The recent review of quantum algorithms is great
Can we have a similar - brutally honest - review on various hardware approaches for quantum computing? with a clear presentation of the various limitations (not always sufficiently advertised in papers?)
@LamiLudovico
among them, those based on the hardness of lattice problem are certainly the most popular and well studied. Their security relies on the assumption that the pb called LWE is hard for a q. computer. The paper (if correct) shows that this assumption may not be as sound as hoped.
@LamiLudovico
Since Shor's algo breaks RSA, cryptographers have started designing new cryptosystems hopefully resistant to quantum computers. The NIST has even launched a competition and recently decided to standardize a few of the "post-quantum" (meaning quantum-resistant) cryptosystems (1/2)
What happened to quantum tomography? A few years ago, it seemed hopelessly costly. Now it seems super easy, and even with simple procedures.
Is this thanks to better analysis or because the questions are not exactly the same as before?
After toying a little with bosonic codes, I've become less and less convinced that future quantum computing will rely on qubits rather than qudits. Until now, I always thought that qudits were just an annoying generalization that we could make, but that didn't bring much (1/3)
When Panteleev and Kalachev introduced their good quantum LDPC codes 6 months ago, the next obvious open question was to design an efficient decoder (1/5)
Random quantum circuits are unlikely to lead to a scalable violation of the extended Church-Turing thesis. Another evidence what we need quantum error correction to have fun!
Geometrically local quantum (LDPC) codes achieving optimal distance in any dimension! Finally!
Proof relying on the Freedman-Hastings code to manifold procedure and the existence of good QLDPC codes.
The 1000 qubits of IBM relegated to a footnote in Scott's description of the "top experimental QC advance of 2023".
This says a lot about the crazy progress of neutral atoms!
.
@preskill
to Alain Aspect: "Yes, we were young guys up in the amphitheater and we were waiting to see the reaction of Feynman in front of the young French guy speaking about Foundations of quantum mechanics!"
Lots of nice anecdotes in this interview
@AndreasAtETH
@Google
@qudev
@ETH_en
Although I admire your work, I'm not sure how I feel when leaders in the field brag about their citations. Even if it's not your intent, this doesn't help to create an inclusive environment in the field.
qLDPC codes are great (obviously), but if you have a lot of cat qubits, you don't need to worry too much about bit flips and can rely on classical LDPC codes instead.
Really fun project with
@Ruiz_Diego1
@CVuillot
Mazyar Mirrahimi and Jérémie Guillaud from
@Alice__Bob
Cat qubits concatenated with local LDPC codes, check out our new preprint in collaboration with
@CVuillot
and
@letonyo
! We demonstrate that we can encode 100 distance-22 logical qubits with 758 cats, more than 5 times better compared to the repetition code
More progress on
#quantum
error correction this time by
@IBM
@IBMResearch
now in
@Nature
and on the cover.
High-threshold and low-overhead fault-tolerant quantum memory
Check it out.
free-to-read pdf:
web site:
Bosonic codes are extremely attractive to reduce the overhead of quantum fault-tolerance, by something like a couple of orders of magnitude compared to more standard approaches
Very interesting paper by the
@MSFTResearch
team:
@krystasvore
While 2D architectures are in principle sufficient, they lead to scary numbers when discussing quantum error correction (1/3)
To have a fighting chance against errors building up in quantum computers, we need highly parallelizable decoding algorithms. With Gilles Zémor, we show that good quantum LDPC admit such fast decoders. (1/2)
So how do you get to enjoy all the promises of good quantum LDPC codes if you're stuck with a 2D connectivity? No need to throw away all your awesome superconducting qubits just yet! Ani Krishna, Chris Pattison and
@preskill
have some ideas!
@ghostway_chess
right, but this is still way too close for comfort. If the paper is correct, I don't think anybody will put much trust into the security of these candidates.
the issue when papers with big claim show up on the arXiv, but details are not provided because of IP issues. The arXiv is here to advance science, it's not an advertising platform.
Congrats to our finalist candidate, Dr. Steve Flammia, and to Virginia Tech and Northrop Grumman. Wishing Steve, Virginia Tech, Northrop Grumman every success with the new
#quantum
center.
"While quantum computers may still prove useful for quantum chemistry, it may be prudent to assume exponential speedups are not generically available for this problem."
So it begins: layoff in the quantum industry. Hopefully quantum hardware companies can find enough incentives for focus on improving gate fidelities, before moving too fast towards elusive applications.
We're delighted to announce our three-year quantum computing roadmap, culminating in a neutral-atom computer with 100 logical error-corrected qubits. Read the full release at:
Great to see that IBM preferred approach to (longish-term) quantum error correction involves quantum LDPC codes: " the efficiency benefits are expected to outweigh the long-range connectivity costs"
Very interesting paper by the
@MSFTResearch
team:
@krystasvore
While 2D architectures are in principle sufficient, they lead to scary numbers when discussing quantum error correction (1/3)
If you want to know more about the recent breakthoughs in the field of quantum LDPC codes (including the good codes of Panteleev and Kalachev), I'm giving a colloquium in a few hours:
Quantum error correction seems kind of obvious nowadays. But it wasn't always like this. Today I learned from Steve Brierley of
@RiverLane_io
that Sir Peter Knight almost wrote a paper back in the 90s to say that quantum computing was doomed because of the lack of QEC (1/3)
The advent of quantum cryptography 40 years ago by Wiesner, Bennett and Brassard led to a new perspective on (low-energy) quantum theory: in particular it is compatible with key distribution but not with bit commitment (1/2)
Super interesting paper on QEC for dynamical codes, and a new no-go thm showing that 2D codes with short-range interactions won't let you get low-depth implementations of non-Clifford gates.
Anne Canteaut, lauréate 2023 du prestigieux prix Joliot-Curie de la "Femme scientifique de l'année" en France pour ses travaux sur le chiffrement, qualifie de "terrible" la place accordée aux femmes dans les sciences
Relaying a statement by Elham Kashefi:
"On the day where all of us in the quantum field are thrilled to celebrate the Nobel prize for the historical achievement of the pioneers of our field, Alain Aspect, John F. Clauser and Anton Zeilinger, (1/2)
after the resolution of the quantum LDPC conjecture by Panteleev and Kalachev, it's now the turn of the (combinatorial) NLTS conjecture to fall. Exciting times in QEC and Hamiltonian complexity!
Why is everyone talking about this paper while there are similar claims all the time, and they never pan out?
Is there any hint of something different this time?
By "not any time soon", I mean "clearly not in the next 5 years", contrary to a number of folks in the AI industry.
Yes, I'm skeptical of quantum computing, particularly when it comes to its application to AI.
Applications of good quantum LDPC codes outside of quantum computing for fooling many levels of Sum-of-Squares. It will be exciting to understand the properties and consequences of these new mathematical objects!
listening to a talk by
@pasqalio
on recent progress in neutral atoms platforms for quantum computing.
The speed of improvement (compared to ions and supraconducting qubits) is rather amazing!
(Plots by Pascal Scholl)
many of my quantum colleagues also joined me to endorse this statement of solidarity with students in Iran that are sacrificing their entire career and life to redefine the history of Iran. Please join us and be their voice." (2/2)
Some thoughts about multimode bosonic codes for quantum computing. Right now, most bosonic codes are single-mode and correspond to a 2D (for qubits) subspace of an infinite-dim Fock space. (1/n)
It's great to see that many leaders of our field are completely open in admitting that they made errors in earlier papers (spoiler: this happens to everyone!). Other inspirational examples: Thomas Vidick
John Martinis' NISQ threshold conjecture:
"quantum computers and simulators can be efficiently (classically) simulated with errors, approx a few percent."
=> how far are we from proving such a statement?
Wonderful tutorial by
@RichardKueng
on the randomized measurement toolbox today at the IQFA colloquium. Richard makes it look so nice and simple that it seems extraordinary that these ideas took so long to emerge! (1/2)
Are you a fan of the surface code?
Did you ever wish for a similar simple code in three dimensions which had optimal code parameters given the locality constraints?
In new work with Nouédyn Baspin, we have just the codes for you!
The overview talk by Victor Albert
@theeczoo
is a wonderful resource on recent developments in quantum error correction!
Maybe the slides will be available soon?
the paper is out: with Aurélie Denys, we design codes with interesting groups of "easy to implement" logical gates (think transversal gates, or Gaussian unitaries for bosonic codes). Basically the only requirement is that this group is a unitary 1-design.
as much as I like to think that I'm working on a hard problem, I certainly don't believe that building a QC is the hardest thing humanity has ever tried!
"Building a quantum computer is probably the hardest thing we've tried as a species"
Our mission is to make
#quantum
computing more useful far sooner than previously imaginable.
▶️Find out more about our mission and how we aim to achieve it here: