Beyond rational imitation: Learning arbitrary means actions from communicative demonstrations

被引:104
|
作者
Kiraly, Ildiko [1 ]
Csibra, Gergely [2 ]
Gergely, Gyorgy [2 ]
机构
[1] Eotvos Lorand Univ, Dept Cognit Psychol, H-1064 Budapest, Hungary
[2] Cent European Univ, Dept Cognit Sci, H-1051 Budapest, Hungary
基金
欧洲研究理事会;
关键词
Rational imitation; Relevance-guided imitation; Teleological stance; Natural pedagogy; Social Learning; Ostensive communication; GOAL ATTRIBUTION; YOUNG-CHILDREN; INFANTS; 12-MONTH-OLD; SELECTION; REASON; AGENCY; CUES;
D O I
10.1016/j.jecp.2012.12.003
中图分类号
B844 [发展心理学(人类心理学)];
学科分类号
040202 ;
摘要
The principle of rationality has been invoked to explain that infants expect agents to perform the most efficient means action to attain a goal. It has also been demonstrated that infants take into account the efficiency of observed actions to achieve a goal outcome when deciding whether to reenact a specific behavior or not. It is puzzling, however, that they also tend to imitate an apparently suboptimal unfamiliar action even when they can bring about the same outcome more efficiently by applying a more rational action alternative available to them. We propose that this apparently paradoxical behavior is explained by infants' interpretation of action demonstrations as communicative manifestations of novel and culturally relevant means actions to be acquired, and we present empirical evidence supporting this proposal. In Experiment 1, we found that 14-month-olds reenacted novel arbitrary means actions only following a communicative demonstration. Experiment 2 showed that infants' inclination to reproduce communicatively manifested novel actions is restricted to behaviors they can construe as goal-directed instrumental acts. The study also provides evidence that infants' reenactment of the demonstrated novel actions reflects epistemic motives rather than purely social motives. We argue that ostensive communication enables infants to represent the teleological structure of novel actions even when the causal relations between means and end are cognitively opaque and apparently violate the efficiency expectation derived from the principle of rationality. This new account of imitative learning of novel means shows how the teleological stance and natural pedagogy two separate cognitive adaptations to interpret instrumental versus communicative actions are integrated as a system for learning socially constituted instrumental knowledge in humans. (C) 2013 Elsevier Inc. All rights reserved.
引用
收藏
页码:471 / 486
页数:16
相关论文
共 36 条
  • [21] Learning Geometric Constraints of Actions from Demonstrations for Manipulation Task Planning
    Yuan, Jinqiang
    Chew, Chee-Meng
    Subramaniam, Velusamy
    2018 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND BIOMIMETICS (ROBIO), 2018, : 636 - 641
  • [22] The Art of Imitation: Learning Long-Horizon Manipulation Tasks From Few Demonstrations
    von Hartz, Jan Ole
    Welschehold, Tim
    Valada, Abhinav
    Boedecker, Joschka
    IEEE ROBOTICS AND AUTOMATION LETTERS, 2024, 9 (12): : 11369 - 11376
  • [23] Temporal Logic Imitation: Learning Plan-Satisficing Motion Policies from Demonstrations
    Wang, Yanwei
    Figueroa, Nadia
    Li, Shen
    Shah, Ankit
    Shah, Julie
    CONFERENCE ON ROBOT LEARNING, VOL 205, 2022, 205 : 94 - 105
  • [24] An imitation learning framework for generating multi-modal trajectories from unstructured demonstrations
    Peng, Jian-Wei
    Hu, Min-Chun
    Chu, Wei-Ta
    NEUROCOMPUTING, 2022, 500 : 712 - 723
  • [25] Zero-shot Imitation Learning from Demonstrations for Legged Robot Visual Navigation
    Pan, Xinlei
    Zhang, Tingnan
    Ichter, Brian
    Faust, Aleksandra
    Tan, Jie
    Ha, Sehoon
    2020 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA), 2020, : 679 - 685
  • [26] Ess-InfoGAIL: Semi-supervised Imitation Learning from Imbalanced Demonstrations
    Fu, Huiqiao
    Tang, Kaiqiang
    Lu, Yuanyang
    Qi, Yiming
    Deng, Guizhou
    Sung, Flood
    Chen, Chunlin
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 36 (NEURIPS 2023), 2023,
  • [27] Best-in-class imitation: Non-negative positive-unlabeled imitation learning from imperfect demonstrations
    Zhang, Lin
    Zhu, Fei
    Ling, Xinghong
    Liu, Quan
    INFORMATION SCIENCES, 2022, 601 : 71 - 89
  • [28] Multi-Modal Imitation Learning from Unstructured Demonstrations using Generative Adversarial Nets
    Hausman, Karol
    Chebotar, Yevgen
    Schaal, Stefan
    Sukhatme, Gaurav
    Lim, Joseph J.
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 30 (NIPS 2017), 2017, 30
  • [29] Self-Adaptive Imitation Learning: Learning Tasks with Delayed Rewards from Sub-optimal Demonstrations
    Zhu, Zhuangdi
    Lin, Kaixiang
    Dai, Bo
    Zhou, Jiayu
    THIRTY-SIXTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE / THIRTY-FOURTH CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE / TWELVETH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2022, : 9269 - 9277
  • [30] Extrapolating Beyond Suboptimal Demonstrations via Inverse Reinforcement Learning from Observations
    Brown, Daniel S.
    Goo, Wonjoon
    Nagarajan, Prabhat
    Niekum, Scott
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 97, 2019, 97