top of page

Mathematical models of curiosity

I am attempting to formalize curiosity [defined as intrinsic reward that is proportional to the progress of learning] in a rigorous mathematical model, such that it can explain curiosity-driven behavior and increase our understanding of the brain mechanisms behind it. This project is based on an information-theoretic model I developed, called Hierarchical curiosity loops (HCL) and later extended to Novelty Management. The model's first implementation was recently published, and its predictions of rodents' exploratory behaviors were corroborated. (Papers: GFA14a. Posters: Model, Animal implementation. Lecture: Redwoord Center, Berkeley, CA).

 

 

Curious robots

I believe that once a neuro-inspired model is developed, it can be implemented in robots to assess and further investigate their behaviors. I first implemented my Hierarchical curiosity loops (HCL) model on a simple robot that learned how to reach, exhibiting emergent behavior, similar to infants' (G11, GA12). In an on-going experiment, I continued with implementation of basic curiosity models in the social robot DragonBot, where it learned which of its social behaviors cause humans to interact longer.

Enhancing children's curiosity with social robots

I am investigating how curious robots can learn and interact with curious children. Can one influence the other's curiosity? How can children's curiosity-driven behavior be modeled, assessed and promoted? Watch this video.

Global literacy project

I am part of the Global Literacy project that distributes tablets with educational apps in various areas around the globe in the attempt to bring literacy to places that has none. By collecting vast amounts of data about the tablet usage, we are trying to learn more about the curiosity-driven behavior of children from Uganda, Ethiopia, South-Africa, India and more.

Infant-Agent interaction

An audacious research project attempts to develop an intelligent mobile-for-infants, that can sense and interact with infants in a guided, language-based manner. The primary goal of the project is to see whether early assessments and pro-active interaction with verbal and non-verbal communication can inform us of normal and non-normal development of infants.

Quantum Computer Games

As a physicist trying to explain the world of quantum physics to non-scientists, I have developed an entertaining and educational tool I called "Quantum Computer Games". These are known computer games with a quantum-twist. The known games, such as Minesweeper or Fox and Hounds, are augmented with quantum-physics based rules, such as superposition, particle-wave duality and quantum entanglement. Thus, while still being fun, the player is introduced to the counter-intuitive concepts of the most successful physical theory to-date. (Quantum minesweeper: Android app, GG10, play online. Schroedinger cat and hounds: GG12, play on-line)

The Research Fountain

In this Blog, I post my multi-disciplinary research ideas. Each post is independent and briefly outlines research projects. Check out:

  • Rewarding derivatives

  • Universal programming language

  • Uterosound

  • Portable motor

  • Adorable screens

  • Wearable bats

  • and more...

bottom of page