🥸 Edge#158: Microsoft KEAR is a Deep Learning Model for Common Sense Reasoning
What’s New in AI, a deep dive into one of the freshest research papers or technology frameworks that are worth your attention. Our goal is to keep you up to date with new developments in AI in a way that complements the concepts we are debating in other editions of our newsletter.Â
💥 What’s New in AI: Microsoft KEAR is a Deep Learning Model for Common Sense ReasoningÂ
Common sense is one of the cognitive qualities of the human brain that are hard to quantify or even explain. Not surprisingly, recreating common sense reasoning in ML models has become one of the most relevant challenges of the entire space. The challenges recreating common sense are at the root of one of the fundamental philosophical dilemmas in ML: the friction between logic and knowledge. Through evolution, humans developed knowledge models of the world first and incorporated logical reasoning. This process was reverted in the software industry in which logical models achieved adoption before neural networks. That created two highly distinctive schools of thought. Nowadays, architectures that combine logical reasoning and neural networks have become far from trivial.Â
From the ML perspective, the discipline that has become the center of common sense reasoning capabilities has been question-answering (QA) models.