We've reverted the title from the submitted “Method shrinks NNs by 90% with sub-1% accuracy dropoff, vs. 3.5% for prior SOA”. Submitters: there's no need to try

We've reverted the title from the submitted “Method shrinks NNs by 90% with sub-1% accuracy dropoff, vs. 3.5% for prior SOA”. Submitters: there's no need to try to pack the article into 80 characters! Titles and articles are two separate things.