Image record for AI

IBM AI breakthrough sees firm smash Microsoft's image recognition record 

Firm has developed distributed deep learning software to speed things up

IBM image record

BIG BLUE (IBM) has beaten Microsoft's AI image recognition accuracy record by developing distributed deep learning (DDL) software.

IBM said that, given any random image from a set of 7.5 million pictures from ImageNet-22K database, the trained AI model achieved 33.8 per cent accuracy, trumping the previous record of 29.8 per cent set by Microsoft back in October 2014.

What's more, IBM's system managed to achieve this in seven hours, while the process that allowed Microsoft to set the previous record took 10 days to complete.

 

Comments are closed.