進階搜尋


 
系統識別號 U0026-0812200914260064
論文名稱(中文) 異質性分類現象之檢驗:單一或多元類別表徵模型
論文名稱(英文) Examination of Heterogeneous Categorization Phenomenon: Single V.S. Multiple Category Representation Models
校院名稱 成功大學
系所名稱(中) 認知科學研究所
系所名稱(英) Institute of Cognitive Science
學年度 96
學期 2
出版年 97
研究生(中文) 王崇宇
研究生(英文) Chung-Yu Wang
電子信箱 punker.wang@gmail.com
學號 u7695106
學位類別 碩士
語文別 中文
論文頁數 89頁
口試委員 召集委員-黃金蘭
口試委員-鄭中平
指導教授-楊立行
中文關鍵字 異質性分類表徵  SUSTAIN  ATRIUM 
英文關鍵字 ATRIUM  SUSTAIN  Heterogeneous category representation 
學科別分類
中文摘要 從過去的類別研究一直發展至今,現行有主要的兩大分類理論:範例表徵理論與
規則表徵理論。兩大分類理論皆假定同一類別空間中的所有刺激皆會被相同的類別表
徵所處理,將類別表徵視為是同質性的,但是在Aha & Goldstone(1992),Erickson &
Kruschke(1998)以及Yang & Lewandowsky(2004)中,觀察到了人類在分類上所
表現出來的異質性分類現象,指的是在同一類別空間中的刺激,可以被不同的類別表
徵所處理。這個發現挑戰了傳統類別理論中的同質性假定,這些研究也發現傳統的類
別模型並沒有辦法解釋異質性類別實驗的結果,在本研究中以上述三個實驗的結果在
兩個具有潛力表現出異質性分類現象的異質性模型上進行檢驗,其中一個是具有多重
類別模組的ATRIUM,另一個是以彈性化單一表徵建構的SUSTAIN。檢驗結果發現
ATRIUM 可以成功的解釋三個實驗中的異質性分類現象,主要原因是因為ATRIUM
具有多重的類別表徵模組,且可以在不同的刺激上調整模組的使用比率。SUSTAIN
則只能有限的表現出異質性分類現象,雖然SUSTAIN 具有彈性化的演算模式,仍會
受到其單一表徵基礎的影響,同時也發現了SUSTAIN 會在相當程度上受到學習順序
的影響。然而從SUSTAIN 對Erickson & Kruschke(1998)實驗的模擬結果中發現,
其實驗中的規則表徵現象可能其實是範例的作用所導致,並且本研究在修改後的實驗
中發現,受試者可能是基於範例的影響而在Erickson & Kruschke(1998)實驗中表現
出規則表徵的現象,此外,電腦模擬結果說明SUSTAIN 會受到注意力調幅值的大小
影響,而產生相似度扭轉的現象。綜合來說,多重表徵模型對異質性分類現象的解釋
力要優於單一表徵模型,這也確立了多重表徵在建構分類模型時的必要性。
英文摘要 Now days, there are two major category representation theories in research of
categorization: Exemplar-based theory and Rule-based theory. Those two theories both
assume that in one single category space, all stimuli would be represented by the same
category representation. This kind of assumption treats category representation as
homogeneous. However in three category research: Aha & Goldstone (1992), Erickson &
Kruschke (1998) and Yang & Lewandowsky (2004), the author found the heterogeneous
phenomena in human categorization behaviors, the word “heterogeneous” means that the
stimuli in one single category space may be represented by different category
representations. This founding challenges the homogeneous assumption in traditional
category theory. Those researchers also found that traditional category model can’t
accommodate the result of category experiments. In this research, we select two category
models that have potential to accommodate the result of heterogeneous categorization
experiments and test the performances of these two models in three heterogeneous
categorization experiments. One of the two models is ATRIUM which build-in multiple
category modules, the other is SUSTAIN which has single but flexible category
representation. The result shows that ATRIUM can accommodate the result of three
experiments well, because ATRIUM able to use different category modules in
categorization process and adjust the ratio of modules’ output value. SUSTAIN can only
perform the limited result of the experiments. It means that even the structure of SUSTAIN
is flexible, it still limited by the assumption of single category representation. Another
finding is that SUSTAIN is affected by the learning sequence. According to the result of
SUSTAIN fit the Erickson & Kruschke (1998), the phenomenon of rule-based
representation might due to the effect of learning example. After modify the category space
in the experiment, we found out that the phenomenon of rule-based representation in
original experiment might due to the example in learning phase but not rule representation.
Also, SUSTAIN’s simulation result of the modified experiment shows that SUSTAIN
might affect by attention tuning and twists the similarity of stimuli. Summarize, the
performance of multiple category representations model is better than single category
representation model, this finding also certain the necessary of multiple category
representations in model fabrication.
論文目次 摘要 ............................................................................................................................. 1
第一章:概念、分類與類別理論 ............................................................................... 1
前言 ....................................................................................................................... 1
早期的類別學習理論 ........................................................................................... 2
古典理論 ....................................................................................................... 2
機率理論 ....................................................................................................... 2
原型理論 ....................................................................................................... 3
近代類別學習理論與模型 ................................................................................... 4
範例表徵理論與模型 ................................................................................... 5
規則表徵理論 .............................................................................................. 11
結論 ..................................................................................................................... 12
第二章:異質性類別實驗 ......................................................................................... 14
異質性類別表徵實驗 ......................................................................................... 14
Aha & Goldstone(1992) ......................................................................... 14
Erickson & Kruschke(1998) ............................................................... 16
Yang & Lewandowsky(2004) ................................................................ 18
結論 ..................................................................................................................... 22
第三章:ATRIUM 和SUSTAIN ............................................................................... 23
多重表徵類別模型 ............................................................................................. 23
ATRIUM ...................................................................................................... 23
彈性化單一表徵類別模型 ................................................................................. 30
SUSTAIN .................................................................................................... 30
結論 ..................................................................................................................... 35
第四章:在ATRIUM 與SUSTAIN 上的檢驗 ......................................................... 37
檢驗方法 ............................................................................................................. 37
結果 ..................................................................................................................... 38
Aha & Goldstone(1992) ......................................................................... 38
Erickson & Kruschke(1998) .................................................................. 45
Yang & Lewandowsky(2004) ................................................................ 49
結論 ..................................................................................................................... 60
第五章:規則與範例表徵實驗 ................................................................................. 61
方法 ..................................................................................................................... 63
受試者 ......................................................................................................... 63
實驗設計 ..................................................................................................... 63
刺激材料與實驗設備 ................................................................................. 63
實驗程序 ..................................................................................................... 65
實驗分析與結果 ................................................................................................. 67
學習階段 ..................................................................................................... 67
測驗階段 ..................................................................................................... 68
電腦模擬 ............................................................................................................. 71
預測結果 ..................................................................................................... 71
注意力調幅值 ............................................................................................. 73
結論 ..................................................................................................................... 76
第六章:綜合討論 ..................................................................................................... 77
主要結果、原因與探討 ..................................................................................... 77
異質性實驗在模型上的檢驗 ..................................................................... 77
SUSTAIN V.S. ATRIUM ............................................................................. 80
單一表徵 V.S. 多重表徵 .......................................................................... 80
模型解釋廣度 ............................................................................................. 81
第五章實驗結果探討 ................................................................................. 81
從模型使用探討研究方法 ......................................................................... 83
結論 ..................................................................................................................... 83
第八章:參考文獻 ..................................................................................................... 84
圖表目次
表次
表一 ATRIUM 中的自由估計參數 ................................................................... 29
表二 SUSTAIN 中的自由估計參數 ................................................................. 35
表三 ATRIUM 預測Aha & Goldstone 實驗的參數表。 ................................ 39
表四 SUSTAIN 預測Aha & Goldstone 實驗的參數表 ................................... 41
表五 ATRIUM 預測Erickson & Kruschke(1998)實驗的參數表 ............... 47
表六 SUSTAIN 預測Erickson & Kruschke(1998)實驗的參數表 .............. 48
表七 ATRIUM 對Yang & Lewandowsky(2004)的參數表。 ..................... 52
表八 SUSTAIN 對Yang & Lewandowsky(2004)的參數表。 .................... 56
表九 學習階段變異數分析摘要表 ................................................................... 67
表十 兩個關鍵刺激上二項式檢定表 ............................................................... 69
圖次
圖一 線性不可分割說明圖。 ............................................................................. 5
圖二 Context Model 使用刺激例圖 .................................................................... 7
圖三 MDS 說明圖。 ........................................................................................... 8
圖四 選擇性注意力。 ......................................................................................... 9
圖五 ALCOVE 模型結構圖 .............................................................................. 10
圖六 GRT 概念圖(Ashby & Townsend(1986))。 ....................................... 11
圖七 Aha & Goldstone(1992)實驗中的刺激與類別結構圖 ....................... 14
圖八 Aha & Goldstone(1992)實驗結果。關鍵刺激以粗體字標示。 ....... 15
圖九 Erickson & Kruschke(1998)實驗中刺激與類別結構圖 .................... 16
圖十 Erickson & Kruschke(1998)的實驗結果圖。 .................................... 18
圖十一 Yang & Lewandowsky(2004)中實驗的類別結構圖 ...................... 19
圖十二 TB 組的測驗階段示意與結果統計圖。 ............................................. 20
圖十三 KP 組的測驗階段結果統計與示意圖。 ............................................. 21
圖十四 Yang & Lewandowsky(2004)中所使用的ATRIUM 架構圖。 ..... 25
圖十五 規則節點激發示意圖。 ....................................................................... 25
圖十六 SUSTAIN 用於Yang & Lewandowsky (2004)模型結構圖。 ...... 31
圖十七 不同注意力值(λ)下的叢集激發值的坡度曲線。 ......................... 32
圖十八 ATRIUM 預測Aha & GoldsStone 實驗的平均預測圖。 .................. 39
圖十九 ATRIUM 最佳的一次估計中在學習階段不同學習範例上使用的規則
模組圖 ......................................................................................................... 40
圖二十SUSTAIN 預測Aha & GoldsStone 實驗的平均預測圖 ...................... 41
圖二十一 SUSTAIN 預測Aha & Goldstone 實驗的平均叢集結構圖。 ....... 42
圖二十二 SUSTAIN 正確分類2 個關鍵刺激的叢集結構圖。 ..................... 43
圖二十三 SUSTAIN 正確分類3 個關鍵刺激的叢集結構圖。 ..................... 43
圖二十四 SUSTAIN 正確分類4 個關鍵刺激的叢集結構圖。 ..................... 43
圖二十五 Erickson & Kruschke(1998)的反推實驗數據。 ........................ 46
圖二十六 ATRIUM 預測Erickson & Kruschke(1998)實驗的平均結果 ... 46
圖二十七 SUSTAIN 預測Erickson & Kruschke(1998)實驗的平均結果 .. 48
圖二十八 SUSTAIN 預測Erickson & Kruschke(1998)實驗的平均叢集結構
圖。 ............................................................................................................. 49
圖二十九 ATRIUM 對KP 組的預測圖。 ....................................................... 51
圖三十 ATRIUM 對TB 組的預測圖 ............................................................... 51
圖三十一 ATRIUM 預測KP 時的區間注意力變化值 ................................... 53
圖三十二 學習區間中刺激在不同情境下模組使用機率變化值 ................... 53
圖三十三 ATRIUM 預測TB 時的區間注意力變化值 ................................... 54
圖三十四 SUSTAIN 對KP 組的預測圖 .......................................................... 55
圖三十五 SUSTAIN 對TB 組的預測圖 .......................................................... 55
圖三十六 SUSTAIN 在Yang & Lewandowsky(2004)中的平均叢集結構圖。
..................................................................................................................... 57
圖三十七 SUSTAIN 修改起始注意力值後對TB 組的預測圖 ...................... 58
圖三十八 整體的RMSD 值分佈圖。 .............................................................. 59
圖三十九 被SUSTAIN 激發的規則範例修改方向圖。 ................................ 62
圖四十 實驗使用的刺激圖形 ........................................................................... 64
圖四十一 實驗的類別結構圖 ........................................................................... 65
圖四十二 實驗程序圖 ....................................................................................... 66
圖四十三 高度為4.5 的矩形刺激圖 ................................................................ 67
圖四十四 兩組受試者在所有區間正確率折線圖 ........................................... 68
圖四十五 未告知規則組的測驗階段結果圖 ................................................... 70
圖四十六 告知規則組的測驗階段結果圖 ....................................................... 70
圖四十七 SUSTAIN 針對規則告知組的最佳平均預測結果圖。 ................. 72
圖四十八 SUSTAIN 針對規則告知組的全部平均預測結果 ......................... 72
圖四十九 參數組合與預測方向關係圖。 ....................................................... 72
圖五十 SUSTAIN 所產生的叢集圖 ................................................................. 74
圖五十一 注意力調幅值為20 下的距離值與激發值關係圖 ......................... 74
圖五十二 在不同注意力調幅值之下,關鍵刺激對規則與異例叢集激發值的
關係圖 ......................................................................................................... 75
參考文獻 楊立行。(民96)認知模擬在類別學習上的應用:以Yang & Lewandowsky (2003)之
研究為例。中華心理學期刊,49卷,3期,285-300。
Aha, D. W., & Goldstone, R. (1992). Concept learning and flexible weighting. In J. K.
Kruschke (Ed.), Proceedings of the Fourteenth Annual Conference of the Cognitive
Science Society (pp. 534-539). Hillsdale, NJ: Erlbaum.
Ashby, F. G., & Townsend, J. T. (1986). Varieties of Perceptual Independence.
Psychological Review, 93(2), 154-179.
Ashby, F. G., & Gott, R. (1988). Decision rules in the perception and categorization of
multidimensional stimuli. Journal of Experimental Psychology: Learning, Memory,
and Cognition. 14, 33-53.
Ashby, F. G., & Maddox, W. T. (1990). Integrating information from separable
psychological dimensions. Journal of Experimental Psychology: Human Perception
and Performance, 16, 598-612.
Ashby, F. G., & Alfonso-Reese L. A., Turken, U., Waldron, E. M. (1998). A
Neuropsychological Theory of Multiple Systems in Category Learning.
Psychological Review, 105(3), 442-481.
Ashby, F. G., & O'Brien, J. B. (2005). Category learning and multiple memory systems.
85
TRENDS in Cognitive Sciences, 9(2), 83-89.
Erickson, M. A., & Kruschke, J. K. (1998). Rules and Exemplars in Category Learning.
Journal of Experimental Psychology: General, 127(2), 107-140.
Flannagan, M. J., Fried, L. S., & Holyoak, K. J. (1986). Distributional Expectations and the
Induction of Category Structure. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 12(2), 241-256.
Gluck, M. A., & Bower, G. H. (1988b). From conditioning to category learning: An
adaptive network model. Journal of Experimental Psychology: General, 117.
227-247.
Homa, D., & Chambliss, D.. (1975). The Relative Contributions of Common and
Distinctive Information on the Abstraction from Ill-Defined Categories. Journal of
Experimental Psychology, 104(4), 351-359.
Homa, D., Cross, I. J., Cornell, D., & Goldman, D., Shwartz, H. (1973). Prototype
abstraction and classification of new iinstances as a function of number of instances
defining the prototype. Journal of Experimental Psychology, 101(1), 116-122.
Jacobs, R. A., Jordan, M. I., Nowlan, S. J., & Hinton, G. E. (1991). Adaptive mixtures of
local experts. Neural Computation, 3, 79-87
Katz, J. J., & Fodor, J. A. (1963). The structure of semantic theory. Language, 39,
86
170-210.
Kruschke, J. K. (1992). ALCOVE: An exemplar-based connectionist model of category
learning. Psychological Review, 99, 22–44.
Love, B. C., Medin, D. L., & Gureckis, T. (2004). SUSTAIN: A network model of human
category learning. Psychological Review, 111, 309-332.
Luce, R. D. (1959). Individual choice behavior: A theoretical analysis. Westport, Conn.:
Greenwood Press.
Malt, B. C., & Smith, E. E. (1983). Correlated properties in natural categories. Journal of.
Verbal Learning and Verbal Behavior, 23, 250-269..
McCloskey, M., & Glucksberg, S. (1978). Natural categories: Well defined or fuzzy-sets?
Memory and Cognitive. 6:462-72
McKinley, S. C., & Nosofsky, R. M. (1996). Selective attention and the formation of linear
decision boundaries. Journal of Experimental Psychology: Human Perception &
Performance, 22, 294-317.
Medin. D. L., & Schaffer, M. M. (1978). Context theory of classification learning.
Psychological Review, 85, 207-238.
Medin, D. L., & Smith, E. E. (1981). Concepts and concept formation. Annual Reviews of
87
Psychology, 35, 113-138.
Medin, D. L., & Schwanenflugel, P. J. (1981). Linear Separability in Classification
Learning. Journal of Experimental Psychology: Human Learning and Memory,
7(5), 355-368.
Medin,D.L., & Edelson, S. M., Freko, D. (1982). Correlated Symptoms and Simulated
Medical Classification. Journal of Experimental Psychology: Learning, Memory,
and Cognition, 8(1), 14.
Nosofsky, R. M. (1986). Attention, similarity,.and the identificationcategorization
relationship. Journal of Experimental Psychology: General 115, 39-57.
Nosofsky, R. M. (1987). Attention and learning processes in the identification and
categorization of integral stimuli. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 13, 87-108.
Nosofsky, R. M., Clark, S. E., & Shin, H. J. (1989). Rules and exemplars in categorization,
identification, and recognition. Journal of Experimental Psychology: Learning.
Memory, and Cognition. 15. 282-304.
Nosofsky, R. M., Kruschke, J. K., & McKinley, M. C. (1992). Combining Exemplar-Based
Category Representations and Connectionist Learning Rules. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 18(2), 211-233.
88
Nosofsky, R. M., Palmeri, T. J., & McKinley, S. C. (1994). Rule-Plus-Exception Model of
Classification Learning. Psychological Review, 101(1), 53-79.
Posner, M. I., & Keele, S. W. (1968). On the genesis of abstract ideas. Journal of
Exprimental Psychoog,.77:353-63
Posner, M. I., & Keele, S. W. (1970). Retention of abstract ideas. Journal of Experimental
Psychology, 83(2), 304-308.
Poggio, W. H., & Girosio, F. (1990). A Theory of Networks for learning. Science, 247,
978-982.
Rosch, E., & Mervis, C. B. (1975). Family resemblances: Studies in the internal
structure of categories. Cognitive Psychology, 7, 573-605.
Rosch, E., & Lloyd, B. B., (Eds.). (1978). Cognition and Categorization. Hillsdale, NJ:
Erlbaum
Rouder, J. N., & Ratcliff, R. (2004). Comparing Categorization Models. Journal of
Experimental Psychology: General, 133(1), 63-82.
Rouder, J. N., & Ratcliff, R. (2006). Comparing Exemplar- and Rule- Based Theories of
Categorization. Current Directions In Psychological Science, 15(1).
Shepard, R. N. (1957). Stimulus and response generalization: A stochastic model
relating generalization to distance in psychological space. Psychometrika,
22,325-345.
89
Shepard, R. N., Hovland, C. L., & Jenkins, H. M. (1961). Learning and memorization of
classifications. Psychological Monographs. 75(13, Whole No. 517).
Strange, W., Keeney, T., Kessel, F. S., Jenkins, J. J. (1970). Abstraction over time of
prototypes from distortions of random dot patterns:. Journal of experimental
psychology, 83(3), 508-510.
Widrow, B., & Hoff, M. E. (1960). Adaptive switching circuits. In IRE WESCON
Convention Record (p. 96-104). New York.
Yang, L. X., & Lewandowsky, S. (2004). Context-gated knowledge partitioning in
categorization. Journal of Experimental Psychology: Learning, Memory, &
Cognition, 30, 1045-1064.
論文全文使用權限
  • 同意授權校內瀏覽/列印電子全文服務,於2009-08-08起公開。
  • 同意授權校外瀏覽/列印電子全文服務,於2009-08-08起公開。


  • 如您有疑問,請聯絡圖書館
    聯絡電話:(06)2757575#65773
    聯絡E-mail:etds@email.ncku.edu.tw