While the exponential growth in data and the increase in compute power have been key drivers of AI innovation, without research — and the tremendous effort by academics and companies that invest in it — these advances wouldn’t go very far.
At his Transform 2019 talk last month, Lyft’s head of product and machine learning, Gil Arditi, focused on this research and ripped through example after example of research in the past year that’s propelled head-turning advances in AI and machine learning. Below are just a few of the examples he covered (watch his entire talk above).
Deep reinforcement learning
Two AI companies have been at the center of deep reinforcement learning and have used video games as their research playground. OpenAI Five has used the technology to defeat humans in Dota 2, a multiplayer online battle arena video game — and has won 4,075 games against human players for a victory rate of 99.4%. Alphabet’s DeepMind is demonstrating similar success in pitting machine against human with its AlphaStar AI agent playing against real gamers in Starcrat II. During a livestream against pros, AlphaStar swept in all 10 matches.
To accomplish those achievements, Arditi said, “OpenAI Five trained for more than 45,000 years of gameplay, and AlphaStar took more than $26 million in the monetized value of the compute resources they used.”
Natural language processing
Until recently, Google-developed BERT solved a wide range of problems in NLP, including things like sentence classification, sentence space similarity, questions, and answers. Arditi asserts there’s a new contender that emerged from a research paper several weeks ago called XLnet. “It’s outperforming BERT on some new benchmarks,” he explained.
GPT-2 model for text prediction
You can think of this as the world’s best text generator where a model is given a sentence, and then it generates subsequent sentences that are supposed to follow in the same theme.
Arditi demonstrated this with the input sentence, “A train carriage containing controlled nuclear materials was stolen in Cincinnati today. Its whereabouts are unknown.” The prediction by the system continued, “The incident occurred on the downtown train line, which runs from Covington and Ashland stations. In an email to Ohio news outlets, the U.S. department of energy said it is working with the federal railroad administration to find the thief.” As Arditi noted, “It’s a huge step forward that also creates some potential for fake news. This is very timely, because of the upcoming election cycle.”
Identifying fake news
Of course every problem created by technology needs a solution. “Fake news is something that’s very hard for humans to understand [and] to see,” said Arditi. Enter Grover, being touted as a state-of-the-art defense against fake news. “This algorithm is doing really well in both generating text and in classifying text as being fake, as being not real, not generated by humans. It’s also looking at the source. It can adapt to the style of a paper like The Washington Post versus a lesser-known blog.”
Arditi also cited generative models, the generation of high-res images and video (which will have significant impact for both video game development and sophisticated productions that currently require armies of animators), style transfer in 3D, and many more.
See Arditi’s entire talk here — and check out all the other sessions you missed at Transform 2019.
Credit: Google News