YES! I found my “leaky triangle” (old idea of mine) in computational neuroscience:

YES! I found my “leaky triangle” (old idea of mine) in computational neuroscience: [I DIDN’T make this graphic btw 🙂 ]

The “Leaky Integrate and Fire” neuron model, illustrated with a leaky bucket. Here’s it’s used as a way to get robots to have short-term memory so they don’t get stuck in corners, but it’s usually shown as a spike that decays slowly.

The benefit of having a leaky bucket is so that you’ll start to forget. If you remembered EVERYTHING you might over-react (think PTSD) to triggers but if you can remember most things for short periods of time and start to forget before remembering again, you get room for new information to be integrated with the old, which then decays itself, allowing you to forget some more (bucket leak), allowing you room to integrate more new information before *it* starts to leak… and so on and so on and so on.

If this doesn’t make sense, the fault is mine for explaining it poorly. [I got this graphic from a robots site but I liked the illustration]. Point is, I’ve been looking for this but didn’t know where it was or what it was called.

leakyintegrator

Attachments

[responsivevoice_button voice="US English Male"]

Leave a comment

Your email address will not be published. Required fields are marked *


8 − six =

Leave a Reply