Showing posts with label complexity theory. Show all posts
Showing posts with label complexity theory. Show all posts

Thursday, July 13, 2023

Plume Tracking and Odor Mapping Algorithms


A deep reinforcement learning model that allows AI agents to track odor plumes
Feb 2023, phys.org

Insects track odor plumes to find mates. (And there was a similar study done recently here.)

"Instead of running a traditional laboratory wind-tunnel experiment, we used a complementary 'in-silico' approach using artificial neural networks," Singh explained. "This helped us develop an integrative understanding of plume tracking across multiple levels, including emergent behavior, neural representation and neural dynamics."

To train their plume-tracking agents using DRL, the researchers first simulated an odor emanating from a source located within a windy arena with a total area of approximately 120 m2. When their agents identified where the source of the odor was located, they received a reward. In contrast, if they lost track of the odor plume and left the arena, they were "punished."

"The behavior that emerges in our trained artificial agents bears a striking resemblance to the behavior modules biologists have previously observed in flying insects performing plume tracking," Singh said.

via University of Washington and University of Nevada: Satpreet H. Singh et al, Emergent behaviour and neural dynamics in artificial agents tracking odour plumes, Nature Machine Intelligence (2023). DOI: 10.1038/s42256-022-00599-w


Saturday, May 19, 2018

Water Chemistry



Thought I could shed some light on a trendy topic here:  water chemistry and baking success.

Everyone knows Brooklyn pizza is the best pizza, and that there is only one place in the world that makes bagels, and it’s New York City. (If you’ve ever had a bagel in Florida, you know what I’m talking about; unless it was literally baked in NY and shipped to FL, which isn’t uncommon.)

We heard recently the claims of a well-meaning group of entrepreneurs that NYC’s water is now available to the rest of the world. The company's claim: They will analyze your local tap water to calibrate a device that will tweak it to be just like New York City’s water.

What does the water taste like and why? Well, it tastes good, according to taste tests. And the reason is because New York City's water comes from natural reservoirs, highly protected from contamination. This allows the water to pick up minerals in the ground that it passes as it makes its way through the water cycle. Also, the ground it passes is low in calcium, which makes things taste bitter (not better, bitter).

Chances are, however, that taste is not the reason why NYC gets so much props for its water. Chances are it's the chemistry of baking, which is a complicated thing, except to say that it's the chemicals, stupid.

The chemicals themselves are thrown into the Maillard arena to break down and transform.  Pyrrolines and pyridines are some of the bread-smelling chemicals created in this culinary laboratory – the “pyro-“ parts of their names indicate their origin in fire.

I just named two chemicals, but baked goods emanate hundreds of chemicals, all dependent not only on the original chemicals present, but on the nature and extent of the heating reactions. Just like the analogy of the butterfly effect, initial conditions can have a drastic effect on results. In a process as complex as the heated Maillard reaction, a slight change in the initial ingredients can have wide-ranging effects on the bun in the oven.

Taste/smell isn’t the only thing that is subject to change here – texture is a big factor as well, and is affected just as much based on initial ingredients.

Is it the case that the same chemicals that make your water taste good also make your bread taste good and also make you bread the perfect combination of fluffy-chewy-crunchy? Maybe not, but in the case of New York City’s water, some have it all.

Post Script
In case you thought NYC had it all, Paris just put bubbly water in it's water fountains. We call that carbonated water.

Wednesday, May 18, 2016

Brainless Intelligence


Many-headed slime mold aka Physarum polycephalum, image via the French National Centre for Scientific Research, 2016 

Some folks made slime think. The lowly slime mold, a single-celled protist, shows evidence of learning. It remembers the particular route that avoids irritants placed in its path by tinkering scientists. Yup. Funny thing is, the organism investigated is commonly called the “many-headed slime.” This turns out to be an ironic name, for this organism, without a central nervous system, acts like it does in fact have a head, or a brain, and maybe more than that – many heads, and many brains.

This isn’t the first time slime mold has done amazing feats. It’s used to recreate roadmaps from ancient cultures, or Tokyo’s rail system, just based on topographical information. Who do these single-celled organisms think they are, acting like they have brains? This raises the following question: Where does intelligence come from? Does it need a brain?

In Hidden Scents, while talking about the evolution of the smelling organism, I suggest that the mind is first, and then comes the body. There is something thinking in the most primitive of organisms, deciding which molecules in its surrounding sea of life, and proto-life, should be taken into it, to become part of it, and which molecules should stay outside. To be alive, one of the most basic requirements is to have a boundary between the living thing and the outside. This defines the body. But how does this body, living in a sea of potential bodyparts, determine which parts to keep, and which ones to leave behind. The body comes from somewhere, doesn’t it? And isn’t a body - a living body - more than just a bunch of molecules? If so, what’s organizing those molecules? Who is running the show?

Chemosensation is the basis of this interface, and is the process by which human olfaction works. The initial decision-making algorithms to run with this chemosensation are also the base-algorithms of human thought. Rational thought is a much more complex affair, but at the base is the limbic system, and in smelling we have a model for the kind of thinking performed by a simple, multicellular organism. Or even a collection of single celled organisms, perhaps?

Our current mode for thinking about intelligence is undergoing a major reboot. In light of developments in artificial intelligence, the boundaries of human intelligence are already blurred – many of the things once considered human, rational thought are now programmed into an "artificial life form," i.e., a computer program.

But that’s ok, because current models of the brain follow the schematics of a computer in the same way the nervous system was initially thought of as a closed network of fluids and the brain a pressure-modulator. This was in the age of hydraulics, before we knew what electricity was. Now we know what a computer is, and so the brain is like a computer. Tomorrow, we may know what life is; will we then compare the brain to it?

Our ideas on thinking and intelligence necessitate a brain (whether it’s a computer or a water pump or a lifeform). It's very counter intuitive to hear that things without brains can think. Who knows, next it will be like “Things without bodies can think.” Does the temperature in a room think? Does it have a memory?

Notes:
May 2016, phys.org

Laura Sanders, Wired, via Science, 2010

Mar 2015, phys.org