sensoryinputpatternstorage感觉输入模式存储

上传人:wux****ua 文档编号:21717779 上传时间:2021-05-07 格式:PPT 页数:20 大小:290KB
收藏 版权申诉 举报 下载
sensoryinputpatternstorage感觉输入模式存储_第1页
第1页 / 共20页
sensoryinputpatternstorage感觉输入模式存储_第2页
第2页 / 共20页
sensoryinputpatternstorage感觉输入模式存储_第3页
第3页 / 共20页
资源描述:

《sensoryinputpatternstorage感觉输入模式存储》由会员分享,可在线阅读,更多相关《sensoryinputpatternstorage感觉输入模式存储(20页珍藏版)》请在装配图网上搜索。

1、A MODEL OF HUMAN MEMORYJuly 2008 Memory patterns are probably not stored in the brains neurons. Why ? They take too long to train. New “on-the-fly” training disturbs old knowledge. Old knowledge must be retrained in order to be maintained. Where would the old knowledge be stored ? Catch 22. Once neu

2、ral networks are trained, the training patterns are discarded. Human memory does not discard training patterns. Human memory records input patterns in great detail.Cite: Bla Juleszs work with random dot stereograms and eidetic subject. SENSORY INPUT PATTERN STORAGE Sensory input patterns come from e

3、yes, ears, tactile, olfactory, vestibular and other sensors. Incoming patterns are stored in empty folders, wherever they are located. Sequences of patterns, like videos are stored in the same folder. Visual, auditory, and other sensory patterns that were received at the same time are stored in the

4、same folder. Only “interesting” input patterns are stored in the memory. They remain for the rest of ones life. An eidetic stores everything, interesting or not. Sensory input patterns go to short term memory and, if interesting, are transferred to main memory and recorded for life. Short term memor

5、y ( of the order of a few seconds ) is used in the determination of what is interesting. The problem solver can also decide what is interesting. Autoassociative neural networks are used in the pattern retrieval process. Pattern retrieval occurs in response to prompt patterns. Prompt patterns may com

6、e from sensory inputs. The memory is organized in the form of independent segments to make possible a very large storage capacity. SEGMENT N NNSENSING STRONG TRAINING MUX SEGMENT N +1 PROBLEM SOLVERHIT? HIT LINESENSORYINPUT NNSENSING STRONG TRAINING MUXHIT?VC AC SC SHORTTERMMEMORY “INTERESTING INPUT

7、” SWITCHSENSORY INPUT LINE MEMORY INPUT LINE VC = VISUAL CORTEXAC = AUDITORY CORTEXSC = SENSORY CORTEX SENSORY INPUT PATTERN STORAGE INBORN PATTERN STORAGE Inborn knowledge in the form of patterns is pre-loaded in the developing brains memory and remains intact throughout ones lifetime. Examples of

8、inborn knowledge : (a) A bird building a nest involves complex construction in “safe” places such as roof tops, tree tops, telephone poles, etc. (b) Baby horse walking and finding lunch within half hour of birth. (c) Human baby sucking, crying, peeing and pooping. It is conjectured that the memory s

9、torage means for inborn knowledge is the same as for sensory knowledge gained during a lifetime. It is conjectured that the memory retrieval means for inborn data is the same as for sensory input data. Inborn patterns are stored in folders in “memory segment 0”. MEMORY INPUT LINE SEGMENT 0 NNSENSING

10、 MUXINBORN PATTERNS SENSORY INPUT LINEHIT LINE SHORTTERMMEMORY “INTERESTING INPUT” SWITCH INBORN PATTERN STORAGEVCAC HIT? SC VC = VISUAL CORTEXAC = AUDITORY CORTEXSC = SENSORY CORTEX THOUGHT PATTERN STORAGE Thought patterns are also stored in memory. Storage means and retrieval means for thought pat

11、terns are the same as for sensory input patterns. Thought patterns come from the “problem solver”. The design of the problem solver is not yet part of this study, but could be thought of as a mechanism based on Arthur Samuels checker player. Thought patterns are always interesting and stored in empt

12、y memory locations. Storage of thought patterns takes precedence over storage of sensory input patterns. SEGMENT N NNSTRONG TRAINING MUX SEGMENT N+1 STRONG TRAINING MUXPROBLEM SOLVER MEMORY INPUT LINE THOUGHT PATTERNS THOUGHT PATTERN STORAGE NN PATTERN RETRIEVAL SYSTEM Patterns stored in memory can

13、be retrieved without knowledge of their storage location. Autoassociative neural networks are part of the retrieval mechanism. Autoassociative neural networks are trained by using their input patterns as both input and desired response patterns.They are trained to produce outputs that are reproducti

14、ons of their inputs. Once trained, autoassociative networks produce small input/output differences when presented with patterns that were trained in, but large differences when presented with patterns that were not trained in. Dj Vu ? Hit or no hit ? Autoassociative networks are trained with all the

15、 patterns stored in the connected memory folders. The autoassociative networks are prompted with sensory input patterns or thought patterns. Visual input patterns for example are rotated, translated, scaled, brightened, contrasted, etc. by the “visual cortex” VC while attempting to make a hit. If th

16、ere is a hit, the hit pattern is saved and compared with the contents of all the connected memory folders.The patterns of the folder containing the hit pattern are retrieved and sent to the problem solver which is the memory output “customer”. These patterns in turn may be used as prompts to retriev

17、e other folders.This type of feedback could cause a “chain reaction” resulting in the retrieval of many interrelated folders.(I have been speaking with someone for ten minutes but what is his name? Oh, now I remember. Its Jonathan Jones.) SEGMENT NNN SENSING STRONG TRAINING PATTERN RETRIEVAL SYSTEM

18、MUX PROBLEM SOLVERHIT?SENSORY INPUT PATTERNSPROMPT LINE VC AC SC THOUGHT PATTERNSOUTPUTBUFFER HIT? SEGMENT N+1NNSENSING STRONG TRAINING MUX HIT? BUFFER HIT? RETRIEVED PATTERNS LINE RETRIEVED PATTERNSPROMPTSIGNAL PROMPT PATTERNSSHORTTERMMEMORYVCAC SC PROMPT LINE VC= VISUAL CORTEXAC = AUDITORY CORTEXS

19、C = SENSORY CORTEX TRAINING DURING NON-REM SLEEP It is speculated that the autoassociative neural networks are trained during non-REM sleep. Multiplexers sense the memory folders, sequentially feeding the pattern contents to the autoassociative networks for training. The training process initiates a

20、utomatically once the brain is in “sleep mode”. This continues throughout the night during periods of non-REM sleep. SEGMENT N NN NORMAL TRAINING MUX TRAINING DURING NON-REM SLEEP SEGMENT N+1 NN NORMAL TRAINING MUX MEMORY ACTIVITY DURING REM SLEEP Every 90 minutes or so during the night, the brain g

21、oes into “REM mode”. Each episode of REM lasts for about 20 - 30 minutes, increasing as the night progresses. It is generally believed that during REM ( Rapid Eye Movement ) sleep, the person is dreaming. The body is paralyzed during REM sleep, probably to prevent the person from acting out the drea

22、m. During REM, contents are pulled from memory prompted by thought patterns from the problem solver. Memory contents provide further prompts to retrieve further related contents. This is a “chain reaction”. The retrieved memory contents are available to the problem solver. The memory contents are ju

23、xtaposed and intermingled in strange ways, creating fantasies that are dreams.The dreams themselves are stored in new memory locations. During REM, the autoassociative neural networks are trained hard when dream patterns are stored and when patterns are drawn from the memory folders. These networks

24、are both sensed and trained during REM. Brain activity during REM is similar to that of wide-awake consciousness, according to EEG and FMRI. The difference is that during REM, the sensory inputs from the eyes, ears,etc are shut off. SEGMENT NNN SENSING STRONG TRAINING MEMORY ACTIVITY DURING REM ( DR

25、EAM ) SLEEP MUX PROBLEM SOLVERHIT?PROMPT LINE VC AC SC THOUGHT PATTERNSBUFFER HIT? SEGMENT N+1NNSENSING STRONG TRAINING MUX HIT? BUFFER HIT? RETRIEVED PATTERNS LINE RETRIEVED PATTERNSPROMPT PATTERNS MEMORY INPUT LINE PROMPT LINE SPECULATION ON THE REM STATE It is speculated that the purpose of REM s

26、leep is problem solving.Uninhibited thought can be highly creative. During a nights sleep, episodes of REM take place about every 90 minutes or so. Upon awaking one is generally unaware of having dreamt unless waking in the middle of a dream. To retrieve an unaware dream from memory, one needs an ap

27、propriate prompt. This is the function of psychoanalysis. It is speculated that schizophrenia is an abnormal condition under which the subject is awake and conscious and in REM sleep at the same time, with fantasized images superposed on top of real-time visual, auditory, etc. inputs.This is halluci

28、nation. The fantasized images are drawn from memory spontaneously, without prompting. Under hypnosis, a normal subject is awake and conscious and in REM sleep at the same time, a state induced by the hypnotist. The subject responds to visual and auditory inputs from the hypnotist that serve as promp

29、ts. The hypnotist can have a two way interaction with the subject, and can store and retrieve information in and from the subjects memory. SPECULATION ABOUT SEEING, HEARING, WALKING, SPEAKING, ETC. Seeing involves processing and recording new visual images and making associations with pre-recorded i

30、mages stored in memory. Vision and memory are intertwined. Hearing and understanding speech involves processing and recording new auditory images and making associations with pre-recorded auditory images stored in memory. Hearing and speech understanding and memory are intertwined. While walking, se

31、nsory signals from all over the body deliver to the brain information about the mechanical state of the body. These sensory signals acts as prompts to the memory that, in turn, provides muscle control signals that enable walking. This works like a lookup table. Muscle control signals are not compute

32、d in real time but are pulled from memory. Control planning is also pulled from memory. While speaking, muscle control of the vocal tract is pulled from memory in response to prompts. The brain does not compute these control signals in real time. SPECULATION ABOUT FEATURE DETECTION Hubel and Wiesels

33、 discovery of cat cortical cells that respond to vertical and horizontal lines suggest importance of feature detection. Juleszs work with random dot stereograms and experiments with an eidetic subject suggest that the visual process involves the total image in full detail (not just features) and tha

34、t pattern association is critical.SPECULATION ABOUT LEARNING Learning involves storing patterns in memory. Supervised learning, unsupervised learning, learning with a critic, bayesian learning are all useful concepts but probably have little to do with human learning. SPECULATION ABOUT MEMORY FAILUR

35、E Ageing causes slow death of neurons and the dendritic tree, with insufficient rate of replacement. This affects memory retrieval as it becomes more difficult to continue training the brains neural networks. Old data and newly recorded data gradually become inaccessible. Old data lasts the longest.

36、 Alzheimers disease with associated plaques and tangles in the neurons and dendritic tree has the effect of accelerated ageing of these brain structures, makes neural training more difficult, gradually becoming impossible, thereby making old and new memories but especially new memories more and more

37、 inaccessible. Brain injury like in the movie Memento cuts the link to the “memory input line” and prevents the formation of new memories. Short-term memory still works. Recall of old memories still works, and neural training still works. The eidetic memory stores everything of interest, like normal

38、, but training autoassociative networks during non-REM sleep is very strong. SPECULATION ABOUT THE MECHANISMS OF STORAGE AND RETRIEVAL At the moment of conception, DNA is taken from the mother and father to form a new cell. That is the start of a new living animal. The DNA of the new cell contains t

39、he information ( the “blue print” ) needed to construct the living animal. The DNA contains the information to construct the body, the internal organs, including the brain. The DNA also contains the inborn information that will be pre-loaded in the developing brain. Inborn information is stored in D

40、NA. The mechanism for storage and retrieval of the information gained during a lifetime is the same as that for storage and retrieval of inborn information. WILD GUESSES: All information stored in memory is stored in DNA. The DNA that stores this information may be located in the glial cells of the

41、brain. Stored information is not stored in the neurons and dendritic tree.The neurons and the dendritic tree play a key role in association and retrieval of stored information. Some Interesting Questions Why do new-born babies sleep so much, and a lot less when they become kids? Why do babies cry so

42、 much, and less as they grow older? Why do adults cry so little? Memory of children and young people is great. Why cant we remember life events that happened in the first few years? What is the connection between schizophrenia and sleep walking? What did Sigmund Freud do that was right? Why does one

43、 often not remember performing acts of habit? What are habits? How are they related to addiction? How are they related to speech and locomotion? How can you change or break habits? How are questions of sexuality, hetero or homo, related to questions about autism? What causes subject wandering in the middle of a conversation?

展开阅读全文
温馨提示:
1: 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
2: 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
3.本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
5. 装配图网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。
关于我们 - 网站声明 - 网站地图 - 资源地图 - 友情链接 - 网站客服 - 联系我们

copyright@ 2023-2025  zhuangpeitu.com 装配图网版权所有   联系电话:18123376007

备案号:ICP2024067431-1 川公网安备51140202000466号


本站为文档C2C交易模式,即用户上传的文档直接被用户下载,本站只是中间服务平台,本站所有文档下载所得的收益归上传人(含作者)所有。装配图网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私,请立即通知装配图网,我们立即给予删除!