Articles Collectés

1,002

Total

941

Non traités

61

Traités

Non traitée 19/02/2019 09:00
AI safety needs social scientists

RSS: OpenAI News

We’ve written a paper arguing that long-term AI safety research needs social scientists to ensure AI alignment algorithms succeed when actual humans are involved. Properly aligning advanced AI syste...

Non traitée 14/02/2019 09:00
Better language models and their implications

RSS: OpenAI News

We’ve trained a large-scale unsupervised language model which generates coherent paragraphs of text, achieves state-of-the-art performance on many language modeling benchmarks, and performs rudiment...

Non traitée 04/02/2019 09:00
Non traitée 19/12/2018 09:00
OpenAI Fellows Summer 2018: Final projects

RSS: OpenAI News

Our first cohort of OpenAI Fellows has concluded, with each Fellow going from a machine learning beginner to core OpenAI contributor in the course of a 6-month apprenticeship.

Non traitée 14/12/2018 09:00
How AI training scales

RSS: OpenAI News

We’ve discovered that the gradient noise scale, a simple statistical metric, predicts the parallelizability of neural network training on a wide range of tasks. Since complex tasks tend to have nois...

Non traitée 06/12/2018 09:00
Quantifying generalization in reinforcement learning

RSS: OpenAI News

We’re releasing CoinRun, a training environment which provides a metric for an agent’s ability to transfer its experience to novel situations and has already helped clarify a longstanding puzzle...

Non traitée 08/11/2018 09:00
Spinning Up in Deep RL

RSS: OpenAI News

We’re releasing Spinning Up in Deep RL, an educational resource designed to let anyone learn to become a skilled practitioner in deep reinforcement learning. Spinning Up consists of crystal-clear ex...

Non traitée 07/11/2018 09:00
Learning concepts with energy functions

RSS: OpenAI News

We’ve developed an energy-based model that can quickly learn to identify and generate instances of concepts, such as near, above, between, closest, and furthest, expressed as sets of 2d points. Ou...

Non traitée 05/11/2018 09:00
Non traitée 31/10/2018 08:00
Reinforcement learning with prediction-based rewards

RSS: OpenAI News

We’ve developed Random Network Distillation (RND), a prediction-based method for encouraging reinforcement learning agents to explore their environments through curiosity, which for the first time ...

Non traitée 22/10/2018 09:00
Learning complex goals with iterated amplification

RSS: OpenAI News

We’re proposing an AI safety technique called iterated amplification that lets us specify complicated behaviors and goals that are beyond human scale, by demonstrating how to decompose a task into s...

Non traitée 11/10/2018 09:00
OpenAI Scholars 2019: Applications open

RSS: OpenAI News

We are now accepting applications for our second cohort of OpenAI Scholars, a program where we provide 6–10 stipends and mentorship to individuals from underrepresented groups to study deep learning...

Non traitée 09/10/2018 09:00
OpenAI Fellows Winter 2019 & Interns Summer 2019

RSS: OpenAI News

We are now accepting applications for OpenAI Fellows and Interns for 2019.

Non traitée 02/10/2018 09:00
Non traitée 10/09/2018 09:00
OpenAI Scholars 2018: Final projects

RSS: OpenAI News

Our first cohort of OpenAI Scholars has now completed the program.

Non traitée 23/08/2018 09:00
The International 2018: Results

RSS: OpenAI News

OpenAI Five lost two games against top Dota 2 players at The International in Vancouver this week, maintaining a good chance of winning for the first 20–35 minutes of both games.

Non traitée 13/08/2018 09:00
Non traitée 06/08/2018 09:00
OpenAI Five Benchmark: Results

RSS: OpenAI News

Yesterday, OpenAI Five won a best-of-three against a team of 99.95th percentile Dota players: Blitz, Cap, Fogged, Merlini, and MoonMeander—four of whom have played Dota professionally—in fr...

Non traitée 30/07/2018 09:00
Learning dexterity

RSS: OpenAI News

We’ve trained a human-like robot hand to manipulate physical objects with unprecedented dexterity.

Non traitée 26/07/2018 09:00