Hypothesis-Test Cycles and Symbol Grounding

DOI Web Site Open Access
  • Hasida Kōiti
    Graduate School of Information Science and Technology, The University of Tokyo
  • Shimada Sotaro
    School of Science and Technology, Meiji University
  • Imai Mutsumi
    Faculty of Environment and Information Studies, Keio University

Bibliographic Information

Other Title
  • 仮説検証サイクルと記号接地
  • カセツ ケンショウ サイクル ト キゴウ セッチ

Search this article

Description

Language acquisition is a process of symbol grounding, which is construction of sym-<br>bol systems adapted to environment. Environmental adaptation defines the values<br> which cognitive agents pursue primarily by means of hypothesis-test cycles encompass-<br>ing both the inside and outside of their bodies. In addition to these directly grounded<br> cycles, there are also hypothesis-test cycles within cognitive agents. Cognitive processes<br> are combinations of these cycles, where cycles embody typical cognitive phenomena such<br> as navigation and language use.<br> Cycles are essentially countable, so that systems comprising cycles necessarily have<br> discrete structures. A cognitive agent is hence formulated as a discrete system consist-<br>ing of cycles including both directly grounded cycles and symbols (indirectly grounded<br> cycles), where each cycle embodies some value or meaning directly or indirectly associ-<br>ated with environmental adaptation. Computational models of cognition as combina-<br>tion of such cycles (values = meanings) are far more efficient (simpler and less prone to<br> overdesign) than traditional models stipulating possibly non-cyclic information flows.<br> Environmental-adaptation cycles operate at multiple spatiotemporal scales, including<br> real-time adaptive behavior, middle-term learning, and evolution across generations. It<br> is vitally important to address real-time adaptation behavior in terms of cycles, which<br> will raise the efficiency of the computational model not only at the level of real-time<br> adaptation but also accordingly at higher levels. Cycle-based (meaning-based) com-<br>putational models are necessary also because cycles derive meta-level constraints such<br> as symmetry bias and naming insight, which are indispensable for abductive reasoning<br> and language acquisition.<br> Existing technologies including deep learning fail to reflect such a value-based<br> (meaning-based) architecture of cognition. For the sake of thorough symbol grounding,<br>novel approaches are necessary which should integrate environmental-adaptation cycles <br>in the entire computational model at multiple levels of meaning and value.

Journal

Related Projects

See more

Keywords

Details 詳細情報について

Report a problem

Back to top