Memristor

Lead author of the study, Dr. Babak Bakhit

AI and Humanoids

Brain-inspired chip could reduce AI energy use by 70%

Replicating brain-scale processing may require thousands of NVIDIA H100 Tensor Core GPUs. At 700 watts each, that would take megawatts. The brain runs on 20 watts. This organ has inspired scientists to create chips that cut energy use by 70%.

Researchers have created a new computer chip that stores data using the movements of individual atoms

Computing

Tiny, high-capacity "atomristor" stores data in single-atom switches

Engineers have created one of the smallest memory storage devices ever, made out of a 2D material measuring 1 nanometer square. The device works on the movements of single atoms, paving the way for memory systems with incredible information density.

Researchers have taken the next step in developing artificial neurons and synapses

Computing

Biological and artificial neurons connect and communicate online

Researchers in Europe and the UK have managed to connect biological and artificial neurons together. The biological neurons were grown in Italy, sent signals through an artificial synapse in the UK to communicate with artificial neurons in Zurich.

Using memristors, scientists could be getting closer to developing bionic brains Shutterstock)

Electronics

Electronic memory may bring bionic brain one step closer

Researchers claim to have constructed the world's first electronic memory cell that effectively mimics the analog process of human memory and may one day lead to the creation of the first bionic brain.

An image of a circuit with 17 memristors, each composed of two layers of titanium dioxide connected by wire, captured by an atomic force microscope. (Image: HP Labs)

Computing

Memristor breakthrough could enable computers that work like human brains

A discovery by HP that memristors can perform logic could change the way computer systems are designed and enable faster more efficient computers.