TheSequence is a summary of groundbreaking ML research papers, engaging explanations of ML concepts, and exploration of new ML frameworks and platforms. We keep you up-to-date with the main AI news, trends, and technology developments. Stay in the know:

SUBSCRIBE FOR 35$/YEAR

💥 What’s New in AI: Meta’s New Architecture for Build AI Agents that Can Reason Like Humans and Animals

Humans and animals have an incredible ability to learn very complex models of the world with just a small amount of observation and experimentation. The key to this cognitive skill is our ability to reuse incredible amounts of background knowledge. Many cognitive psychologists classify this skill as “common sense”. Even though it is poorly understood, it remains a key to our cognitive architecture. From that perspective, common sense can be seen as a collection of models of the world that can guide our decision-making process.

The architecture of human cognition highly contrasts with the current generation of supervised AI models in which every task needs to be re-learned from scratch. In recent years, we have seen techniques such as self-supervised learning (SSL), which provide an alternative way to learn models of a given environment with minimum supervision. Meta (Facebook) AI Research (FAIR) has been the main champion behind the SSL school of thought. In their recent event, Inside the Lab, Meta AI’s Chief AI Scientist Yann LeCun took these efforts a step further →read further to understand the core of Yann LeCun’s architecture