From MIT Technology Review: algorithms are making American inequality worse. My takeaway from this piece is not that it is the algorithms themselves but the thought processes behind them. Algorithms simply optimize the achievement of the goals you set out for. Based on unfair and unjust starting premises, algorithms just make the injustice more efficient.
From IBM Research: Automated analysis of free speech predicts psychosis onset in high-risk youths. This technique uses AI to analyze speech patterns of clinical high risk youths to mark for “poverty of speech” and “flight of ideas.”
AI with 18,000 GPUs develops its own neural networks. The more I read about stuff like this the more I think we’re getting closer to the future of Asimov’s Foundation.
a deep learning technology to reconstruct the true genome sequence from HTS sequencer data with significantly greater accuracy than previous classical methods
DeepMind taught itself to play chess and then went on to beat all other chess bots. What’s interesting about this AI is that it’s willing to give up material advantage for positional advantage ten moves down the line.
JPL has built an AI-controlled autonomous quadcopter drone. They put it through its paces on a racing course against a pro pilot. The human opponent still came out three minutes ahead, but it’s still impressive considering the drone generated the map entirely by itself.
‘Surprisingly accurate’ models of political leanings based on cars in a neighborhood. This bothers me a lot. A whole lot.
AI is now designing games. And it’s not just level design either.
DARPA Lifelong Learning Machines (L2M), highly adaptable AI that can continuously learn and improve. SkyNet in the making?