Lightnews — Scholar-powered news

Iftach Yakar

@quantumyakar.bsky.social

260 followers 270 following 270 posts

Quantum computing + information | Computer science | postgraduate @ HUJI

Posts Media Videos Starter Packs

Pinned

Iftach Yakar @quantumyakar.bsky.social · Jan 9

Sometimes I just add “eigen” in front of a word and hope for the best.

1 2 14

Iftach Yakar @quantumyakar.bsky.social · 16h

Almost forgot to summarize the paper in a meme (hi @dulwichquantum.bsky.social )

9/8

Iftach Yakar @quantumyakar.bsky.social · 16h

🔗 Interactive results: deduckproject.github.io/global-epp-r...
📄 Paper: arxiv.org/abs/2510.06737

Big thanks to my advisor, Michael Ben-Or, for being patient and giving thorough explanations on everything.

8/8

Advantages of Global Entanglement-Distillation Policies in Quantum Repeater Chains

Quantum repeaters are essential for achieving long-distance quantum communication due to photon loss, which grows exponentially with the channel distance. Current quantum repeater generations use enta...

arxiv.org

Iftach Yakar @quantumyakar.bsky.social · 16h

🛰️ Global policies can crucially extend achievable QKD distances.
Local policies may not be optimal in real repeater networks.
Since these global policies are deterministic, they have no real-time overhead.
Next steps: adaptive policies or other EPPs.
7/8

Iftach Yakar @quantumyakar.bsky.social · 16h

For large chains (>500 repeaters), global policies boost SKR by two orders of magnitude.
In some regimes, they make secret communication possible when local ones yield none.
6/8

Taken from Fig. 4 of our paper. These plots show how advantagous the global strategy is over the local ones. The further the curve drop - the larger the advantage for the global policy.

Different colours represent different local polices; different markers represent different amounts multiplexing - parallel links being created.

Iftach Yakar @quantumyakar.bsky.social · 16h

We take a *global* view: optimize distillation across the entire chain to maximize SKR.
By randomly sampling global policies, we find "near-optimal" policies that consistently outperform the local ones, for given parameters.
5/8

Iftach Yakar @quantumyakar.bsky.social · 16h

Recent protocols (e.g. Mantri et al. 2024, arXiv:2409.06152) use *local* policies - each repeater decides whether to distill based on local information (like estimated fidelity or Secret Key Rate - SKR)
4/8

Iftach Yakar @quantumyakar.bsky.social · 16h

All this introduces noise, which can be mitigated through entanglement distillation (EPP): turning many noisy EPR pairs into fewer, high-fidelity ones.
The question is: when to distill?
Too early - we waste good EPR pairs; Too late - noise kills us.
3/8

A schematic depicting the process of entanglement distillation

Iftach Yakar @quantumyakar.bsky.social · 16h

Quantum communication over large distances is hard due to photon loss, which grows exponentially with distance. Quantum repeaters solve this by dividing the channel into short segments, entangling neighbors, and recursively performing entanglement-swapping.
2/8

Figure from Mantri et al. 2024 http://arxiv.org/abs/2409.06152

Figure 3: Schematic of a nested swapping protocol. We start with N elementary segments each of length l0. The red boxes represent a swapping operation. Based on the decision variable Di, distillation may be performed before a swapping operation. Each swapping operation doubles the length of the link. This process is repeated until an end-to-end link is established.

Iftach Yakar @quantumyakar.bsky.social · 16h

Excited to share my first paper (w/ Michael Ben-Or)!
Quantum repeater protocols must decide when to perform entanglement-distillation - but should that decision be local or global?
We find that global policies can make or break secure communication. 🔑
1/8
arxiv.org/abs/2510.06737

AI slop featuring two quantum repeaters sharing entanglement. Just to get your attention... 😬

1 1

Iftach Yakar @quantumyakar.bsky.social · 4d

lol

Iftach Yakar @quantumyakar.bsky.social · 7d

“I know it sounds woo” 🤣

2 5

Iftach Yakar @quantumyakar.bsky.social · 17d

Just published a Stim implementation of the [[16,4,4]] Tesseract code - the one used in Microsoft+Quantinnum's QECC experiment last year.
Simulation can be used to quickly benchmark the code under various parameters.

Contributions are welcome.
github.com/DeDuckProjec...

A plot showing the acceptance rate of the code under different noise parameters. X axis - number of rounds. Y axis - Acceptance rate

A plot showing a comparison between the logical fidelity of the code under different noise parameters, with of without applying the error correction. X axis - number of rounds. Y axis - average fidelity

1 6

Iftach Yakar @quantumyakar.bsky.social · 17d

Error correction codes for ages 8+
@quantuminthings.bsky.social

1 5

Iftach Yakar @quantumyakar.bsky.social · Sep 7

Hate these.. probably better off spraying, though, at this stage of infestation

Iftach Yakar @quantumyakar.bsky.social · Sep 7

Shot from my iPhone (and a very basic telescope)

Iftach Yakar @quantumyakar.bsky.social · Sep 7

Just defended my thesis 😱

Iftach Yakar @quantumyakar.bsky.social · Sep 3

Makes sense.
Now you just gotta make sure your IDE highlights that keyword and you’re golden!

Iftach Yakar @quantumyakar.bsky.social · Sep 2

That’s fair

Iftach Yakar @quantumyakar.bsky.social · Sep 2

Is there anything more suspenseful than running a simulation when you don’t even know if the code you wrote is correct?

1 6

Iftach Yakar @quantumyakar.bsky.social · Sep 2

I didn’t we could keep such a macro system coherent!

Iftach Yakar @quantumyakar.bsky.social · Aug 30

Maybe the quantum advantage was these friends all along

Iftach Yakar @quantumyakar.bsky.social · Aug 30

Nice one :)

Iftach Yakar @quantumyakar.bsky.social · Aug 11

Do you think Miles would’ve felt the same regarding quantum information? ;)

BTW, regarding the “cool” - he has an album titled “birth of the cool”

Iftach Yakar @quantumyakar.bsky.social · Aug 11

Overheard people at the office talking about Poisson variables.
Turns out they were talking about croissants.
🤦🏻‍♂️

Iftach Yakar @quantumyakar.bsky.social · Jul 30

If I had to guess, I'd say that "high" favors high quality but can drop when your network connection is insufficient, while always-high insists on high quality anyway.