View on GitHub

Hi! I'm Lana.

I research Artificial Life and AI in Sony CSL's Kyoto Lab.


Artficial Life - Open Science - Artificial Intelligence

Who am I?

I am an associate researcher (and founding member) of Sony Computer Science Labs’ new Kyoto Lab. Read more about my motivations for starting the lab by clicking on my picture here (English and Japanese). I am also the Research Chair on the International Society for ALife’s board of directors. I research learning and predictive systems, and I am interested in sanitizing science practices. You can read about my career path on my CV (tl;dr: I went to Engineering School in France then moved to Japan where I got my Master’s and PhD. I used to speak ~9 languages and can do just about anything that involves programming. Sometimes I get awards for scientific work or non scientific work. My favorite scientist is Frans De Waal. I like food.)

Follow me

Message me on Twitter @sina_lana

Read me

I wrote about changing science publishing through the collaborative open science platform I am developing:

Learn about Artificial Life in my Introduction to Artificial Life for People who Like AI on The Gradient.

I write about my ideas and curate exciting science on Medium

This very old blog has scientific ideas as well as more personal stories.

Most of my papers are listed in my Google Scholar profile

What am I working on?

I am interested in systems that learn in unconventional ways, as well as unconventional systems that learn in mainstream ways. Currently I work a lot on predictive coding. Some of my projects:

I research visual illusions in neural networks, biological and artificial, on the principle that systems that “fail” as the brain does teach us more than systems that superficially succeed as humans do. I am not so much interested in perfect artificial vision as I am interested in charmingly and accurately faulty artficial perception.

I also work on measures of complexity and measures of “life”. How could you detect life on faraway planets? What kind of data and how much of it do you need to make a reasonable assumption about whether a system is alive or not?

I am building a platform for better science.

Other projects of mine include putting predictive networks in mobile phones and recording weird behavior in plants.

bw_1 bw_2 bw_3

c_1 c_2

Motion illusions produced by my evolutionary generator.