munawar@vaxb.acs.unt.edu (03/25/91)
Hi, If this is not the right bb to post the below mentioned question forgive me. I was wondering whether dreams are our brain's method of garbage collection(like done with computer memory) Has anybody done any research in to it? -shri
schraudo@beowulf.ucsd.edu (Nici Schraudolph) (03/27/91)
munawar@vaxb.acs.unt.edu writes: >I was wondering whether dreams are our brain's method of >garbage collection(like done with computer memory) >Has anybody done any research in to it? Crick & Mitchison have speculated that the function of dream sleep might be to remove parasitic modes -- "false memories" that occur in some neural net architectures under heavy storage loads. The idea is that in REM sleep the brain "runs" in reverse, _unlearning_ from experience. Since there's no sensory input, experiences in REM sleep (ie. dreams) are dominated by the parasitic modes. This unlearning is a rough connectionist analog to garbage collection in symbolic computation, and is used in at least one neural net learning algorithm (Boltzmann Machine). Should it bear even the slightest resemblance to what happens in humans, dream content would offer fascinating insights into the cogs & gears of our cognitive machinery. Question: has this hypothesis by C&M had any impact on dream research? Have there been any developments in this area since they made it eight years ago? References: Crick & Mitchison (1983), The Function of Dream Sleep, Nature 304, 111-114. Hinton & Sejnowski (1986), Learning and Relearning in Boltzmann Machines, in: Rumelhart & McClelland (eds.), Parallel Distributed Processing, Vol. 1, 282-317, MIT Press. -- Nicol N. Schraudolph, CSE Dept. | "I don't know about your dreams, but mine Univ. of California, San Diego | are sort of hackney: same thing night after La Jolla, CA 92093-0114, U.S.A. | night, just this repetitive. And the color nici%cs@ucsd.{edu,bitnet,uucp} | is really bad..." - Laurie Anderson.
ttoupin@diana.cair.du.edu (Tory Toupin) (03/28/91)
In article <schraudo.670039946@beowulf> schraudo@beowulf.ucsd.edu (Nici Schraudolph) writes: >munawar@vaxb.acs.unt.edu writes: > >>I was wondering whether dreams are our brain's method of >>garbage collection(like done with computer memory) >>Has anybody done any research in to it? > >Crick & Mitchison have speculated that the function of dream sleep might be >to remove parasitic modes -- "false memories" that occur in some neural net >architectures under heavy storage loads. The idea is that in REM sleep the >brain "runs" in reverse, _unlearning_ from experience. Since there's no >sensory input, experiences in REM sleep (ie. dreams) are dominated by the >parasitic modes. How about this: the brain is attempting to piece together memories to for temporary concepts -- a form of forethought? That is to say, memories are symbols of the state of the body when the memory was formed, and the brain tends to superimpose(?) these symbols (I suppose it chooses symbols to put together by a correlation of states -- a similar "smell", a similar "image") and see if they are physically possible situations based on the knowledge of situations which can/have occurred by trying these memories out on the "virtual body" (i.e.: the so-called mind's eye, but the entire body as well) and if it is not physically possible, it mutates the symbols slightly so that they are not associated as well with one another. [If this is not clear, it is because I have only begun to write this stuff down. If anyone cares to comment (critically, abusively, whatever :), it would be most apreciated.] -- Tory S. Toupin | ttoupin@diana.cair.du.edu | Existence toward perfection... Unversity of Denver | Life of mediocrity. Undergraduate: Math & Computer Sciences| Denver, CO 80208 | -Tory Toupin ----- C'est ne pas un fichier de <<.signature>>
prune@athena.mit.edu (Paul Berland) (03/30/91)
In article <1991Mar27.162850.4397@mercury.cair.du.edu>, ttoupin@diana.cair.du.edu (Tory Toupin) writes: |> In article <schraudo.670039946@beowulf> schraudo@beowulf.ucsd.edu (Nici Schraudolph) writes: |> >munawar@vaxb.acs.unt.edu writes: |> > |> >>I was wondering whether dreams are our brain's method of |> >>garbage collection(like done with computer memory) |> >>Has anybody done any research in to it? There have been many theories about the role of dreaming but none have been scientifically verified yet. Your analogy might be useful at some stage of the investigation, but it can only hope to describe part of what dreaming does. There is much of dreaming which is a creative process just as in the awake state. In this creative process, new patterns and structures are formed and stored; something which falls outside of garbage-collection. Think of all the inspirations people have when dreaming. |> >Crick & Mitchison have speculated that the function of dream sleep might be |> >to remove parasitic modes -- "false memories" that occur in some neural net |> >architectures under heavy storage loads. The idea is that in REM sleep the |> >brain "runs" in reverse, _unlearning_ from experience. Since there's no |> >sensory input, experiences in REM sleep (ie. dreams) are dominated by the |> >parasitic modes. There is some evidence that would make us think that things resembling brains use processes resembling dreaming for garbage-collection, but still we are dealing with analogies. Useful? Maybe. |> How about this: the brain is attempting to piece together memories to form |> temporary concepts -- a form of forethought? That is to say, memories are |> symbols of the state of the body when the memory was formed, and the brain |> tends to superimpose(?) these symbols...and see if they are physically |> possible situations based on the knowledge of situations which can/have |> occurred by trying these memories out on the "virtual body" (i.e.: the |> so-called mind's eye, but the entire body as well) and if it is not |> physically possible, it mutates the symbols slightly so that |> they are not associated as well with one another. This idea I find less plausible than the garbage-collection one, even as a sub-process. Dreams work with symbols not physical objects and situations. Dreams progress using logic much different than the rules of our world so that the "physically possible situations" cannot be determined in the dream state. Many physically impossible situations occur routinely without benefit of critical analysis, which is performed by a part of the mind which is usually asleep. This idea would be more plausible to me if it described not the test of physical processes but the test of symbolic forms. That is a much more common theory in dream psychology.
ttoupin@diana.cair.du.edu (Tory Toupin) (03/31/91)
In article <1991Mar30.023454.12481@athena.mit.edu> prune@athena.mit.edu (Paul Berland) writes: >In article <1991Mar27.162850.4397@mercury.cair.du.edu>, ttoupin@diana.cair.du.edu (Tory Toupin) writes: >|> How about this: the brain is attempting to piece together memories to form >|> temporary concepts -- a form of forethought? That is to say, memories are >|> symbols of the state of the body when the memory was formed, and the brain >|> tends to superimpose(?) these symbols...and see if they are physically >|> possible situations based on the knowledge of situations which can/have >|> occurred by trying these memories out on the "virtual body" (i.e.: the >|> so-called mind's eye, but the entire body as well) and if it is not >|> physically possible, it mutates the symbols slightly so that >|> they are not associated as well with one another. > >This idea I find less plausible than the garbage-collection one, even as a >sub-process. Dreams work with symbols not physical objects and situations. >Dreams progress using logic much different than the rules of our world so >that the "physically possible situations" cannot be determined in the dream >state. Many physically impossible situations occur routinely without >benefit of critical analysis, which is performed by a part of the mind >which is usually asleep. This idea would be more plausible to me if it >described not the test of physical processes but the test of symbolic >forms. That is a much more common theory in dream psychology. I did not mention PHYSICAL objects, but, in fact, memories: "...memories are symbols of the state of the body when the memory was formed [the experience was 'had']..." Now, as I stated above, in the process of dreaming, these symbols are put together (based on some "strong" association--be it smell, image, temperature, or other) to form a temporary symbol which is applied to the virtual body. What is the virtual body? Perhaps one way to look at it is as a "device" in the body which has the duty of taking stimuli and "return- ing" feedback from that stimuli--just like the real body, but no REAL action is performed. For example: visualizing a landscape. When one remembers a land- scape, the symbol for that landscape is applied to the virtual body--the feed- back that is given in reply to this symbol (stimuli) is, though not exactly the same, somewhat similar to the feedback that was stored when the landscape was first experienced (Is this making any sense? If not, I'll try another example). Certainly, the virtual body is quite closely associated with the real body since it is able to return the feedback that the real body would have had it received the symbol. Thus, the virtual body should be able to recognize when a symbol it has been given is physically possible ("virtually possible" is perhaps a better way to phrase it) or physically not possible in the sense that, if the body has never experienced the symbol before (say, sustained flight without aid) it must not be possible. If the symbols that were superimposed to produce this impossible temporary symbol, then the symbols are forced to diverge so that they are not associated with one another in the future (This is, I'd guess, a form of garbage collection). However, that temporary associated is STILL a valid memory and it may or may not be retained (by what criterion it is kept or "deleted", I'm not quite sure). Anyway, see if you think this is any more plausible with the added explan- ation... -- Tory S. Toupin | ttoupin@diana.cair.du.edu | Existence toward perfection... Unversity of Denver | Life of mediocrity. Undergraduate: Math & Computer Sciences| Denver, CO 80208 | -Tory Toupin ----- C'est ne pas un fichier de <<.signature>>
prune@athena.mit.edu (Paul Berland) (04/05/91)
Tory Toupin writes: >|> How about this: the brain is attempting to piece together memories to form >|> temporary concepts -- a form of forethought? That is to say, memories are >|> symbols of the state of the body when the memory was formed, and the brain >|> tends to superimpose(?) these symbols...and see if they are physically ^^^^^^^^^^ >|> possible situations based on the knowledge of situations which can/have ^^^^^^^^ >|> occurred by trying these memories out on the "virtual body" (i.e.: the >|> so-called mind's eye, but the entire body as well) and if it is not >|> physically possible, it mutates the symbols slightly so that ^^^^^^^^^^^^^^^^^^^ >|> they are not associated as well with one another. My objection was with the criterion "physically possible". I would use "satisfactorily representative" which is much more subjective to other forces besides the "physical" such as desirability towards satisfaction of some ego-related need, compatibility with surrounding memory "cluster," simplicity of storage, and other equally important criteria. > Now, as I stated above, in the process of dreaming, these symbols are put > together (based on some "strong" association--be it smell, image, > temperature, or other) to form a temporary symbol which is applied to > the virtual body... > Certainly, the virtual body is quite closely associated with the real body > since it is able to return the feedback that the real body would have had it > received the symbol. Thus, the virtual body should be able to recognize when > a symbol it has been given is physically possible ("virtually possible" is > perhaps a better way to phrase it) or physically not possible in the sense > that, if the body has never experienced the symbol before (say, sustained > flight without aid) it must not be possible. This sounds more like a "testing ground" theory of dreaming rather than a "garbage collection" theory of dreaming. I mentioned before that I thought the "garbage collection" theory was limited in that it failed to account for the creative process that occurs during dreaming. We COULD call the "testing ground" theory a "sophisticated garbage collection" theory since it has both creative and destructive (at some level) processes, but this seems to me only to highlight a minor point of what is otherwise a totally different theory. The "testing ground" theory would claim to make an account for Winson's 2nd and 3rd steps in the learning process described in Fiona Oceanstar's post "Re: Function of Dreams": > 2) associating new info. with memories of past experiences. > 3) formulating strategies to govern future behavior. The "testing ground" theory is certainly compatible with Winson's theories. Is it necessary and sufficient within that framework?