Modeling "pain"

So many ways to model pain. Could one use the notion of a breakdown in trust between entitities?

Update: I read some clever guys modeling pain as a difference from equilibrium.

forced completion

Suppose a new pattern emerges from an old one when in perceiving the new we decompose it into an incomplete old pattern completed in a new way. My thought experiment involves seeing a dumbbell shape as a violation of the circle - a forced completion of the circle creating the new pattern.

This should result in storing the dumbbell shape nearby the circle shape, either as a subtopic or as a brother topic to the circle in a more general category.

But, as David Levy reminds me, if there is no value attached to identifying a new pattern, it will not become learned. For example, I got lazy above and drew something that was not a dumbbell, in order to make my point. In fact it looks like a couple of bubbles inadvertently glued together - a different forced completion than the one I was trying to illustrate. In that case, we either ignore identifying a new pattern or we start duplicating the original circle.

Physical Community and Online Community

Woods Hole is a place with a strong sense of community in the winter and it has a pretty strong sense in the summer as well - when the town is packed with people from NY, Texas, Paris, etc.
Recently I was having an online exchange with someone, having trouble discussing "community" because I was talking about physical community and he was talking about online community.

Why can't online community and physical community be the same? It leads to this question: what resources could be put in front of an online community to help people in a physical community? I am doing an aggregated calendar. Other answers range from: help with home maintenance and construction to babysitting, catering, eldercare, shopping, transport, laundry, and maybe anything else that one person regards as a chore and another person is happy to offer.

The idea is taking form slowly and I am wondering about minimum useful capabilities. One idea I like is that accounts are formed around "my project", with possible "ask and bid" supported executing the project AND a passive view mode allowed where the public can watch. A kind of live TV.

In some ways, though, you want to eliminate a direct quid-pro-quo for low budget projects. For example, I would be happy to do some baby sitting, assuming that sooner or later I am going to need some eldercare.

A version of the spliting principle

From an old Italian woman in a TV show Happy: "Once you know the Truth you can't un-know it".

Feels like a million dollar idea

A couple different thought threads intersect.

I had a slightly stressful morning conversation with the civil engineer on our home addition project. He is doing a great job facilitating things like the contractor/plumber relation and he has been of immense help, knowing everyone and moving the project along. I could not get the project done without him.

Separately, I have been doing this woodsholecalendar.com and trying to get some nice young people to throw some energy into helping me expand - with better music content; better source-to-aggregate automation tools; and, who knows?

So I am walking into town, thinking about adding resources like: free-form written content or classifieds, or whatever makes a newspaper useful. But online. What kind of resources would be valuable to a community? I am remembering a most recent text message from one of these younger guys, saying he was interested in "ways people can come together and form community". So I thought: rather than pitching online newspaper ideas to the guy [who I am trying to get interested in helping me] why not reflect his idea back at him. And I message back: "Like what? Besides what Facebook does...."

Around this time, thoughts of my morning conversation with the civil engineer intrude and I start to wonder: how could you create a community resource [sub-part of the calendar] that would help people like me and my civil engineer get a project going, scheduled, and completed? And in a way that encouraged community values? In a broader context I have been exposed to the philosophical discussions of building large homes and why they should or should not be allowed in different places. Right now, Woods Hole is listed (perhaps a bit inaccurately) as one of the top ten mostly costly home locations in America. Oversized houses are a problem here and in places like Concord and Chilmark on the Vineyard [We saw a great documentary from there called "One big house"]. The interesting thing is that the dynamic driving houses to be larger is not only coming from people seeking status. It is coming from feature creep in the design and construction of the home. It is a natural dynamic when homeowners are uncertain and the designers have a conflict of interest: they desire to please and they desire to make profit.

The 'community value' precept comes both from the discussion with my young colleague and from the discussion of oversized houses. They converge on the idea of doing something

To facilitate the interaction of builders and homeowners during the project while serving community values*

Think about doing that online and add a spark: non participants watching the progress of the project, like a reality TV show. These viewers could add a lot of value back into the project.

So the sub-webpage could be called "My Home Project". You go there as a homeowner and start a project. Then builders etc go there to find work, or to track it. What is publicly viewable is a negotiation between homeowners and potential builders. Hopefully this social medium becomes a useful networking tool for the project. I am not sure what is the incentive of a builder, other than the homeowner calling the shots and trying to manage things though this website. SOUNDS TOO AMBITIOUS SO: is there a minimal version of this sort of thing that is doable?

* Beauty, energy efficiency, neighborhood character, feature integration (rather than sprawl), use of local resources

Life is Language

I had a strange thought while half asleep last night: that the nature of planning an action, inherent in any action, involves a linguistic component of the brain. Now maybe not for plants that do not act [really??]. But thinking of a wolf, smelling a sick bison bull a mile a way and heading for it in a straight line (ignoring all other bison ….you've seen the video) makes me believe that the wolf has an abstract understanding of a straight line AND some form of mental grammar allowing the straight line to be implemented as the wolf decides what to do. I cannot see how an action can exist without a plan, or how a plan could exist without a grammar for combining geometry and purpose.

An unlucky clam

You wouldn't think that a clam would have good or bad luck. I assume, for the most part, that clams are born, grow in the mud, then die. So I went clamming yesterday and one of the clams I pulled up had a limpet (a single shelled mollusk) growing across the two shells of the clam. Now, the clam has little leverage for opening itself up - as needed for its feeding and rudimentary navigational capabilities- and where the limpet crossed from one side to the other, it was probably keeping the clam shut and constricted, preventing the clam from doing much of what it needed to survive. The limpet would have killed the clam after a while.

So I pried the limpet off and was about to pocket the clam when I noticed there was also a small oyster drill (yet another type of mollusk) starting a hole through the clam's shell, in a different place. In a little while the drill would have gotten through, inverted its stomach through the hole, and started to digest the clam from the inside.

And now I am about to toss that poor clam into boiling water. I mean: how many bad things can happen to one clam? Talk about bad luck. Possibly, since it is now resting in a bucket of sea water, freed from the constrictions of the limpet, and not yet punctured by the oyster drill, it might be experiencing a brief moment of relief before the disaster that's coming tomorrow at lunch time. Sad. At least it had an exciting life.

Update: I decided to have mercy on that clam. Broke the oyster drill off and returned the clam to its home waters. The others got cooked and eaten. Whose laughing now? On the other hand, I have read that the seabirds - gulls and diving ducks - pay attention to where people are clamming, and go around afterwards to collect any clams that fell off the clam rake. Smart birds. In fact there were gulls and mergansers swimming where I was clamming the other day, so I payed close attention to where I dropped the unlucky clam back in the water. I am afraid the birds did too. I'll go look for the clam in the future, but its run of bad luck may continue inside the stomach of a bird.

AI - know it by knowing its failures

I was surprised to read an academic study showing AI did not have human-like shape recognition. They use the trope of a teapot with golf-ball texture and come to the long-since obvious conclusion that the software is prioritizing texture over shape. I first heard about this with sea turtle texture identification over a wide range of non-sea turtle shapes.

I construe Neural Nets as providing pieces-wise linear coordinate rectangle mapping which, if done at sufficiently hi resolution will be able to approximate any smooth curve. There are no explicit provisions for multi-curve problem spaces. There is no understanding of differentiability. One real problem is that the neural layers are essentially training on small groups of pixels. The larger the group the more time it takes to "learn" and one approach is to toss on new layers for larger and larger groups. But -GOD- mathematicians knew long ago that shape was a global property, and common sense says there is rarely a guarantee of "smoothness" or the sorts of convexity that are assumed in their vague theories of interpolation/ extrapolation. Trying to deduce shape from texture is complete nonsense. Trying to use raw pixels is the stupidest thing imaginable, from the point of view of position-invariant shape. What has become clear is that often the data->result mapping is discontinuous.

So I wonder a couple things, in a snide sort of way: why don't these guys go learn some math? And what about all those other AI "successes" we are hearing about? Like medical diagnostics. How do we know these aren't subject to the same errors - correlating a global outcome with an easy to compute local property that is- in fact coincidental? I see some risk ahead when no one can check if the emperor's new clothes are real - when the problem domain is harder to check than image recognition.

Went to a WHOI seminar where the man in charge of AI spending at MIT Lincoln Labs gave a superficial summary of AI that touched on its successes and the failures. Weirdly, he claimed "98% accurate" image recognition in one part of the talk, then mentioned later how easy it is to trick the system. Leading to the question: 98% of what?

Anyway, short of smart feature extraction (of global features) there will be no pixel-based image recognition.

Bagel Recipe

Adapted from: Perfect Bagels from TIQA

Makes 6 bagels. I am using King Arthur Bread Dough flour, soft water, and ~66 deg. F room temp.

Day 1: make the "pre-ferment"; let rest overnight
Day 2: make final dough by adding flour, etc to pre-ferment; let rest overnight
Day 3: form bagels, boil, then bake

Day 1 - make the pre-ferment
Mix in bowl

• 1 3/4 cups flour
• 1/2 tsp yeast (Active Dry)
• 1/2 tbs of MALT syrup (the dark brown colors the bagels, see if can find light color, or powder?)
• 1.5+ cups room temp water

Cover bowl and put in fridge overnight. 

Day 2 - make the dough

Take bowl from fridge and let approach room temperature.

Mix in:

• 1.5 cups flour
• 1/2 tsp yeast
• 1 tsp of salt [I increased this and eliminated additional sugar from TIQA recipe]

Stir in the flour and kneed in more flour as needed until the dough is not tacky and smooth.  [5-10 minutes, by hand] [Actually: kneed in more flour until you cannot kneed in any more]

Let rest a few (~15) minutes then cover bowl and put in fridge overnight

Day 3 - make the bagels

Take bowl out of fridge. 

Put large pot of water on to boil vigorously (which takes ~1/2 hour on my crappy stove top)

Preheat oven to 500 deg. F (should be ready when water is boiling)

After bowl has warmed slightly (during the 1/2 hour) and after the night resting, dough will have expanded a bit. 

Cut dough into 6 equal sized pieces, roll each in a ball, puncture, and form into bagel. (Dough will be a bit sticky, so an alternative is to form bagels the previous night, which I have not tried yet. Since the dough is getting warmer and stickier I put some dough back in fridge to keep it cool while making the rings. Don't let dough get too warm or start rising, as then bagels will flatten and not boil well.

Take baking pan, cover with baking paper and put the formed dough rings on pan, about 1 inch apart. (This is just while waiting to boil them.)

Three at a time, boil the bagels for 45 sec each side. [they may drop to bottom then float back up at first]. Take out with slotted spoon. Roll them briefly on a towel to remove extra moisture. Then put them back on the baking pan.

Bake at 475, ~8 minutes then turn pan and do another 8 - 10 minutes. Should be golden brown and not burnt on the bottom.

Notes:  I made two errors: (1) not enough flour kneaded into the dough, so no dropping to the bottom of the boiling water and a "crumb" with oversized bubbles. (2) Baked at 500 for too long: burnt the bottoms. I will be more careful with the flour and put the baking pan on the top rack - monitoring more carefully at 450-475. They are delicious anyway....but not good enough.

Google News is Un-Balanced

I wish they would "figure it out" -I am seeing one story after another about celebrities: politicians, comedians, celebrity shows, celebrity Royals, celebrity companies (Apple, Samsung, Tesla, etc...), not to mention celebrity CEOs. Then we get to scientific news and there are no celebrities, just pictures of asteroids and stories of hemorrhoids. At least since Steven Hawking is dead. Every day this week there was story about a single Netflix movie - a celebrity show.

I don't give a f*ck about your celebrities. I don't care about your iPhone updates and new OS talk. It is a whole problem with society.

My first fried clams

I spent an hour collecting at the yacht club and got ~24 clams. They are really close to the restaurant experience (no yuck factor) and delicious.

Got the recipe from YouTube: dip in cream/egg and shake well in inflated bag with flour/paprika/salt. Fry in peanut oil.

Later, with friends of Joe (Tyler and Jessie) we had a feast with more of these fried beauties.