Openai

Running Codex safely at OpenAI

How OpenAI runs Codex securely with sandboxing, approvals, network policies, and agent-native telemetry to support safe and compliant coding agent adoption.

Advancing youth safety and wellbeing in EMEA

Explore OpenAI’s European Youth Safety Blueprint and EMEA Youth & Wellbeing Grants, advancing safe, responsible AI for teens, families, and educators.

How ChatGPT learns about the world while protecting privacy

Learn how ChatGPT safeguards your privacy, reduces personal data in training, and gives you control over whether your conversations improve AI models.

Introducing OpenAI

OpenAI is a non-profit artificial intelligence research company. Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return. Since our research is free from financial obligations, we can better focus on a positive human impact.

Weight normalization: A simple reparameterization to accelerate training of deep neural networks

Team++

We've had some fantastic people join over the past few months (and we're still hiring). Welcome, everyone!

Welcome, Pieter and Shivon!

We have two more team updates.

OpenAI Gym Beta

We’re releasing the public beta of OpenAI Gym, a toolkit for developing and comparing reinforcement learning (RL) algorithms. It consists of a growing suite of environments (from simulated robots to Atari games), and a site for comparing and reproducing results.

Adversarial training methods for semi-supervised text classification

Team update

We’d like to welcome the latest set of team members to OpenAI (and we’re still hiring!)

Generative models

This post describes four projects that share a common theme of enhancing or using generative models, a branch of unsupervised learning techniques in machine learning. In addition to describing our work, this post will tell you a bit more about generative models: what they are, why they are important, and where they might be going.

OpenAI technical goals

OpenAI’s mission is to build safe AI, and ensure AI’s benefits are as widely and evenly distributed as possible.

Concrete AI safety problems

We (along with researchers from Berkeley and Stanford) are co-authors on today’s paper led by Google Brain researchers, Concrete Problems in AI Safety. The paper explores many research problems around ensuring that modern machine learning systems operate as intended.

Special projects

Impactful scientific work requires working on the right problems—problems which are not just interesting, but whose solutions matter.

Team update

We’ve hired more great people to help us achieve our goals. Welcome, everyone!

Machine Learning Unconference

The latest information about the Unconference is now available at the Unconference wiki, which will be periodically updated with more information for attendees.

Infrastructure for deep learning

Deep learning is an empirical science, and the quality of a group’s infrastructure is a multiplier on progress. Fortunately, today’s open-source ecosystem makes it possible for anyone to build great deep learning infrastructure.

Transfer from simulation to real world through learning deep inverse dynamics model

Report from the self-organizing conference

Last week we hosted over a hundred and fifty AI practitioners in our offices for our first self-organizing conference on machine learning.

Semi-supervised knowledge transfer for deep learning from private training data

Extensions and limitations of the neural GPU

Variational lossy autoencoder

RL²: Fast reinforcement learning via slow reinforcement learning

A connection between generative adversarial networks, inverse reinforcement learning, and energy-based models

On the quantitative analysis of decoder-based generative models

#Exploration: A study of count-based exploration for deep reinforcement learning

OpenAI and Microsoft

We’re working with Microsoft to start running most of our large-scale experiments on Azure.

Universe

We’re releasing Universe, a software platform for measuring and training an AI’s general intelligence across the world’s supply of games, websites and other applications.

Faulty reward functions in the wild

Reinforcement learning algorithms can break in surprising, counterintuitive ways. In this post we’ll explore one failure mode, which is where you misspecify your reward function.

PixelCNN++: Improving the PixelCNN with discretized logistic mixture likelihood and other modifications

Team update

The OpenAI team is now 45 people. Together, we’re pushing the frontier of AI capabilities—whether by validating novel ideas, creating new software systems, or deploying machine learning on robots.

Adversarial attacks on neural network policies

Attacking machine learning with adversarial examples

Adversarial examples are inputs to machine learning models that an attacker has intentionally designed to cause the model to make a mistake; they’re like optical illusions for machines. In this post we’ll show how adversarial examples work across different mediums, and will discuss why securing systems against them can be difficult.

Third-person imitation learning

Prediction and control with temporal segment models

Emergence of grounded compositional language in multi-agent populations

Learning to communicate

In this post we’ll outline new OpenAI research in which agents develop their own language.

Distill

We’re excited to support today’s launch of Distill, a new kind of journal aimed at excellent communication of machine learning results (novel or existing).

One-shot imitation learning

Evolution strategies as a scalable alternative to reinforcement learning

We’ve discovered that evolution strategies (ES), an optimization technique that’s been known for decades, rivals the performance of standard reinforcement learning (RL) techniques on modern RL benchmarks (e.g. Atari/MuJoCo), while overcoming many of RL’s inconveniences.