Massachusetts Institute of Technology (MIT) researchers have designed a technique to mass manufacture robots as small as a cell that can be employed to observe conditions within a gas or oil pipeline, or to study out disorders while floating via the bloodstream.
The solution to making such tiny robots, which the team dubs as “syncells” (abbreviation for synthetic cells), in huge quantities lies in managing the natural fracturing procedure of brittle and atomically-thin materials. The procedure, dubbed as “autoperforation,” manages the fracture lines so that they create miniscule pockets of a predictable shape or size.
Embedded within these pockets are electronic materials or circuits that can record, collect, and output info, as per to the research published in the Nature Materials journal. The system, designed by scientists at the MIT in the U.S., employs a 2D form of carbon dubbed as graphene, which makes the outer layer of the minute syncells.
On a similar note, while the speed of ML (machine learning) has increased over the past decade, the underlying hardware allowing ML tasks has not altered much. This includes tons of conventional processing chips, such as GPUs (graphics processing units) and CPUs (computer processing units) equipped in huge data centers.
But on the edge of processing is a field dubbed as neuromorphic computing, which looks to make computer processors work more akin to the human brain. Scientists at MIT earlier revealed a revolutionary neuromorphic processor design that can represent the next step for AI tech.
The secret behind this is a design that generates a synthetic synapse for “brain on a processor” hardware. Digital processor of modern day makes computations on the basis of on/off and binary signaling. Neuromorphic processors rather work in a similar fashion, exchanging bunches of electric signals at different intensities, similar to the neurons present in the brain.