OpenAI News

OpenAI News · February 15, 2018 08:00

Interpretable machine learning through teaching

We’ve designed a method that encourages AIs to teach each other with examples that also make sense to humans. Our approach automatically selects the most informative examples to teach a concept—for instance, the best images to describe the concept of...

OpenAI News · February 07, 2018 08:00

Discovering types for entity disambiguation

We’ve built a system for automatically figuring out which object is meant by a word by having a neural network decide if the word belongs to each of about 100 automatically-discovered “types” (non-exclusive categories).

OpenAI News · January 31, 2018 08:00

Requests for Research 2.0

We’re releasing a new batch of seven unsolved problems which have come up in the course of our research at OpenAI.

Topics: Models

Entities: OpenAIModels

OpenAI News · January 18, 2018 08:00

Scaling Kubernetes to 2,500 nodes

We’ve been running⁠(opens in a new window) Kubernetes⁠(opens in a new window) for deep learning research for over two years. While our largest-scale workloads manage bare cloud VMs directly, Kubernetes provides a fast iteration cycle, reasonable...

OpenAI News · December 06, 2017 08:00

Block-sparse GPU kernels

We’re releasing highly-optimized GPU kernels for an underexplored class of neural network architectures: networks with block-sparse weights. Depending on the chosen sparsity, these kernels can run orders of magnitude faster than cuBLAS or cuSPARSE. We’ve...

Topics: Infrastructure

Entities: Infrastructure

OpenAI News · December 04, 2017 08:00

Learning sparse neural networks through L₀ regularization

We propose a practical method for L₀ norm regularization for neural networks: pruning the network during training by encouraging weights to become exactly zero. Such regularization is interesting since (1) it can greatly speed up training and inference, and...

Topics: Infrastructure

Entities: Infrastructure

OpenAI News · November 02, 2017 07:00

Interpretable and pedagogical examples

Teachers intentionally pick the most informative examples to show their students. However, if the teacher and student are neural networks, the examples that the teacher network learns to give, although effective at teaching the student, are typically...

Topics: Infrastructure

Entities: Infrastructure

OpenAI News · October 26, 2017 07:00

Learning a hierarchy

We’ve developed a hierarchical reinforcement learning algorithm that learns high-level actions useful for solving a range of tasks, allowing fast solving of tasks requiring thousands of timesteps. Our algorithm, when applied to a set of navigation problems,...

OpenAI News · October 19, 2017 07:00

Generalizing from simulation

Our latest robotics techniques allow robot controllers, trained entirely in simulation and deployed on physical robots, to react to unplanned changes in the environment as they solve simple tasks. That is, we’ve used these techniques to build closed-loop...

OpenAI News · October 18, 2017 07:00

Sim-to-real transfer of robotic control with dynamics randomization

Simulations are attractive environments for training agents as they provide an abundant source of data and alleviate certain safety concerns during the training process. But the behaviours developed by agents in simulation are often specific to the...

Topics: AgentsInfrastructure

Entities: AgentsInfrastructure

OpenAI News · October 18, 2017 07:00

Asymmetric actor critic for image-based robot learning

Deep reinforcement learning (RL) has proven a powerful technique in many sequential decision making domains. However, Robotics poses many challenges for RL, most notably training on a physical system can be expensive and dangerous, which has sparked...

Topics: PolicyInfrastructure

Entities: PolicyInfrastructure