Over 100 girl geeks attended a Girl Geek Dinner at OpenAI HQ in San Francisco to learn about their recent work in reinforcement learning, robotics, AI policy, and more! https://girlgeek.io/girl-geek-x-openai-lightning-talks-and-panel-video-transcript
Learn more about sponsoring your own Girl Geek Dinner: https://girlgeek.io/sponsor
0:00 Event Highlights
2:12 “Welcome to OpenAI Girl Geek Dinner”
Gretchen DeKnikker, COO, Geek X
Sukrutha Bhadouria, CTO, Girl Geek X
4:13 “Intro to OpenAI”
Ashley Pilipiszyn, Technical Director
6:29 “Building the first esports world-champion-defeating AI”
Brooke Chan, Software Engineer, Dota team
Early this year, OpenAI Five became the first AI to defeat the world champions in an esports game. We trained a system to understand and play Dota 2 at a high level, demonstrating how pairing self-play with reinforcement learning can lead to superhuman performance at very difficult tasks. In this talk, Brooke will give some background on the 2 year project as well as an inside look at challenges that come with training an AI on a very complex game.
21:42 “Using reinforcement learning (RL) to learn dexterous in-hand manipulation policies”
Lilian Weng, Research Scientist, Robotics team
In this talk, Lilian will introduce how the Robotics team uses reinforcement learning (RL) to learn dexterous in-hand manipulation policies which can perform vision-based object reorientation on a physical Shadow Dexterous Hand. Our policies are trained in a simulated environment in which we randomize many physical properties of the system and an object’s appearance – the method is known as Domain Randomization. Although the training happens entirely in a simulator, our policies transfer to the physical robot well.
36:50 “Using AI to create and collaborate with music”
Christine Payne, Research Scientist, Music Generation team
We’ve created MuseNet, a deep neural network that can generate 4-minute musical compositions with 10 different instruments. It can combine styles from country to Mozart to the Beatles. MuseNet was not explicitly programmed with our understanding of music, but instead discovered patterns of harmony, rhythm, and style by learning to predict the next token in hundreds of thousands of MIDI files. In this talk, Christine will share how it can be used to push the boundaries of AI creativity, both as an independent composer, and as a collaboration tool with human artists.
50:51 “Ensuring artificial general intelligence (AGI) benefits all of humanity”
Mira Murati, RL Team Manager, Reinforcement Learning team
At OpenAI, our mission is to ensure that artificial general intelligence (AGI) benefits all of humanity, primarily by attempting to build safe AGI and share the benefits with the world.
1:02:27 “What is AI Policy?”
Amanda Askell, Research Scientist, Policy team
What is AI policy? AI systems can fail unexpectedly, be used in ways their creators didn’t anticipate, or having unforeseen social consequences. But the same can be said of cars and pharmaceuticals, so why should we think there are any unique policy challenges posed by AI? In this talk, Amanda will point out that the primary mechanisms for preventing these sorts of failures in other industries are not currently well-suited to AI systems. Amanda will then discuss the ways that engineers can help meet these policy challenges.
1:21:41 Q&A
Want more Girl Geek X?
ATTEND: https://girlgeek.io/attend
CONFERENCE: http://elevate.girlgeek.io
READ: https://girlgeek.io/read
LISTEN: https://girlgeek.io/podcasts
WATCH: https://girlgeek.io/watch
SPONSOR: https://girlgeek.io/sponsor
SUBSCRIBE: https://girlgeek.io/subscribe
About OpenAI: At OpenAI, our mission is to ensure that artificial general intelligence (AGI) benefits all of humanity, primarily by attempting to build safe AGI and share the benefits with the world.
This OpenAI Girl Geek Dinner was held on September 10, 2019.
