brains, bees, and sea

Several more days of neuromorphic fun.  There are lots of things we don't understand about the brain: very little of the brain is active at one time, neurons send other neurons encapsulated mRNA messages, split brain patients become functions extremely quickly, dendrite processing is an important process. I led a session on embodied physical reservoir computing.  A session on how LLMs work was illuminating.  There is work on building biotransistors which can be made on flexible substrates, and work in water.  The brain includes a "neuromorphic twin" model of the world, which it uses to predict sensory data; if there is a mismatch, there can be peculiar sensations.  The brain has as many glial cells, including astrocytes, as neurons; a model of these at one extreme of a spectrum of structure looks a bit like a dense associative memory, what capabilities might other places on the spectrum provide? AI basically hasn't incorporated any discoveries from neuroscience in the last 50 years.  Deep learning, with its billions of parameters, works because a particular Riemannian metric is more likely to exist in these high dimensions giving a unique global minimum.  Insects diverged from other animals before there were brains, so have rather different structures. Bee navigation establishes an absolute reference frame by using the sun and time of day.  

Whew!  That's been a lot.  I'm flying home tomorrow, so today I spent time walking, decompressing by watching the sea up close (it's profoundly weird to stand close to the shore for quite a while, and the tide neither comes in nor goes out), and a bit of paddling.

There's a lovely sandy beach in front of the hotel, which ends at these rocks.  Standing listening to the water gently lapping against them is incredibly relaxing.

more brains, and coastal views

We've had multiple days of great talks and conversations, and lots of walks.  We've heard about evolving computer programs to become replicators and machine learning programs, evolving self-organising developmental pathways, song learning in zebra finches, navigation and learning "place" maps (when you move to a different room, you switch to a different mental map; might that be why you forget why you moved to the new room?), computing with spikes, future neuromorphic hardware, making brain-like circuits with very simple components (resistors, capacitors, batteries and memristors), a mnemonic for understanding the effect of different operations and memories (if one bit is a 1 ft box, then an adder is a room, a multiplier is the whole hotel, the closest memory cache is down on the sea shore, DRAM is Corsica, a hard drive is in Rome, and the cloud is on Jupiter!). Today the focus was on robots with neuromorphic brains, and the key role played by the environment.

May brain is full to overflowing, and we've still got a couple of days to go!  To decompress a bit, I went for a walk along a costal path.  Low rocky cliffs, with a few beaches scattered along the length.

some of the coves are sandy; this one is splendidly rocky

brains and trees

Today at CapoCaccia 2025 we had some discussion sessions on "Life with and without a cortex".  We learned that you can take the cortex out of a mouse and it makes hardly any difference: although the mouse cannot learn new behaviours, the majority of behaviours it uses are in fact innate.  One speaker said that looking at mice doesn't help to understand our brains: we need to look at more nearby species, eg monkeys, and even so, that only gets you so far.  And there was the fun idea that each of the seven deadly sins is associated with a particular hormone.

More conversations, then I went exploring a bit further afield that yesterday.  Unfortunately my phone's 4G isn't working, and there's no wifi outside.  So I downloaded a map of the vicinity, and took great care to remember where I had come from.  I found some good trails through the local pine forest.  Again, stunning.

view from a hotel window : CapoCaccia 2025

I arrived in northern Sardinia last night, for the two week Capo Caccia workshop on Neuromorphic Intelligence.  This morning I saw the view from my window: this is on the "non seaview" side of the hotel, which means the room is a little cheaper.  It's still a lovely view.

The restaurant looks over the sea, and the view is indeed stunning.

In the morning were some great "discussions" (more informal than lectures) on evolution of neural circuits, and emotions in fish.  I then had some excellent conversations with participants in the afternoon.  Then off for a stroll to photograph the sea view.

It promises to be a fantastic time: great talks, great conversations, great food, great venue, great weather!