We employ a combination of recent developments in semi-supervised learning
for automatic speech recognition to obtain state-of-the-art results on
LibriSpeech utilizing the unlabeled audio of the Libri-Light dataset. More
precisely, we carry out...
As artificial intelligence increasingly influences our world, it becomes
crucial to assess its technical progress and societal impact. This paper
surveys problems and opportunities in the measurement of AI systems and their
impact, based on a...
Jack ClarkRelevant paper here:
Measurement in AI Policy: Opportunities and Challenges
Research in NLP lacks geographic diversity, and the question of how NLP can
be scaled to low-resourced languages has not yet been adequately solved.
"Low-resourced"-ness is a complex problem going beyond data availability and
Jack ClarkData distribution VS data representation - there are huge swathes of human experience which are not represented in digitized datasets. From a Masakhane paper: Participatory Research for Low-resourced Machine Translation: A Case Study in African Languages arxiv.org/abs/2010.02353pic.twitter.com/1Ga3vNZDa0
Much as replacing hand-designed features with learned functions has
revolutionized how we solve perceptual tasks, we believe learned algorithms
will transform how we train models. In this work we focus on general-purpose
learned optimizers capable...
Facebook is moving to curb internal debate around divisive political and social topics, Chief Executive Mark Zuckerberg said, after a spate of disputes and criticism that has fueled discord among staff.
Google is expanding its moderation system for internal message boards as it tries appease employees' desire for transparency and open dialogue while still cracking down on divisive distractions, as staffers work from home through summer 2021.
Jack ClarkAlso, just to nerd out a bit, but I've been reading Acemoglu's work for almost a decade, and I remember walking past their office when I was at MIT a few years ago but not knocking on the door. Now I get to learn from them!