Hacker News new | past | comments | ask | show | jobs | submit login
The robotic worm (oreilly.com)
40 points by monort on Nov 20, 2014 | hide | past | favorite | 4 comments



Really exciting research!

I've been trying to replicate this work. So far haven't succeeded. The code for the system described in the article isn't available. Only a "newer" version is available. In either case, both implementations will behave differently depending on how quickly they execute, so reproducing the described results is proving tricky. Also the motor output program isn't available, nor well defined.

When next I get time, I planned on posting to the Google Group for this project and see if they are willing to enlighten. I re-implemented the code as a single C program, and re-architected it to use a synchronous tick, moving from one state to the next. So my program has re-producible results, regardless of speed. It just doesn't appear to behave correctly yet. More tinkering...


This will be much harder with other species since so little is known about the neural system of other species when compared to C elegans. It is very easy to study the wrom since it is clear and you can track the development of its neurons from an early stage. It is also easy to run mutagenesis experiments on them as well.


Hopefully in doing these explorations we'll glean higher level understanding of neural system that can then be applied to other species rather than having to start from scratch again.


Here are some more info including a video. http://www.i-programmer.info/news/105-artificial-intelligenc...




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: