Cloud Hosting

One of the most significant leaps came in the realm of artificial reasoning. For years, artificial neural networks have struggled with tasks requiring analogical thinking, relying instead on brute-force statistical analysis. However, a groundbreaking March 2023 research from IBM Zurich upended this paradigm. By leveraging the power of hyperdimensional computing, a technique that encodes information in complex, multi-dimensional vectors, researchers successfully tackled the Ravens progressive matrix, a notoriously difficult abstract reasoning test. With more focused work over the upcoming years, in this field , Researchers like Abbas Rahimi hopes that this approach of vector-driven AI could not only be faster, and more transparent, but can also prove to be more energy efficient. Through which we can substantially reduce energy consumption, lowering it up to 10 or even 100 times, which can directly transfer to lowering down the carbon footprint, which current AI platforms lack. This achievement marks a crucial step towards building AIs that can truly reason, understand analogies, and navigate the complexities of the real world.

Here is the original post:

BREAKTHROUGHS IN COMPUTER SCIENCE 2023: A YEAR OF AI - Medium