Can you hear this gif?
Minimizing trust in perception
@ertwroImages referenced or modified from Google's labeling for reuse.
Based on Dr. Matthew green's blog[1] and Dr. Elliot D. Freeman's[2] research
| Highlights: 1 → Visually Evoked Auditory Reponse: Our senses work together, in a multimodal fashion. Unreliable yet useful. 2 → Alterations in Perception: When the input of that stimuli is not clear artifacts appear 3 → Reflexes: There are different stages of feedback for such stimuli, all of them automated by optimization 4 → Zero knowledge proofs: A mathematical bypass for the implications of proving to someone who sees the world in a different way that something is true 5 → Conclussion |
|
n the opening gif I asked you if you could hear that gif. Around 70% people in a twitter survey responded that they can. This was made famous by Dr Lisa DeBruine who twitted about it.
This evoked response is a mild form of synesthesia. First described in 2008.[3] Not properly it but a reflex form of it. This is what is called a Visually-Evoked Auditory Response (vEAR). Are these people just agreeing on it or are they really hearing this? There are neurologic reasons for believing they actually do.[4]
Neurologic feedback loops compete for a definite answer to multimodal sensory inputs. One everyday example of this is the McGurk effect, where the same piece of information is affected by the auditory stimuli being combined with a similar competing signal.[5]
What changes is the visual input. From Ba, to Pa, to Fa
It has some interesting consequences. E.g. Desynchronizing audio from video as a learning tool. For instance, we know that delay in auditory stimuli helps improve comprehension. Useful in the case of people learning a new language.[6]
We are sure this happens due to Electrophysiologic recordings of Visually evoked potentials. Evoked potentials are monitored the objective response of an individual to sensory stimuli. Like is the case for hearing loss or multiple sclerosis(Auditory Brainstem Response) and alteration in the visual pathway (Visually evoked potentials).
Such tests would allow finding that a person indeed is observing such a thing. Although useful for us it would not help the patient on its own.
Alterations in perception are far more common than people would normally realize.
|
Suppose for a moment you are in the market buying fruits. You smell apples, turn and see a group of them. they look great so you pick a couple for an apple pie. You head to the counter and as you try to pay, you notice there's something weird about the price.
It's actually more expensive than you thought. The seller kindly tells you - there's no mistake, you have green and red apples. Their price is not the same - which surprises you as don't understand what the person in front of you is telling you. All the apples have the same color. You checked back when you took them and they are in front of your own eyes! You argue and people around you think you are crazy as they can perfectly see the difference. What's going on? A prank? At home, you tell the story just to find everybody confirms that, in fact, the apples are not the same color. Are they deceiving you? How can you trust them? So you go and check the internet, learn about color blindness and check a doctor, Everything confirms it, but can you really be sure? |
yschromatopsia a.k.a. Color blindness is a deficiency in color vision. It can be acquired due to trauma or diseases[7] that affect the retina (e,g. glaucoma) yet most of the time is hereditary.[8]
The problem is due to an alteration -most of the time altered distribution in the fovea- in the color sensing cells in your retina, photoreceptors known as cones. 3 of them for short(B), medium(G) and long(R) wavelengths. You most probably can see some colors unless you have monochromacy, in which case you have no functional cones at all.
An important distinction we must make is that the most common form is anomalous trichromacy - 1 in 20 people - In this condition, people have the 3 kinds of cones but the react to a different wavelength than they would normally do. Confusing red and green.
In deuteranomaly, green colors react to wavelengths that are too high giving rise to a light brown color when stimulated by red light. In protanomaly, red cones detect wavelengths that are too low and give rise to a brown color in response to green light.
The regular test for this condition is known as the Ishihara test. A group of pseudo-isochromatic plates that have an image that would be hidden depending on the type of color alteration a person has.
Normally is performed under conditions where there's high light with high CRI, at least >90% and 5600k for color temperature, so normal LEDs don't work at the moment.
Those light properties I mentioned are fairly known for people that record video. As light rendition through a lens changes depending on the quality of the light. Just like is a problem it can also be used as a tool.
There are glasses that block portions of the spectrum that creates confusion, allowing for an easier identification of differences in color.
This only works for people that have the 3 receptors but is an overall improvement. The brain can even adapt to this changed stimuli and learn to process even better color. Not curing the alteration but giving improving the tool.
This shows that confusion can be filtered and one interesting take is that you can prove the existence of differences at the sensory level.
or sensory alterations, we use evoked potentials. A form of electrophysiologic recordings. We present a stimulus and then see if the patient's body responds to the stimulus. We use it to determine if, for instance, a child is deaf or blind as a baby. In order to correct the alteration before the opportunity windows for learning in the brain closes.
Most responses are five-steps systems:
When the processing is fast and without loops, we call this a reflex arch. It's particularly fast in the case of responses to pain as some of the processing is done at the spine level without going to the brain, saving vital milliseconds.
Conditional responses are adpatative reflexes. Being the pavlovian response to a skiner box the best known. Is a mechanism for saving energy in response to stimuli.
Slower responses are a product of active learning. This is a change in behavior based on previous similar experiences. The main difference with a reflex is the complexity of the processing. So a learned experience needs the same mechanism as pain but slower. The experience must not activate a reflex in order to get to a relay center in the brain. If the signal arrives at a decision-making center in the brain where it can loop new experiences with feedback, learning occurs.
It has been shown[n], that memories thought to be lost can be restored. As apparently the pathway to access them is the thing that gets lost due to protein accumulation.
That stress (e.g. pain, anxiousness, time constraints) fix memories better in the long term with reduced exposure. The problem with such exposures is that they "fix" in memory whatever occurred. If what a person experienced was lack of mastery, that's what gets imprinted. On the other hand, low-stress environments (no time constraints, depression, pleasure, psychoactive drugs) allow for quick exploration of stimuli and removal of fixed old pathways.
Intermittent cycles are a must if learning is bound to occur. By removing obstacles from personal paths.
|
This is important in the sense of discussing from first principles. Even when someone experiences something that goes against their core beliefs if done in a way that allows for safe experimentation of decision making. Without engaging the person through interaction and thought experiments no learning and desensitization can occur. It must be gradual in order to build growing coordination.
s like a magic trick in a theater. How do you know the magician and the person selected at random from the public are not colluding, that the person in the public is not actually a rival magician trying to steal the trick/ruin it in front of everyone else, that there's really a trick in the first place?
From an observer's perspective, how could you trust the prover and the verifier?
This is something that mathematicians solved back in 1985,[y] in the form of interactive proof systems with zero-knowledge. Later to advance in the form of non-interactive zero-knowledge; 3-coloring; Schnorr, Groth-Sahai, and zk-SNARKs.
Zero-knowledge proofs are methods where one party can prove another interested actor that one statement is true without conveying different information to the fact the statement is true.
Zero-Knowledge proofs:
in Goldwasser, Micali and Rackoff's paper.
One of the classic explanations on this subject is the 3 coloring scheme. You can find an in-depth intuitive review by Dr. Matthew green, but roughly speaking:
Suppose you hire google to solve the distribution of bandwidth in 3 channels for a huge cellphone network so that no two adjacent colors the same.
How can you be sure that they have computed the problem without them revealing the full answer so you can check? The original solution involved a type of closed room. Where google would the solution covered by hats. Then you would enter there and uncover at random 2 hats. If the adjacent colors are equal you would know they are lying. If they are different you would gain a little more trust in them. This process would repeat until you, the verifier, are satisfied (probabilistically speaking).
Although the context is different and an analogy might be an overstretch. When scientists teach non-experts that are skeptics or in denial of a subject, where the body of evidence is sound but is against the core of the verifier's interests, what the scientist hopes to perform is a zero-knowledge proof.
Normally in a zero-knowledge proof, the prover has a bigger computational capacity and preparation than the verifier.
He can't teach in minutes the years of evidence and expertise in order for the person to understand or accept the proofs. Yet he can share portions of the information that he wouldn't be able to obtain otherwise by faking it. 1 If the scientist is honest (appeals to nothing except evidence) he will eventually convince them 2 He can only convince them if the science is true 3 the person accepts the information as true regardless of how it affects their self-interests.
The concept is so anti-intuitive that cryptographers struggle to find a good analogy.
Yet that's maybe the most telling aspect of this. Science is anti-intuitive. If someone lacks the ability to determine if something is true or not, there are still ways to approach this problem. Avoiding the pitfalls of triggering erroneous impulses or activating non-rational responses.
here are many people that have strong opinions against choices based on science. Many stimuli trigger similar activation pathways, confounding the subject but he doesn't understand the confusion.
Identifying the mixed signals and cleaning the channels of interaction is the first step.
This is like an immune system response, but at the behavior, level rooted in the nervous system. Just like in the immune system desensitization is the first step to allow continuous exposure to a new element.
We sometimes create more harm than good by engaging in conversations where computation from the verifier is limited. This is disingenuous and creates more harm in the long run.
This desensitization only requires the verifier to bypass their senses and obtain the proof that what is being shared is true. Any further understanding will require resources from the verifier that now has a party he can trust a little more. Giving them the gift of experiencing reality in a little more vivid way.
REFERENCES:
[1] Zero Knowledge Proofs: An illustrated primer. Matthew Green in fundamentals November 27, 2014. e-Pub
[7]Hart, W. M., Jr. (1987). Acquired dyschromatopsias. Survey of Ophthalmology, 32(1), 10–31.
[8]Gordon, N. (1998). Colour blindness. Public Health, 112(2), 81–84.
