Google Quantum AI Update 2022

The Quantum AI team's mission is to build a useful quantum computer and to discover novel applications that could one day help solve challenging, real-world problems for humanity that would otherwise be impossible.

In our latest Quantum AI update, hear Hartmut Neven's thoughts on potential quantum applications and Erik Lucero's update on building an error-corrected quantum computer.

Speaker: Harmut Neven and Erik Lucero

#QuantumAITalks

fullname: Hartmut Neven, Erik Lucero;

[afiacco]

It's interesting because when you learn a task. Such as electrical equipment testing. Someone who has done it for years was trained by someone who had done it for years, and He or she trains you and so on. Using your knowledge you also add to the job and then train the next guy. Each person improving on the previous person's work. Quantum computing is new to everyone, yet we expect regular updates lol. Keep working we need a Faraday or a Howard Hughes crazy person to take it to the next level.

[GMTX-qjor]

Awesome presentation, loved the music analogy, big hug from Australia we all know that without vibration(waves) there is nothing

[jaychristteves]

"Success here means nothing less than a source of clean, near-limitless energy." — These would be a dream come true for humanity, when it happens!!

[TheTrophyStore]

I have had the pleasure to talk to a quantum mechanic over dinner, I was trained as a motor mechanic and found a calling for computers, self taught programmer, problem solver, tenatious explorer. We chatted indept and on the surface chasing pathways and dreaming of when it will be real. The team she is in are level pegging with the world, they claim they have cracked error correction. I understand the the musical reference to the program, I like the idea of quantum eyes. I would like to think about that for awhile. Thanks for the discussion.
Roy

[fernbear]

Honestly I think it'll help most with large language models (LMs) since the exponential scaling on information density scales linearly with the exponential increase in LM computation, meaning you should.... hopefully? Need just linear scaling for whatever level of error tolerance you desire. Which is non-quantized to begin with, so hopefully the dependence of outputs based upon some faulty-with-x%-chance system wouldn't be too terrible.

I guess I could be really failing to understand here though if it's something like the exponential divergence you'd see in a chaotic system, but the fixed latent space of the residual branch of the network should constrain that divergence to excessively manageable levels.

Could be wrong and currently still on the very much classical side of computer research, but once quantum ML takes off I plan on being near the front wave (hopefully!) Of it. :D

[atulsalgaonkar]

Very exciting.Thank you for making this available.

[Aldraz]

16:36 and you can see there's Vsauce Michael :D