MIT’s CSAIL masters the art of teaching AI “unlearning” to cleanse biased data- a sci-fi plot turned reality. Like a diet, substituting biased data changes AI behavior overtime.
Deep into the world of tech acronym soup, the MIT-IBM Watson AI Lab is cooking up a brain-inspired computing dish called CRAFINS, thanks to a $7.6mil helping hand from Uncle Sam.
MIT’s wizards have created an AI-training haven within your mobile, keeping private data private. It’s like a polite, well-behaved guest; no data spillage, just magic!
MIT’s clever clogs have hatched a nifty method to monitor the energy guzzling of AI’s matrix multiplication and vector adding, henceforth known as “energy-aware pruning”. Cheers MIT!
AI models, the digital know-it-alls, are learning humility. MIT and ETH Zurich are teching them to say “I don’t know” when in doubt, humanizing our robotic compatriots.
