

@JustinYouens I like the sphere - it would work really well for the little image used in tabs and everything

Anonymous

The mods can make it a "featured" post

James Wootton

18:28

My internet is terrible, so I am mostly not here :(

But I do lend my vote to the sphere as the new logo

Andrew O

@JamesWootton I mentioned earlier I would like to send out some t shirts to mods if you could please send me your shirt size and address to [email protected]

Anonymous

Hi @JamesWootton!

Anonymous

It's 100 kbps here :P

James Wootton

Thanks! Will do

Anonymous

Anyhow, any new progress in terms of quantum computing research? Witten's paper was interesting, which @glS shared

Andrew O

18:31

Yes. I saved that as soon as I saw it. haha

Anonymous

I am still reading the first few pages

Andrew O

Not exactly new but I found the Oak Ridge paper on cloud quantum computing of an atomic nucleus a good read.

arXiv:1801.03897v1 [quant-ph] 11 Jan 2018

Anonymous

Oops, @Semiclassical was the one who shared the paper. @glS showed us Preskill's twitter post :P

Semiclassical

ty

seems to be reviewing entropy stuff

I find Preskill's twitter to be more relevant, tbh

in the sense of it suggesting that the Witten paper is going to lead to more stuff by Witten

Andrew O

I just received @JohnWatrous 's book theory of quantum information

Anonymous

18:36

I am currently trying to make sense of Rebentrost/Lloyd's paper on quantum Hopfield networks: arxiv.org/abs/1710.03599

Anonymous

I found it interesting how they're motivating the energy function from the Ising model

Mithrandir24601

@AndrewO Nice!

Anonymous

@AndrewO Which one?

Anonymous

cambridge.org/core/books/theory-of-quantum-information/…

Anonymous

Ah, I see

Anonymous

18:38

Hadn't heard of it, before

Andrew O

@Blue Yes that one

Anonymous

But should be good :)

Andrew O

If anyone has book recommendations please let me know. I'm trying to build up the Strangeworks library

Anonymous

@AndrewO Books on quantum computation?

Andrew O

Quantum Computation / Information

I also received Wittek's book on quantum machine learning this week

It was payday so I went on a bit of a textbook shopping spree haha

Now to work through them.

Anonymous

18:41

Here's a book on information theory which I have been reading for the past few days and found it to be extremely interesting. It isn't quantum though. But lot of the algorithms (especially the neural network ones) do have quantum implementations.

Anonymous

"Information Theory, Inference and Learning Algorithms
Textbook by David J. C. MacKay"

Mithrandir24601

@AndrewO Hmm... I think I've read Quantum Error Correction before (it was a while ago) and I've also read some of Quantum Networking (Networks and Telecommunications)

Andrew O

I've got Lidar's book.

user1271772

Anyone still here?

Anonymous

@user1271772 Hey!

Mithrandir24601

18:46

@Blue I've got Elements of Information Theory

@user1271772 Yep

user1271772

I guess I missed the interesting parts. Now we're just recommending books?

Andrew O

Can anyone recommend Quantum Information Theory by Wilde?

Anonymous

@Mithrandir24601 Thanks, I'll see. McKay's book is quite unconventional in the sense that there are a lot of places where motivation is drawn from physics and other subjects like biology :)

Anonymous

@user1271772 It started with Semiclassical sharing Witten's paper

Anonymous

We are all still here though

user1271772

18:49

I saw that.
Didn't seem to be much interesting in that paper that I saw.
However it is nice to know that Witten is now one of us.

Anonymous

So you could feel free to lead the discussion

Mithrandir24601

@Blue That sounds good actually

Semiclassical

Well, so far he's only done the one mini-intro

so it remains to be seen what else he's got up his sleeve

user1271772

It would be amazing to see him contribute something major to quantum information, I'm waiting for it.

Semiclassical

the place I heard about this, to clarify, is from the latest blog post at Peter Woit's blog

user1271772

18:52

Ok if quantum computing took off and a Nobel Prize was awarded for it, who would it go to?

David Deutsch ?

Peter Shor ?

Semiclassical

and he mentioned that in the context of the upcoming PITP 2018 program at ISA : pitp.ias.edu

Anonymous

@user1271772 Feel free to refrain but it would be good to have some brief introduction from you (about yourself and what you do/study), since you have made several contributions to the site. I wanted to thank you for the contributions to the site. Keep it up!

Andrew O

I'd say Deutsch. Or maybe if quantum annealing was significantly improved, nishimori or farhi

Anonymous

Also, btw your username is quite difficult to remember, lol :P

user1271772

Thank you @Blue. My work started in quantum foundations theory, then moved to open quantum systems modeling (decoherence), then quantum algorithms for adiabatic quantum computing. My PhD was in quantum chemistry.

Mithrandir24601

18:54

@user1271772 Hmm.. Depends on context - I think Deutsch and Josza for showing that it could be potentially faster than a classical computer, Shor for potentially breaking encryption

Whoever came up with QKD deserves a mention as well

user1271772

Bennett and Brassard?

Andrew O

prizes tend to go to people that open new fields of research so I'm not sure if Shor would be eligible.

user1271772

Absolutely you're right.

Mithrandir24601

@user1271772 ... Yes. Obviously...

Anonymous

@user1271772 Wow, that's quite a lot of experience in "quantum" stuff. Good to hear. I'm supposing you're a post-doc/professor now? :)

user1271772

18:56

That's right @AndrewO. He certainly popularized quantum computing and got a lot of new people interested in it, but he's not a pioneer at the same level as Deutsch, Bennett, and Brassard.

Andrew O

There was a bit in Scott Aaronson's book about how Shor built on the work of others.

user1271772

@Blue: I've been working in quantum information since 2008. In that year I also took the course taught by Michele Mosca and John Watrous. I was working with Raymond Laflamme at the time.

Andrew O

@user1271772 How did you like his course? I just received his new textbook

user1271772

Yes Shor's algorithm makes use of things like "eigenvalue kick-back" which came from Deutsch-Josza algorithm and the Cleve-Eckert-Machievello-Mosca version of it

@AndrewO: Do you mean "Introduction to Quantum Computing" by Kaye, Laflamme, Mosca?

Andrew O

@user1271772 No, it's "The Theory of Quantum Information"

It was released in April

user1271772

18:59

Oh Watrous,
I thought you meant Laflamme or Mosca.

I haven't got John's book,
These books tend to be quite elementary, and the really interesting stuff to me nowadays is in research papers.

But that book is based on John's lecture notes which he's had online since the early 2000s. It is excellent for beginners. He is a great teacher.

Anonymous

@AndrewO Your PhD was in which subject? I forget

user1271772

You can see his 2006 and 2011 lecture notes here: https://cs.uwaterloo.ca/~watrous/LectureNotes.html

And a PDF of his book here:
https://cs.uwaterloo.ca/~watrous/TQI/

I am a big fan of people who make their books available for free.

Andrew O

@Blue Computational Physics

Mostly algorithms for spin-glasses since that's what the D-Wave essentially solves .

Anonymous

@AndrewO Ah! Your LinkedIn profile says you have experience with Spin Glasses. I was stuck with that thing along with the mean field energy theory stuff for long

Anonymous

Maybe if you have some time I could ask you a few questions which I was confused about :P

Anonymous

19:07

You happen to know the variational methods, right?

Anonymous

Or I could ask them on the main site where you could answer when you're free

Anonymous

I am currently working on the Hopfield networks and there the minimization of energy function comes up a lot

Andrew O

Sure

Anonymous

Which is motivated from spin glass systems

Andrew O

I'm not familiar with Hopfield models but I can read up a bit and try to help you out.

Anonymous

19:11

I can give a brief summary:

Anonymous

Hopfield network

A Hopfield network is a form of recurrent artificial neural network popularized by John Hopfield in 1982, but described earlier by Little in 1974. Hopfield nets serve as content-addressable ("associative") memory systems with binary threshold nodes. They are guaranteed to converge to a local minimum, but will sometimes converge to a false pattern (wrong local minimum) rather than the stored pattern (expected local minimum). Hopfield networks also provide a model for understanding human memory. == Structure == The units in Hopfield nets are binary threshold units, i.e. the units only take on two...

Anonymous

So basically a Hopfield network consists of a single layer of neurons

Anonymous

And their activations can be 1 or -1

Anonymous

In some notations they take it to be 0 or 1

Anonymous

19:13

You know the "neurons which fire together, wire together" thingy, yes?

Anonymous

Here it is similar, you increase the weights connecting the neurons which have same activation

user1271772

Oh @AndrewO: You worked with Helmut?

Andrew O

Sounds like an Ising model but I'm don't know the fire together wire together thing.

@user1271772 Yes I did

user1271772

Helmut got me in touch with Zheng Zhu a couple years ago, but Zheng would not share his code.

Anonymous

@AndrewO Yep, it is very similar to the Ising model! :) The weights in a neural network is similar to what you call "coupling" in spin glasses and Ising models. Perhaps it will be better if I ask this an elaborate question on the main site. You could read it meanwhile (if you get time)

Andrew O

19:15

@user1271772 Correct. Helmut does not allow us to share code.

user1271772

It made me wonder whether anything out of Helmut's lab is true, or just faked.

Really?

Andrew O

@user1271772 That's a very bold claim. Helmut takes great care in maintaining data for provenance.

user1271772

So the only way to reproduce the data in the papers is to get the authors to reproduce it? So if someone dies it can never be reproduced?

Andrew O

@user1271772 Simulation parameters are in the publications and data is required to be maintained by the NSF.

Anonymous

Yeah, authors who don't share code sort of annoy me :P

user1271772

19:18

But the code is not required to be maintained by the NSF?

Andrew O

All of our code is backed up as well.

Helmut has shared code in the past with collaborators. But it is rare for someone outside to receive our code.

user1271772

Don't get me wrong, I have been very impressed by Helmut's results that he's presented at various conferences, and in papers.

Andrew O

May I ask who you are?

user1271772

I am going to change my SE name to my real name at some point, but haven't decided how to do it and when. I've been using this account for several years.

Almost a decade now.

Andrew O

Proposing Helmut's data is faked is not something I take lightly, especially under anonymity. If you'd like to continue this conversation feel free to email me [email protected]

user1271772

19:24

@AndrewO: There's an easy solution: show us the code.

Mithrandir24601

@user1271772 As @AndrewO just said, it's maybe not the best idea to go around saying this kind of thing here without serious evidence. Not liking that someone doesn't share this or that is fine, but implying that they're possibly faking data is quite serious

user1271772

I think my statement was interpreted much more strongly than I intended, which isn't uncommon in a text-only based chat like this, so I am not complaining about that. But in all honesty, I respect Helmut's work and have collaborated with him on projects.

Andrew O

@user1271772 That's something you'll have to take up with Helmut as he's the PI. I'm more than happy to explain my implementation of algorithms.

user1271772

It's up to him.
Actually we had also asked for compiled binaries in case they didn't want to show the source, but they wouldn't give compiled binaries either.

Andrew O

Sounds familiar. Did you work for Mattias?

user1271772

19:32

No I did not.

Anonymous

Hi @agaitaarino!

agaitaarino

Arriving late, but still catching some people I see!

Anonymous

Most are still hanging around :)

Anonymous

@agaitaarino Which area of quantum computing do you work on? I see you ask a lot of questions regarding solid-state

agaitaarino

I'm actually a chemist... I try to understand magnetic molecules in a QC context.

user1271772

19:38

What are some of the uses of molecular magnets?

agaitaarino

I uploaded my project here: uv.es/gaita/pdf/ERC_CoG_2014_B2_20-05.pdf

Anonymous

Ah, cool stuff!

agaitaarino

The ideas are getting a little old, but it's still the basics of what we're doing.

Working with magnetic molecules to try to understand quantum decoherence in the solid state better.

And trying to use them also to connect several qubits in a controlled way.

All very fundamental stuff, actually, nothing I see as close to an application

Anonymous

@agaitaarino I would actually be interested to know about that. I still don't properly understand how exactly measurements in solid state are conducted or rather what even constitutes "measurement". Pretty much a noob in experimental aspects

Andrew O

@agaitaarino I love Fig. 1

agaitaarino

19:42

We do a lot of pulsed EPR (more on that in a moment if needed), but in fact it is preliminary, since it is bulk/ensemble.

Anonymous

@AndrewO lol...it's nice

Andrew O

I like cats. a lot

agaitaarino

Kind of like the old days pulsed NMR-QC: EPR is NMR with electron spins

Anonymous

@AndrewO "epitome of indifference" ;)

agaitaarino

I'm allergic: the one for Fig1 was only semi-domestic. It took a good morning of chasing him around to get a good picture xD

But, on measurements, there are plans to couple single molecules to superconducting circuits

Anonymous

19:44

wtf...that's a real cat picture you took? LOL

Anonymous

@agaitaarino How exactly?

agaitaarino

pubs.rsc.org/en/content/articlelanding/2016/dt/…

This is a 2-year ago summary of the plans (I was not in the paper, but am currently collaborating with Luis and Majer in a QuantERA project based on something similar).

Anonymous

Looks interesting, but as usual I don't think I have the necessary background to understand it

Anonymous

But I'll try, thanks for sharing

agaitaarino

Yay, real cat! The plan all along was to be able to tell people: "To get the grant, you have to use a picture of a cat as Figure 1; worked for me!", to help remember people that anecdotal advice is not necessarily useful :-P

user1271772

19:48

I have some colleagues in quantum chemistry that always say at the beginning of their presentations that the motivation for all their work, is to one day be able to study molecular magnetics using ab initio methods. Do you know what the motivation for this would be? Any useful applications of molecular magnets?

agaitaarino

Feel free to ask question on the background, I'll be delighted to understand to the best of my ability :-D

Andrew O

Will these chats be every week or every other week?

agaitaarino

I thought for a long time that molecular magnets were dead, but last year a colleague/competitor jumped to hysteresis at 60K, which is surprisingly hot

Anonymous

Actually tbh, I don't understand what experimentally measurement means, in different quantum computer architectures like superconducting QCs, photonic QCs, etc. I just know some of the mathematical details. So I was looking for some simple introduction to "architecture of quantum computers"...preferably accessible at the undergrad level

agaitaarino

(every other week, right?)

Anonymous

19:50

Biweekly, yes

Anonymous

Once in every two weeks

Andrew O

Ok thanks.

I have a meeting shortly

So I'll see you all in two weeks. I'll work on the logo and getting a circuit generator going.

Anonymous

To be specific, it will be held on every alternate Tuesday

Andrew O

Have a good week

Anonymous

See you, it was nice having you here :)

agaitaarino

19:51

If there is a photon that can be transmitted only if the molecule is in a certain spin state, then that can be a way to measure the spin state (qubit), does that sound right?

Anonymous

@agaitaarino Yep

user1271772

yes

What is the significance of having hysteresis at 60K ?

agaitaarino

Or, with quantum dots: if current can pass only through a quantum dot depending for a certain charge state of the quantum dots we use as qubit, that0s another measurement.

Mithrandir24601

@AndrewO Bye and thanks!

agaitaarino

We have been trying to rise this temperature (from around 2-4K) since 1993 or so.

user1271772

19:53

@AndrewO: See you in two weeks. Sorry for all that, I didn't realize what I said would be taken so strongly.

agaitaarino

There was a inredible jump to like 14K a few years ago.

And suddenly, 60K, almost unprecedented in the >20K range.

So, not useful, but intriguing: we started to think it could not be done.

Anonymous

@agaitaarino Ah, I see

Anonymous

Interesting

agaitaarino

This is hysterisis at the single molecule level (or from single molecular origin) by the way

user1271772

But what would the molecular magnets be used for?

agaitaarino

19:55

So, not measured at a single spin, but not arising from magnetic domains

user1271772

Are molecular magnets just cool, or they are useful for something?

agaitaarino

If 60K -> 300-400K, as classical bits

but I'd say just cool

user1271772

No potential practical use of molecular magnets?

agaitaarino

They are also an origin for molecular spin qubits (which on their own: mostly just cool)

Some talk about applications in cryogenics

user1271772

molecular spin qubits like in NMR experiments?

agaitaarino

19:56

To cool stuff below 1K.

user1271772

What is the application in cryogenics?

Is it better than laser cooling or evaporative cooling?

agaitaarino

Following the Dalton paper above ( pubs.rsc.org/en/content/articlelanding/2016/dt/… ), potentially much nicer than NMR

Not specialist on magnetic refrigeration, but it's this: en.wikipedia.org/wiki/Magnetic_refrigeration

Anonymous

It was a great session. Thank you all for coming. It's quite late here, so I got to leave now. See you people again in two weeks!

user1271772

See you in 2 weeks.

Mithrandir24601

@Blue Bye! I think it went rather well :)

agaitaarino

19:58

Yes, sorry I arrived so late but I think I'll leave soon. I'lltry to catch up with what other said before I arrived though!

Anonymous

@agaitaarino I'll later ask you more questions about the measurement part, but thanks for explaining. :)

user1271772

@agaitaarino: interesting paper. I like the Figure showing coherence times as a function of year published. Very promising. However how many qubits have they been able to do QC with?

agaitaarino

From the little I've seen, I also think it was really good, thanks all!

Mithrandir24601

@Blue You could certainly ask lots of questions on the main site about measurements

@user1271772 Number of qubits isn't really a great number (by itself) to characterise, well, anything really

agaitaarino

not real QC with this approach, actually... the best thing was 2-qubit Grover with a related approach, journals.aps.org/prl/abstract/10.1103/PhysRevLett.119.187702

user1271772

20:01

@Mithrandir24601: What would you say is the best number to characterize a quantum computer?

Mithrandir24601

@user1271772 There is no single 'best number'. It depends on so many things. Connectivity, fidelity, number of qubits, gate operation time, how long ago did [group] start looking at [method] for implementing a QC, scalability, ...

user1271772

Yes, at the IBM conference in December they introduced a metric called "Quantum volume" which combines many of these things. It is like a "capacity to do quantum computation". Bigger "quantum volume" = better.

However I think there's nothing wrong with asking "how many qubits" as a starter question.

It is just what I was interested in, not necessarily the best measure of anything.
If he said "1000 qubits" then I'd start asking about connectivity, if he said "complete graph" I'd ask about types of possible gates (only sigma_z ? or sigma_z and sigma_x can be used together?). If he said "universal gate set possible" I'd ask about decoherence times.

But I didn't need to ask any of that because the answer was 2 qubits!

Mithrandir24601

@user1271772 I don't know... I prefer 'How many qubits can you use to do an actual quantum algorithm with a high success probability', as the likes of Google and IBM are getting very good at announcing 50 or 72 qubit chips that, by the sounds of things, aren't of any actual use

user1271772

Yes I agree about Google and IBM. At least IBM has the 5-qubit and 16-qubit machines available for public use, but Google on the other hand ......

Mithrandir24601

@user1271772 Yeah, true

@user1271772 Google does have (at least) a working 9 qubit chip

(as of a few years ago now)

user1271772

20:10

Not one that we can do experiments on, as far as I know (?)

Well Google published in July 2016 the paper where they used a "9-qubit" quantum chip to calculate the ground state energy of the hydrogen molecule, but that quantum computation only required 2 qubits, so I have never actually seen any 9-qubit "actual" algorithm done with high success probability --- you can correct me if I'm wrong though, I haven't read all their papers.

Mithrandir24601

@user1271772 Ahh, that was the one I was thinking of, yeah

Although I did stumble across this, which is something

user1271772

Just looked back at the paper, they also have an ancilla, so 3 physical qubits, 2 logical qubits. No evidence that they ever had a 9 qubit quantum computer

Mithrandir24601

Although it gets complicated for photonics - if you asked me how many qubits we've got, my answer would be "I'm not really sure" :P

user1271772

Very cool paper! Thanks for sending it!
So they do have a 9-qubit quantum computer which can actually do something (preserve states). No "actual" algorithm, but perhaps a step towards a 2D surface code.

As for photonics, I'm not sure either, how many qubits have been used for "actual algorithms with high success probability". How about for boson sampling? How far are we for that? I've seen a paper where they used a classical computer to simulate boson sampling up to about 50 bosons I think.

Mithrandir24601

@user1271772 Yeah... We don't really work in 'qubits', that's the problem

@user1271772 We can do more than 50 now - 60s is possible to simulate classically

As for numbers of photons, um, like 10?

@user1271772 Oh and it gets worse - as soon as you add (currently) realistic noise, the circuit either becomes small enough to easily simulate classically or large enough that the noise dominates and makes the light classical, which makes it efficiently simulatable...

user1271772

20:36

Yes I heard about these things. Do you work with Ian Wamsley?

Mithrandir24601

@user1271772 I've met him before, but no - I'm supervised by Anthony Laing and Yogesh Joglekar, although I've done projects with other people in Bristol (Peter Turner as well as QComms group)

Ah, this is what I was thinking of - "No imminent quantum supremacy by boson sampling"

Mithrandir24601

20:56

@user1271772 So, the most recent paper published by my supervisors/PhD-friends was 4 photons over 6 modes. How many qubits this is equivalent to, I'm not sure, but if you have number resolving detectors, somewhere between 2.5 and 9.5?

Don't take my word for it though (or quote me on this!), as mapping between number of modes in the system and how many qubits this is equivalent to is not entirely straightforward and I'm not sure if applying the unitary on the 6 modes allows for implementing a unitary on all possible states, or just a subset of those :/

(with a single input photon, I believe it's equivalent to $\log_2 6 \approx 2.58$ qubits)

user1271772

Using the formula for the best classical algorithm for exact boson sampling: O(m2^n + mn^2), for n photons in m output modes, I'd say 4 photons is more than 4 qubits, because n qubits would be O(2^n) cost, but with 6 modes we have O(6*2^n + 6*n^2) which is larger than O(2^n), but I don't know the constant under the big O.

heather

21:24

hello.

sorry i missed the chat session =) seems to have gone well.

Quantum Computing

21:43

0

Q: Is it possible to "calculate" the absolute value of a permanent using Boson Sampling?

In boson sampling, if we start with 1 photon in each of the first $M$ modes of an interferometer, the probability of detecting 1 photon in each output mode is: $|\textrm{Perm}(A)|^2$, where the columns and rows of $A$ are the first $M$ columns of the interferometer's unitary matrix $U$, and all o...

quantum-information classical-computing optical-quantum-computing boson-sampling photonics

Transcript for

The Classical Channel