🚨 RESEARCHERS HAVE DEMONSTRATED DIRECT TRAINING OF A QUANTUM NEURAL NETWORK ON REAL QUANTUM HARDWARE.A team from IonQ, Université Paris Cité, QC Ware, and others has developed a new training framework that makes gradient-based optimization of quantum neural networks significantly more efficient.The approach uses three key innovations: a specialized “Butterfly” circuit architecture, layer-wise training, and a parallelized gradient estimation method. Together, these reduce the number of quantum c

Exact time evolution of a free massless scalar field! This is a theoretically very boring thing (day 1, week 1 of a quantum field theory course), but... how do I explain this... to a lattice field theorist, a lot of algorithms scale really badly. Many PhD theses have been written about a 32x32 simulation domain. It goes like: you get your 32x32 domain, you run it on a supercomputer for a day, and you've only generated 10 independent samples. I'm exaggerating a bit, but you do have to set your si

People often talk about quantum computing like it's some kind of magic.But when you break it down, it's really just math.Today's computers use bits, and each bit can only be a 0 or a 1.Quantum computers use qubits.Instead of being only 0 or 1, a qubit can exist in a range of probabilities between the two. This gives quantum computers a completely different way of processing information.That's why they're so powerful, and why people are concerned about their impact on cryptography.Many of the sys