Role-Governed Categorization and Category Learning as a Result from Structural Alignment: The RoleMap Model
The paper presents a symbolic model for category learning and categorization (called RoleMap). Unlike the other models which implement learning in a separate working mode, role-governed category learning and categorization emerge in RoleMap while it does its usual reasoning. The model is based on several basic mechanisms known as reflecting the sub-processes of analogy-making. It steps on the assumption that in their everyday life people constantly compare what they experience and what they know. Various commonalities between the incoming information (current experience) and the stored one (long-term memory) emerge from those comparisons. Some of those commonalities are considered to be highly important, and they are transformed into concepts for further use. This process denotes the category learning. When there is missing knowledge in the incoming information (i.e. the perceived object is still not recognized), the model makes anticipations about what is missing, based on the similar episodes from its long-term memory. Various such anticipations may emerge for different reasons. However, with time only one of them wins and is transformed into a category member. This process denotes the act of categorization.
[1] Markman, Arthur B., and Brian H. Ross. "Category use and category learning." Psychological bulletin 129.4 (2003): 592. W.-K. Chen, Linear Networks and Systems (Book style). Belmont, CA: Wadsworth, 1993, pp. 123–135.
[2] Goldstone, Robert L., Alan Kersten, and Paulo F. Carvalho. "Concepts and Categorization” In I. B. Weiner (Ed.), Handbook of Psychology, Second Edition (2012).
[3] Yamauchi, Takashi, and Arthur B. Markman. "Learning categories composed of varying instances: The effect of classification, inference, and structural alignment." Memory & Cognition 28.1 (2000): 64-78.
[4] Gentner, Dedre, and Kenneth J. Kurtz. "Relational categories." Categorization inside and outside the lab (2005): 151-175.
[5] Markman, Arthur B., and C. Hunt Stilwell. "Role-governed categories." Journal of Experimental & Theoretical Artificial Intelligence 13.4 (2001): 329-358.
[6] Lin, Emilie L., and Gregory L. Murphy. "Thematic relations in adults' concepts." Journal of experimental psychology: General 130.1 (2001): 3.
[7] Goldwater, Micah B., Rebecca Bainbridge, and Gregory L. Murphy. "Learning of role-governed and thematic categories." Acta psychologica 164 (2016): 112-126.
[8] Asmuth, Jennifer A., and Dedre Gentner. "Context sensitivity of relational nouns." Proceedings of the Annual Meeting of the Cognitive Science Society. Vol. 27. No. 27. 2005.
[9] Nosofsky, Robert M. "Attention, similarity, and the identification–categorization relationship." Journal of experimental psychology: General 115.1 (1986): 39.
[10] Love, Bradley C., Douglas L. Medin, and Todd M. Gureckis. "SUSTAIN: a network model of category learning." Psychological review 111.2 (2004): 309.
[11] LeCun, Yann, Yoshua Bengio, and Geoffrey Hinton. "Deep learning." Nature 521.7553 (2015): 436.
[12] Rips, Lance J., Edward E. Smith, and Douglas L. Medin. "Concepts and categories: Memory, meaning, and metaphysics." The Oxford handbook of thinking and reasoning (2012): 177-209.
[13] Kuehne, Sven, et al. "SEQL: Category learning as progressive abstraction using structure mapping." Proceedings of the 22nd annual meeting of the cognitive science society. Vol. 4. 2000.
[14] Falkenhainer, Brian, Kenneth D. Forbus, and Dedre Gentner. "The structure-mapping engine: Algorithm and examples." Artificial intelligence 41.1 (1989): 1-63.
[15] Doumas, Leonidas AA, John E. Hummel, and Catherine M. Sandhofer. "A theory of the discovery and predication of relational concepts." Psychological review 115.1 (2008): 1.
[16] Goldwater, Micah B., Arthur B. Markman, and C. Hunt Stilwell. "The empirical case for role-governed categories." Cognition 118.3 (2011): 359-376.
[17] Kemp, Charles, et al. "A probabilistic model of theory formation." Cognition 114.2 (2010): 165-196.
[18] Kokinov, B. "The context-sensitive cognitive architecture DUAL." Proceedings of the Sixteenth Annual Conference of the Cognitive Science Society. 1994.
[19] Gentner, Dedre, and Mary Jo Rattermann. "Language and the career of similarity." Perspectives on language and thought: Interrelations in development 225 (1991).
[1] Markman, Arthur B., and Brian H. Ross. "Category use and category learning." Psychological bulletin 129.4 (2003): 592. W.-K. Chen, Linear Networks and Systems (Book style). Belmont, CA: Wadsworth, 1993, pp. 123–135.
[2] Goldstone, Robert L., Alan Kersten, and Paulo F. Carvalho. "Concepts and Categorization” In I. B. Weiner (Ed.), Handbook of Psychology, Second Edition (2012).
[3] Yamauchi, Takashi, and Arthur B. Markman. "Learning categories composed of varying instances: The effect of classification, inference, and structural alignment." Memory & Cognition 28.1 (2000): 64-78.
[4] Gentner, Dedre, and Kenneth J. Kurtz. "Relational categories." Categorization inside and outside the lab (2005): 151-175.
[5] Markman, Arthur B., and C. Hunt Stilwell. "Role-governed categories." Journal of Experimental & Theoretical Artificial Intelligence 13.4 (2001): 329-358.
[6] Lin, Emilie L., and Gregory L. Murphy. "Thematic relations in adults' concepts." Journal of experimental psychology: General 130.1 (2001): 3.
[7] Goldwater, Micah B., Rebecca Bainbridge, and Gregory L. Murphy. "Learning of role-governed and thematic categories." Acta psychologica 164 (2016): 112-126.
[8] Asmuth, Jennifer A., and Dedre Gentner. "Context sensitivity of relational nouns." Proceedings of the Annual Meeting of the Cognitive Science Society. Vol. 27. No. 27. 2005.
[9] Nosofsky, Robert M. "Attention, similarity, and the identification–categorization relationship." Journal of experimental psychology: General 115.1 (1986): 39.
[10] Love, Bradley C., Douglas L. Medin, and Todd M. Gureckis. "SUSTAIN: a network model of category learning." Psychological review 111.2 (2004): 309.
[11] LeCun, Yann, Yoshua Bengio, and Geoffrey Hinton. "Deep learning." Nature 521.7553 (2015): 436.
[12] Rips, Lance J., Edward E. Smith, and Douglas L. Medin. "Concepts and categories: Memory, meaning, and metaphysics." The Oxford handbook of thinking and reasoning (2012): 177-209.
[13] Kuehne, Sven, et al. "SEQL: Category learning as progressive abstraction using structure mapping." Proceedings of the 22nd annual meeting of the cognitive science society. Vol. 4. 2000.
[14] Falkenhainer, Brian, Kenneth D. Forbus, and Dedre Gentner. "The structure-mapping engine: Algorithm and examples." Artificial intelligence 41.1 (1989): 1-63.
[15] Doumas, Leonidas AA, John E. Hummel, and Catherine M. Sandhofer. "A theory of the discovery and predication of relational concepts." Psychological review 115.1 (2008): 1.
[16] Goldwater, Micah B., Arthur B. Markman, and C. Hunt Stilwell. "The empirical case for role-governed categories." Cognition 118.3 (2011): 359-376.
[17] Kemp, Charles, et al. "A probabilistic model of theory formation." Cognition 114.2 (2010): 165-196.
[18] Kokinov, B. "The context-sensitive cognitive architecture DUAL." Proceedings of the Sixteenth Annual Conference of the Cognitive Science Society. 1994.
[19] Gentner, Dedre, and Mary Jo Rattermann. "Language and the career of similarity." Perspectives on language and thought: Interrelations in development 225 (1991).
@article{"International Journal of Information, Control and Computer Sciences:77687", author = "Yolina A. Petrova and Georgi I. Petkov", title = "Role-Governed Categorization and Category Learning as a Result from Structural Alignment: The RoleMap Model", abstract = "The paper presents a symbolic model for category learning and categorization (called RoleMap). Unlike the other models which implement learning in a separate working mode, role-governed category learning and categorization emerge in RoleMap while it does its usual reasoning. The model is based on several basic mechanisms known as reflecting the sub-processes of analogy-making. It steps on the assumption that in their everyday life people constantly compare what they experience and what they know. Various commonalities between the incoming information (current experience) and the stored one (long-term memory) emerge from those comparisons. Some of those commonalities are considered to be highly important, and they are transformed into concepts for further use. This process denotes the category learning. When there is missing knowledge in the incoming information (i.e. the perceived object is still not recognized), the model makes anticipations about what is missing, based on the similar episodes from its long-term memory. Various such anticipations may emerge for different reasons. However, with time only one of them wins and is transformed into a category member. This process denotes the act of categorization.
", keywords = "Categorization, category learning, role-governed category, analogy-making, cognitive modeling.", volume = "12", number = "8", pages = "578-8", }
