Cognitive research essentially requires the introduction of models, and methodologies for empirically verifying those models. This is principally because this academic field intrinsically involves the issue of how to understand entities that are unobservable, or difficult to observe. Particularly for higher-order cognitive processes, many different factors must be related, and thus understanding those processes are virtually not possible without introducing models whose complexities match the complexity caused by those factors. This article argues, while summarizing previous efforts of forerunners in exploring cognitive research, that explorations of new methodologies that combine adequate models and advanced technologies for instrumentation and analysis will be required for making breakthroughs in higher-level cognitive research. The article also points out that important long-term actions for such breakthroughs must include the rescue of the current education from the large gap between humanities/social sciences and natural sciences/engineering, and the strong promotion of governmental policies for supporting the creation of new academic fields. Science takes a long time. Herbert A. Simon

抄録全体を表示

 The model-based approach, along with the experimental approach, is a primary research methodology in cognitive science. Cognitive scientists have contributed to the development of psychological science by the benefits of building computational cognitive models. The authors have examined another aspect of the benefits of cognitive modeling as a learning tool by the practices of cognitive science classes in which university students are instructed to build computational cognitive models. In this paper, we introduce class practice examples implemented over the past 10 years, and discuss the possibilities and limitations of a learning paradigm, “Learning by Building Cognitive Models.”

抄録全体を表示

 Users often observe anomalous behaviors of systems, such as machine failures, autonomous agents, and natural phenomena. We analyze the features and the benefits of the memory-based strategy, which focuses on memorization of instances to predict anomalous and regular behaviors of the system. In this study, we develop our previous research and investigate the cognitive processes and the benefits of the memory-based strategy with ACT-R model simulations. We set the parameters defining the encoding processes of anomalous instances and regular instances in the model of the memory-based strategy and performed simulations to verify how these two parameters influence prediction performance. The results of simulations showed that (1) anomalous instances are encoded and regular instances are not encoded in the memory-based strategy and that (2) such inactivity on regular instances suppresses commission errors of regular instances and does not suppress commission errors of anomalous instances and omission errors, which leads to correct prediction of systems' behaviors.

抄録全体を表示

 One of the challenges on developing intelligent tutoring systems in collaborative learning is to providing adaptive feedbacks with adequate facilitations. This study focuses on collaborative learning involving a knowledge integration activity, whereby learner dyads explain each other’s expert knowledge supported by a Pedagogical Conversational Agent. The goal of this paper was to investigate how collaborative process and learning gain can be determined by the degree to which learners synchronize their gaze (gaze recurrence) and use overlapping language (lexical alignment) during their interaction. This study conducted a laboratory-based eye-tracking experiment, wherein thirty-four learners’ gazes and oral dialogs were analyzed. Through this experiment, the author investigated how gaze recurrence and lexical alignment can predict collaborative learning process and learning gain. Multiple regression analysis was conducted, wherein learning performance was regressed on the two independent variables and shows how the model predicts both collaborative process and gain. The results also showed that both gaze recurrence and lexical overlap significantly predicted learning performance. These results indicate that the two variables might be useful for developing detection modules that enable a better understanding of learner-learner collaborative learning.

抄録全体を表示

 Recent studies on causal learning have shown that people use covariation information to infer casual structure, and that people have prior assumptions about causal structure and causal strength. Although covariation between two variables is insufficient to induce causal direction, learners give various interpretations to covariation information. Whereas necessity of causality assumes low base rate of the effect, sufficiency of causality expects high causal strength. These viewpoints result in opposite interpretations in causal structure learning. The purpose of the present study is to investigate prior assumptions in inferring preventive causal structure. Participants were asked to observe the states of bacteria (present or absent) and to infer their causal direction. The results found that judgments of causal structure varied as a function of covariation information, and that participants interpreted covariation according to sufficiency of causation. These findings are explained by asymmetries in generative and preventive causal relations. Theoretical implications and future directions are discussed.

抄録全体を表示

 This article reviewed the free-energy principle, proposed as the general and unified brain theory by Friston, K. et al. (2006), its powerful framework, and recent expansions. This theory developed a mathematically precise theory to explain the computational neural mechanisms for optimizing posterior beliefs of the world in the brain. The freeenergy principle consisted of two major inferences: the unconscious inference and active inference. In addition, to optimize posterior beliefs or to select and execute behaviors, this theory proposed the precision of signals and its optimization as important computations; it also predicted the aberrant optimization of precision triggered various psychopathological syndromes. Furthermore, the free-energy principle theoretically demonstrated the composition of values from intrinsic (or epistemic) and extrinsic values. Intrinsic values were considered to involve curiosity and play fundamental roles in decision making and behavioral selection. This article expounded how the free-energy principle gave the theoretical explanations for brain functions such as perception, motor behavior, behavioral selection, and insight.

抄録全体を表示

 落語の鑑賞中に観客が示す自発性瞬目の時間当たりの生起回数に着目すると，その増減のタイミングが観客間で一致する現象がしばしば観察される．この瞬目同期(blink synchronization) が生じる箇所は，口演の意味内容と対応していていたことから，演者の表現が観客一人ひとりの注意の配分・解放をガイドし，観客の注意過程に連動する瞬目も結果的に同期するのだと考えられている(野村・岡田, 2014,『認知科学』, 21(2), 226-244)． 研究1 では，瞬目同期の程度に観客の視聴経験が影響するか否かを心理実験で検討した．その結果，視聴経験が多い群では，視聴経験が少ない群よりも同期の程度は高く，視聴経験が瞬目同期を促進することが示唆された．研究2 では，集団実験と個人実験の比較から，観客間相互作用が瞬目同期に与える影響を検討した．その結果，集団実験の場合には，個人実験の場合に比べてズレ30%から60%が小さくなり，観客間相互作用が瞬目同期を促進することが示唆された． 数理的にみれば落語は，非線形力学系として応答する観客群に演者からの共通入力が印加される状況としてモデル化することができる．だが通常，こうした状況下の共通入力は未知であり，客観的に測定することは容易ではない．そこで以下の研究では，観客の瞬目から共通入力を再構成する手法を提案した．この手法ではまず，共通入力を複数の力学系に入力し，出力の点過程時系列を得る．次に，点過程時系列から時間当たりの生起数を算出し，一定間隔の遅れ時間をとって多次元空間に埋め込む．その後，力学系の数だけリカレンスプロットを作成し，これらの和集合をとる操作により，力学系に固有のダイナミクスを相殺する．これを本研究では，重畳リカレンスプロットと命名した．最後に，リカレンスプロットから多次元空間内の距離を算出する手法(Hirata et al., 2008) を重畳リカレンスプロットに適用し，共通入力を再構成して各時点の振幅を得る．研究3 では，パラメータを変えて分岐させた神経細胞の数理モデルに共通入力を与え，発火時刻の時系列の情報のみを用いて未知の共通入力を再構成できることを示し，提案手法の妥当性を確認した．研究4 では，提案手法を実験で得た瞬目時刻の時系列に適用し，演者からの共通入力を再構成した．本提案手法は，観客の瞬目に着目するため，落語に限らず，演劇，授業，映像作品といった視覚表現の訴求力を客観的に評価する手法として応用が期待できる．

抄録全体を表示