Dear Duolingo: Why is spoken language so hard to understand?
Understanding the difficulties of comprehending spoken language in a new language.
Announcing the Cohere For AI Research Grant Program
Introducing the Cohere For AI Research Grant Program, promoting collaboration between academia and industry for responsible innovation.
Modular: What’s the difference between the AI Engine and Mojo?
Exploring the distinctions between the AI Engine and Mojo in the Modular system.
On the Stepwise Nature of <br> Self-Supervised Learning
Exploring the stepwise nature of self-supervised learning
Google at ACL 2023
This blogpost provides an overview of Google's participation at ACL 2023, including showcasing their latest research, active involvement in workshops and tutorials, and various papers presented in the field of natural language processing and understanding.
Complete Personalization, Complete Control: The Composable CDP
A blogpost about how organizations compete for consumer attention in a crowded retail marketplace and the role of a Composable CDP in achieving complete personalization and control.
3 Questions: Honing robot perception and mapping
Exploring how future robots enhance perception and mapping capabilities for better interaction with the environment, as discussed by Luca Carlone and Jonathan How of MIT LIDS
How Instacart’s item availability evolved over the pandemic
This blogpost explores how Instacart's item availability has evolved over the pandemic, outlining the challenges faced and the innovations in product, machine learning, and engineering that helped address them.
Google at ACL 2023
A blogpost about Google's participation and research at ACL 2023, showcasing over 50 publications and active involvement in workshops and tutorials.
Special report: Which country studies English the most?
Analysis of English language learners worldwide to determine which country studies English the most, along with surprising reasons behind their high levels of engagement.
On the Stepwise Nature of Self-Supervised Learning
This blogpost discusses the stepwise behavior observed in self-supervised learning, presenting a mathematical model that describes the training process of large-scale SSL methods and demonstrating its applicability in realistic settings.
Minetester: A fully open RL environment built on Minetest
An overview of a fully open RL environment built on Minetest and preliminary work