❓ Did you know what makes robots intelligent?
💎 In 1983, a developmental psychologist named Howard Gardner proposed that humans possess multiple (currently, nine) intelligences, thereby challenging the traditional view of intelligence based on I.Q. Tests.
💎 While this theory has influenced #pedagogy and educational sciences, the intelligence of artificial agents like #robots is rarely studied or discussed in these terms.
💎 Instead, the focus has been on the basic cognitive capabilities needed to exhibit intelligence and the development of artificial methods (computational models) to realize these cognitive capabilities. Some of these #artificialintelligence (A.I.) capabilities that are relevant for present-day robotics include:
🎥 Perception: Robots need A.I. to understand what they are seeing, hearing, or touching. For this, they use models and methods from #computervision, automatic speech recognition, natural language understanding, and #machinelearning in general.
🗝️ Knowledge representation: Robots need A.I. to represent and update its knowledge about the world. For this, it often relies on domain-specific #ontologies and semantic networks.
🔮 Reasoning: Robots need A.I. to reason based on what it knows, to derive new conclusions that are valid within the given context. Inductive, deductive, and abductive reasoning is the commonly used #logicalreasoning approaches.
🎯 Decision-making: Robots need A.I. to learn what would be the best action to perform, in a given situation. They use #reinforcementlearning to learn an optimal sequence of actions to maximize the expected rewards in the long run.
📋 Planning: Robots need A.I. to plan how to achieve their goals, i.e., to reach any state that satisfies the conditions specified in the goal (e.g., the robot is at a specific location, the table is clean, etc.). They use #automatedplanningmethods to determine the robot's actions and the sequence to arrive at one of the goal states.
❗To generate and exhibit intelligent behavior, robots integrate several of these A.I. models and methods within a robot control architecture. The intelligent behavior of artificial agents typically results from integrating different A.I. models and methods.
Check out the Gardner’s Theory of Multiple Intelligences: ▶️ https://lnkd.in/dvHgPKJn
To learn more about the cognitive (as well as physical) capabilities needed by robots, check out this Handbook of Robotics, especially chapters 3, 4, 8, and 9: AI-perspectives: the #Turing option
For more fun: Check out this collection of robots
Contributing Editor: Dr. Teena Hassan
Comments