engblogs

summaries of the latest blog articles from your favorite tech companies.
OpenAIOpenAI

Scaling laws for reward model overoptimization

Understanding the scaling laws involved in overoptimization of reward models.

10/19/2022
DoorDash MLDoorDash ML

Augmenting Fuzzy Matching with Human Review to Maximize Precision and Recall

Using human review in combination with a classification model to improve precision and recall in fuzzy matching

10/18/2022
SoundCloudSoundCloud

What's New with SoundCloud, October 2022

Updates to the SoundCloud iOS and Android apps to improve user experience based on user feedback.

10/17/2022
Google DeepMindGoogle DeepMind

Stopping malaria in its tracks

Developing a better malaria vaccine with the help of AI that could save hundreds of thousands of lives every year

10/13/2022
The New York TimesThe New York Times

Collective Decision-Making with AHP

The blogpost discusses how the NYT Identity team used the Analytic Hierarchy Process (AHP) to make collective decisions on selecting a user ID format.

10/13/2022
SoundCloudSoundCloud

Android Image Loading at SoundCloud

Exploring the new image loading features in the recent SoundCloud Android app redesign.

10/13/2022
Google DeepMindGoogle DeepMind

Measuring perception in AI models

Introducing the Perception Test, a multimodal benchmark using real-world videos to evaluate perception capabilities of AI models.

10/12/2022
Microsoft AIMicrosoft AI

From Hot Wheels to handling content: How brands are using Microsoft AI to be more productive and imaginative

Brands are leveraging Microsoft AI to increase productivity and creativity

10/12/2022
Google DeepMindGoogle DeepMind

Measuring perception in AI models

Introducing the Perception Test: a multimodal benchmark to evaluate the perceptual capabilities of AI models using real-world videos.

10/12/2022
Google DeepMindGoogle DeepMind

How undesired goals can arise with correct rewards

Exploring the emergence of undesired goals despite correct rewards in AI systems

10/7/2022
Google DeepMindGoogle DeepMind

How undesired goals can arise with correct rewards

Understanding how AI systems can unintentionally pursue undesired goals due to goal misgeneralisation, even when trained with correct specifications.

10/7/2022
Microsoft AIMicrosoft AI

Microsoft open sources its ‘farm of the future’ toolkit

Microsoft open sources its ‘farm of the future’ toolkit.

10/6/2022