Published in News

Finnish brainiacs invent futuristic quantum algorithms

by on14 May 2024


You must be nicer to the cats

A team of Finnish boffins claim to be close to creating algorithms for quantum computers so that they can solve real world problems

According to Pat Brans from Pat Brans Associates/Grenoble Ecole de Management the method could be a game changer.

The Finnish Technical Research Center (VTT) teamed up with CSC, the folks running LUMI (Europe's speediest supercomputer), to hook up ever-bigger quantum computers to LUMI. This lets users mess around with algorithms that play nicely with quantum and classical computers. The old-school computer holds onto most of the data and does the heavy lifting, while the quantum computer gets the fancy steps and spits back the results.

VTT's quantum journey started in 2021 with a modest 5 qubits. They levelled up to 20 qubits last year and are gunning for 50 by the end of 2024. Some universities and research nerds already use this hybrid setup to crack simple problems.

Ville Kotovirta, the head honcho of VTT's quantum algorithm and software squad, says it's like a dress rehearsal for the quantum future. It's a playground for supercomputing algorithm whizzes to dream about the cool stuff they'll do when quantum and classical computers join forces. They are getting their hands dirty with algorithm writing as we speak.

VTT was the first in Europe to rock this hybrid vibe, but now others are hopping on the bandwagon. Last June, the European HPC joint undertaking gave the thumbs up to six more projects to build similar setups in Czechia, France, Germany, Italy, Poland, and Spain.

Now, Kotovirta's crew is busy whipping up all sorts of algorithms for these hybrid beasts. They've got this thing called quadratic unconstrained binary optimisation (QUBO), which quantum annealing or quantum approximate optimisation algorithms (QAOAs) can solve.

They're also diving into graph data analysis, peeping at the community vibes of networks, and tackling quantum chemistry to make molecule simulations less of a headache. Plus, they're playing god with synthetic biology, cooking up new proteins with all the right stuff.

They are all about that quantum generative machine learning life, teaching models to learn from existing data and spit out fresh samples.

Kotovirta said that while most people are using generative AI makes all those fake pics, texts, and sounds, he was using that same magic with quantum computing to pop out new proteins.

Rate this item
(2 votes)

Media