Unit News

Drawing showing a toddler and a child-like robot back-to-back investigating abstract 3D shapes with magnifying glasses.

Research Update

Curious robots mimic how children can learn to understand language

Rewarding AI for playful behavior helps them learn tasks much more efficiently than rote instruction and provides valuable insights into human language acquisition.

A human outline in neon pink sits on a background of wire patterns in bright neon colors. Code in speech bubbles emerges from the back of the outline

Research Update

AI learns better when it talks to itself

Inner speech combined with working memory helps AI to learn and generalize across tasks

28 January 2026

Illustration of an anthropomorphized robot stacking colored blocks back to back with a human toddler.

Research Update

New, embodied AI reveals how robots and toddlers learn to understand

With a brain-inspired architecture, the model offers insights into cognitive development and information processing in neural networks.

23 January 2025

Efficiency versus adaptability behavior in AI and humans header image

Research Update

Simplicity versus adaptability: Understanding the balance between habitual and goal-directed behaviors

New study on learning provides insights into adaptive behaviors.

Research Update

Expectations vs reality: How robots update their behavior based on their experience

The unpredictability of everyday life can inform our expectations of the future, says Professor Jun Tani, who is revealing the underlying neural networks of predictions by exploring robotic minds.

13 December 2021

Research Update

Expecting the Unexpected: A New Model for Cognition

Researchers develop computer model of prediction and recognition

21 October 2019