5 ms·

It seems likely to me that “seeing things” is consistent with the predictive processing model from cognitive neuroscience. To summarize: the retina error-correc

by xpe 3y ago

It seems likely to me that “seeing things” is consistent with the predictive processing model from cognitive neuroscience. To summarize: the retina error-corrects a simulation provided from the cortex. Let’s say some visual details are fuzzy. With a strong enough Bayesian prior (e.g. belief in UFOs), the human perceptual apparatus could subjectively see things that look like UFOs. Mind bending, but this seems to me a credible and popular theory of brain functioning. I’m no expert and new to the theory. Please correct me if I’m butchering it. P.S. Like anyone, I often detect motion in my peripheral vision without much detail. Maybe once a year (or less), I see weird things that seem implausible, but only for brief instant. When I look, there is nothing unusual. Yes, bring on the “what are you smoking?” comments… but I assure you the answer is “nothing”.

gitfan86 3y ago
Yes, that seems right. Humans fill in the gaps. AI is similar, where it "hallucinates", but are humans that much different?
friendzis 3y ago
While this is not a well studied phenomenon, as in to the extent of having well formed theory, we do know that eyes constantly perform micromovements. One of the prevailing hypotheses states that due to low inherent resolution of the retina these micromovements are used to create sort of sub-pixel resolution. Hence, visual cortex inherently works in predictive/generative mode with constant feedback correction. This hypothesis is sometimes used to explain phenomena like pareidolia and might as well be used to explain perception of UFOs.
Loquebantur 3y ago
You are hallucinating complete nonsense here. Just think for a second, if such a thing was possible, why would it be restricted to "UFOs"? Obviously, such an illness would be well known.
- George83728 3y ago
  > if such a thing was possible, why would it be restricted to "UFOs"? Who said it is? Tons of people report seeing all manner of nonsense.
  - Loquebantur 3y ago
    The prevalence of mental illnesses leading to visual hallucination is very low actually. The idea of whole departments in the US government being staffed with them is absurdist nonsense.
    - gitfan86 3y ago
      "normally functioning" brains see things that are not actually there ALL the time. https://en.wikipedia.org/wiki/Optical_illusion https://en.wikipedia.org/wiki/Optical_illusion Most UFO reports fall into the Optical Illusion category rather than extreme schizophrenia.
      - Loquebantur 3y ago
        Optical illusions are not at all the same thing as visual hallucinations. Your allegation about UFO reports is accordingly simply incorrect.
        gitfan86 3y ago
        Right, so your point about people not having hallucinations frequently is valid. But most people do see optical illusions. So it seems strange that you have jumped to the conclusion that these must be real UFOs since most people do not hallucinate.
        Loquebantur 3y ago
        I did not jump to that conclusion, you are putting words in my mouth. The whistleblower reports of the existence of a research program that he was briefed about. He testifies this program was stuying non-human technology, among that functioning craft, aka "flying saucers". The determination of it being non-human in orgin was due to extensive material analysis, evident morphology and working characteistics of those devices.
- xpe 3y ago
  First, your tone comes across as unnecessarily uncivil. Your comments could be rephrased to better promote curious discussion. Second, I will respond in terms of the charitable substance of your comment. I think there is some useful back-and-forth to be had. Please be specific: part is “complete nonsense” to you; which of the following applies? 1. You reject the predictive processing theory? Why? 2. You reject my characterization of the theory? 3. You reject my application of the theory to the example of seeing UFOs? For the reason you gave? Is there more? 4. What research have you done to support your knowledge and claims? 5. Have you read “Hallucinations and Strong Priors” in Trends in Cognitive Science [1]? I referenced it in nearby comment as well. Penultimate point: correctly or incorrectly, I am using facts and reasoning in my above comment; I don’t see how this qualifies as “hallucination” in terms of an AI or human phenomenon. Final point: there is considerable nuance and skill required to criticize effectively. In asking questions 1 through 3, I’m offering an example of a better way. [1] https://doi.org/10.1016/j.tics.2018.12.001 https://doi.org/10.1016/j.tics.2018.12.001
  - Loquebantur 3y ago
    Curious discussion is engaging on the topic at hand, not on completely unrelated nonsense, that you are no expert on anyway (as you professed yourself). As already stated, the whistleblower does not fall at all in the context you are discussing here, as he reports having been briefed on research programs that study non-human technology. Now, in the case of those sightings by military pilots, you have multiple sensor systems and multiple pilots involved simultaneously. Clearly impossible to be explained by your idea here. But even in cases of single eye-witnesses, their sightings are often over extended periods of time, entirely incompatible with your paper.
    - xpe 3y ago
      > Curious discussion is engaging on the topic at hand, not on completely unrelated nonsense… 1. No. This interpretation of one HN Guideline is not accurate. Such a view would justify being non-charitable, which is incompatible with the guidelines more broadly. Curious and charitable conversation should not stop because you think my comment is off-topic. 2. Regarding topicality; as I explained at length in another comment, as conversations evolve in a tree of comments, there may be considerable distance (in terms of a topic vector space) from the root. This is a normal and useful. I’m certainly open and appreciative of someone adding to the conversation. You made a connection from my points to what you see is the central point. Your connection was ‘negative’ in the sense that you suggested my points about perceptions and priors does not apply to the original post. This sense of negative is not a problem, even though I disagree with some of your logic. But the way you did it — quite harshly — was unnecessary. 3. Each time you say “utter nonsense” or “bogus” (and the like) it comes across as insulting. Are you aware of this? There are clearer phrases and terms to use that are not insulting. Depending on what you are trying to say, you might try: - you could say you disagree (in the sense of a value judgment or a subjective matter) - you could say that I’m making a logical or rational error - you could say that something is not true, factually - you could say that particular claim of mine is inconsistent with another claim - you could say that I am not providing evidence or support for a claim Onto a different point. You wrote: > not on completely unrelated nonsense, that you are no expert on anyway (as you professed yourself). 4. Stating that I am not an expert in a particular cognitive science theory does not inherently disqualify the logic of my argument. It merely suggests an openness to learning if there are mistakes. Such an openness is indicative of humility and a tendency to not overreach. As you can see in my other comments, I have recognized and agreed with some of your points. But these points have not contradicted or disproved what I wrote.
xpe 3y ago
> Recent empirical work from independent laboratories shows strong, overly precise priors can engender hallucinations in healthy subjects and that individuals who hallucinate in the real world are more susceptible to these laboratory phenomena. https://www.cell.com/trends/cognitive-sciences/fulltext/S1364-6613(18)30282-1 https://www.cell.com/trends/cognitive-sciences/fulltext/S136... Trends in Cognitive Science OPINION | VOLUME 23, ISSUE 2, P114-127, FEBRUARY 2019 Hallucinations and Strong Priors Philip R. Corlett, Guillermo Horga, Paul C. Fletcher, Ben Alderson-Day, Katharina Schmack, Albert R. Powers III Open Access Published: December 21, 2018 DOI: https://doi.org/10.1016/j.tics.2018.12.001 https://doi.org/10.1016/j.tics.2018.12.001
- Loquebantur 3y ago
  First of, this is completely off-topic, since the whistleblower reports being briefed on non-human technology in possession of the US, which is being intensively studied. Further, their related "sightings" are corroborated by multiple sensor systems simultaneously. And multiple pilots as well. Now, your idea of hallucinations being an explanation for such sightings, obviously impossible due to said sensors systems in the first place, is utter bogus since not only do you not have "strong priors" in such a situation, you also do not employ people suffering from hallucinations as pilots for mulit-million dollar fighter jets to begin with. Your article finally describes an entirely different phenomenon than "seeing an object consistently for an extended period of time". Which is what is relevant here.
  - xpe 3y ago
    You are not the final arbiter of what is on topic. Topics are subjective and connected cognitively differently by different people. Also, in my view, my comment is directly on topic to the message it replies to. [1] I’m curious about your user experience with Hacker News. Do you use the normal web interface or some other UI? I’m sorry to ask the obvious question, but I do want to rule it out: do you realize that comments are situated relative to the parent comment, right? Now, after factoring this in, do you see how my comment is a response to its parent? If you’d like to think about it mathematically, try this. The original post could be characterized as having a topic vector in a many dimensional space. Each subsequent reply can be characterized similarly. As you get deeper in the conversation tree, the topic vectors diverge significantly. This does not mean a somewhat distant (in topic vector space) comment is “off-topic”. Quite to the contrary, it means the original post has generated a rich variety of discussion. This said, I take your point that transitory perceptual errors do not explain all UFO sightings, nor do they directly bear on the whistleblower’s main claims. One can make such a point without being prickly about it. I’m sorry that you seem upset here and in your comments generally across the threads here. I can relate to some degree; there is is a lot to be concerned about when it comes to the topics of evidence and people’s ability to evaluate truth claims. I can also relate to prioritizing factual assessments at the expense of tact and empathy towards others. For many years of my life I did this, and it did not serve me well. It is your responsibility to not take out on your frustrations on me or anyone else. Treat us with respect or your reputation may suffer. If you manage a civil tone, I will engage with you further. Otherwise, I don’t see it being productive enough to warrant the effort. To be specific, I suggest reviewing the HN Guideliness. In particular, strive to ask charitable clarification questions and avoid emotionally charged claims; e.g. that what I’m saying is bogus. Lastly, you are making many assumptions which you then use to knock down straw men arguments. Notes [1] I say this based on a larger number of upvotes than I would expect on a comment this deep in the tree. This information is asymmetric; I recognize that you do not have access to it. I wonder what your assessment of this will be. - If you think I’m lying, this would confirm what you want to believe about my comment being off topic and bogus. - If you think I’m telling the truth, it would challenge your claims.
xpe 3y ago
> I often detect motion in my peripheral vision without much detail. Maybe once a year (or less), I see weird things that seem implausible, but only for brief instant. When I look, there is nothing unusual. I want to elaborate on one personal experience I alluded to above. I am curious if others have experienced anything similar. While driving at about 20 mph across some train tracks, I noticed in my left peripheral vision a parked vehicle with its rear door open. For a tiny instant, I perceived it roughly like dirt was “flowing out” of the car. When I looked more closely — and again, this happened in a fraction of a second — it was clear that a dog, probably a golden retriever, was jumping out. So my first fleeting perceptual experience got the motion roughly correct (movement out of a back seat onto the ground) and color (light brown) correct. It got the object wrong. From what I can tell, this is a consistent subjective experience with the predictive processing model. It seems to me if one did not subscribe to the predictive processing model, one would expect the brain to only register a vague blur. I think it is also worth mentioning that in the days leading up to that experience I had a DIY project where I dug a deep hole, and I put a bucket in the bottom which I filled before lifting it out and dumping it. (This was more efficient than doing one shovel load at a time.) This digging was not an expedient process which could have led my brain to spend energy processing better ways to do it.

by xpe 3y ago

gitfan86 3y ago

Yes, that seems right. Humans fill in the gaps. AI is similar, where it "hallucinates", but are humans that much different?

friendzis 3y ago

While this is not a well studied phenomenon, as in to the extent of having well formed theory, we do know that eyes constantly perform micromovements. One of the prevailing hypotheses states that due to low inherent resolution of the retina these micromovements are used to create sort of sub-pixel resolution. Hence, visual cortex inherently works in predictive/generative mode with constant feedback correction. This hypothesis is sometimes used to explain phenomena like pareidolia and might as well be used to explain perception of UFOs.

Loquebantur 3y ago

You are hallucinating complete nonsense here. Just think for a second, if such a thing was possible, why would it be restricted to "UFOs"? Obviously, such an illness would be well known.

George83728 3y ago
> if such a thing was possible, why would it be restricted to "UFOs"? Who said it is? Tons of people report seeing all manner of nonsense.
- Loquebantur 3y ago
  The prevalence of mental illnesses leading to visual hallucination is very low actually. The idea of whole departments in the US government being staffed with them is absurdist nonsense.
  - gitfan86 3y ago
    "normally functioning" brains see things that are not actually there ALL the time. https://en.wikipedia.org/wiki/Optical_illusion https://en.wikipedia.org/wiki/Optical_illusion Most UFO reports fall into the Optical Illusion category rather than extreme schizophrenia.
    - Loquebantur 3y ago
      Optical illusions are not at all the same thing as visual hallucinations. Your allegation about UFO reports is accordingly simply incorrect.
      - gitfan86 3y ago
        Right, so your point about people not having hallucinations frequently is valid. But most people do see optical illusions. So it seems strange that you have jumped to the conclusion that these must be real UFOs since most people do not hallucinate.
        Loquebantur 3y ago
        I did not jump to that conclusion, you are putting words in my mouth. The whistleblower reports of the existence of a research program that he was briefed about. He testifies this program was stuying non-human technology, among that functioning craft, aka "flying saucers". The determination of it being non-human in orgin was due to extensive material analysis, evident morphology and working characteistics of those devices.
xpe 3y ago
First, your tone comes across as unnecessarily uncivil. Your comments could be rephrased to better promote curious discussion. Second, I will respond in terms of the charitable substance of your comment. I think there is some useful back-and-forth to be had. Please be specific: part is “complete nonsense” to you; which of the following applies? 1. You reject the predictive processing theory? Why? 2. You reject my characterization of the theory? 3. You reject my application of the theory to the example of seeing UFOs? For the reason you gave? Is there more? 4. What research have you done to support your knowledge and claims? 5. Have you read “Hallucinations and Strong Priors” in Trends in Cognitive Science [1]? I referenced it in nearby comment as well. Penultimate point: correctly or incorrectly, I am using facts and reasoning in my above comment; I don’t see how this qualifies as “hallucination” in terms of an AI or human phenomenon. Final point: there is considerable nuance and skill required to criticize effectively. In asking questions 1 through 3, I’m offering an example of a better way. [1] https://doi.org/10.1016/j.tics.2018.12.001 https://doi.org/10.1016/j.tics.2018.12.001
- Loquebantur 3y ago
  Curious discussion is engaging on the topic at hand, not on completely unrelated nonsense, that you are no expert on anyway (as you professed yourself). As already stated, the whistleblower does not fall at all in the context you are discussing here, as he reports having been briefed on research programs that study non-human technology. Now, in the case of those sightings by military pilots, you have multiple sensor systems and multiple pilots involved simultaneously. Clearly impossible to be explained by your idea here. But even in cases of single eye-witnesses, their sightings are often over extended periods of time, entirely incompatible with your paper.
  - xpe 3y ago
    > Curious discussion is engaging on the topic at hand, not on completely unrelated nonsense… 1. No. This interpretation of one HN Guideline is not accurate. Such a view would justify being non-charitable, which is incompatible with the guidelines more broadly. Curious and charitable conversation should not stop because you think my comment is off-topic. 2. Regarding topicality; as I explained at length in another comment, as conversations evolve in a tree of comments, there may be considerable distance (in terms of a topic vector space) from the root. This is a normal and useful. I’m certainly open and appreciative of someone adding to the conversation. You made a connection from my points to what you see is the central point. Your connection was ‘negative’ in the sense that you suggested my points about perceptions and priors does not apply to the original post. This sense of negative is not a problem, even though I disagree with some of your logic. But the way you did it — quite harshly — was unnecessary. 3. Each time you say “utter nonsense” or “bogus” (and the like) it comes across as insulting. Are you aware of this? There are clearer phrases and terms to use that are not insulting. Depending on what you are trying to say, you might try: - you could say you disagree (in the sense of a value judgment or a subjective matter) - you could say that I’m making a logical or rational error - you could say that something is not true, factually - you could say that particular claim of mine is inconsistent with another claim - you could say that I am not providing evidence or support for a claim Onto a different point. You wrote: > not on completely unrelated nonsense, that you are no expert on anyway (as you professed yourself). 4. Stating that I am not an expert in a particular cognitive science theory does not inherently disqualify the logic of my argument. It merely suggests an openness to learning if there are mistakes. Such an openness is indicative of humility and a tendency to not overreach. As you can see in my other comments, I have recognized and agreed with some of your points. But these points have not contradicted or disproved what I wrote.

xpe 3y ago

> Recent empirical work from independent laboratories shows strong, overly precise priors can engender hallucinations in healthy subjects and that individuals who hallucinate in the real world are more susceptible to these laboratory phenomena. https://www.cell.com/trends/cognitive-sciences/fulltext/S1364-6613(18)30282-1 https://www.cell.com/trends/cognitive-sciences/fulltext/S136... Trends in Cognitive Science OPINION | VOLUME 23, ISSUE 2, P114-127, FEBRUARY 2019 Hallucinations and Strong Priors Philip R. Corlett, Guillermo Horga, Paul C. Fletcher, Ben Alderson-Day, Katharina Schmack, Albert R. Powers III Open Access Published: December 21, 2018 DOI: https://doi.org/10.1016/j.tics.2018.12.001 https://doi.org/10.1016/j.tics.2018.12.001

Loquebantur 3y ago
First of, this is completely off-topic, since the whistleblower reports being briefed on non-human technology in possession of the US, which is being intensively studied. Further, their related "sightings" are corroborated by multiple sensor systems simultaneously. And multiple pilots as well. Now, your idea of hallucinations being an explanation for such sightings, obviously impossible due to said sensors systems in the first place, is utter bogus since not only do you not have "strong priors" in such a situation, you also do not employ people suffering from hallucinations as pilots for mulit-million dollar fighter jets to begin with. Your article finally describes an entirely different phenomenon than "seeing an object consistently for an extended period of time". Which is what is relevant here.
- xpe 3y ago
  You are not the final arbiter of what is on topic. Topics are subjective and connected cognitively differently by different people. Also, in my view, my comment is directly on topic to the message it replies to. [1] I’m curious about your user experience with Hacker News. Do you use the normal web interface or some other UI? I’m sorry to ask the obvious question, but I do want to rule it out: do you realize that comments are situated relative to the parent comment, right? Now, after factoring this in, do you see how my comment is a response to its parent? If you’d like to think about it mathematically, try this. The original post could be characterized as having a topic vector in a many dimensional space. Each subsequent reply can be characterized similarly. As you get deeper in the conversation tree, the topic vectors diverge significantly. This does not mean a somewhat distant (in topic vector space) comment is “off-topic”. Quite to the contrary, it means the original post has generated a rich variety of discussion. This said, I take your point that transitory perceptual errors do not explain all UFO sightings, nor do they directly bear on the whistleblower’s main claims. One can make such a point without being prickly about it. I’m sorry that you seem upset here and in your comments generally across the threads here. I can relate to some degree; there is is a lot to be concerned about when it comes to the topics of evidence and people’s ability to evaluate truth claims. I can also relate to prioritizing factual assessments at the expense of tact and empathy towards others. For many years of my life I did this, and it did not serve me well. It is your responsibility to not take out on your frustrations on me or anyone else. Treat us with respect or your reputation may suffer. If you manage a civil tone, I will engage with you further. Otherwise, I don’t see it being productive enough to warrant the effort. To be specific, I suggest reviewing the HN Guideliness. In particular, strive to ask charitable clarification questions and avoid emotionally charged claims; e.g. that what I’m saying is bogus. Lastly, you are making many assumptions which you then use to knock down straw men arguments. Notes [1] I say this based on a larger number of upvotes than I would expect on a comment this deep in the tree. This information is asymmetric; I recognize that you do not have access to it. I wonder what your assessment of this will be. - If you think I’m lying, this would confirm what you want to believe about my comment being off topic and bogus. - If you think I’m telling the truth, it would challenge your claims.

xpe 3y ago

> I often detect motion in my peripheral vision without much detail. Maybe once a year (or less), I see weird things that seem implausible, but only for brief instant. When I look, there is nothing unusual. I want to elaborate on one personal experience I alluded to above. I am curious if others have experienced anything similar. While driving at about 20 mph across some train tracks, I noticed in my left peripheral vision a parked vehicle with its rear door open. For a tiny instant, I perceived it roughly like dirt was “flowing out” of the car. When I looked more closely — and again, this happened in a fraction of a second — it was clear that a dog, probably a golden retriever, was jumping out. So my first fleeting perceptual experience got the motion roughly correct (movement out of a back seat onto the ground) and color (light brown) correct. It got the object wrong. From what I can tell, this is a consistent subjective experience with the predictive processing model. It seems to me if one did not subscribe to the predictive processing model, one would expect the brain to only register a vague blur. I think it is also worth mentioning that in the days leading up to that experience I had a DIY project where I dug a deep hole, and I put a bucket in the bottom which I filled before lifting it out and dumping it. (This was more efficient than doing one shovel load at a time.) This digging was not an expedient process which could have led my brain to spend energy processing better ways to do it.