Buying and Target-Based Behavior Control for Human Similar Artificial Agents

Cognitive neuroscience explains human behavior with two different mechanisms: routine, habits made without thinking and intentionally oriented actions. A new research aims to reveal the ability to exhibit more realistic and human-like behavior by concrete this dual structure in artificial intelligence agents.

Current human-like agent frames often focus on these features:

• Hedef-oriented behavior modeling weight loss; planned and informed decision-making processes Home
• largely ignored the habit-based behaviours, if such behaviors form most of everyday life
• Although it is successful in computer vision and natural language processing, limited progress in behavior control
• Single mechanism systems in artificial agent design

The approach of the research group develops a “multiple behavior control” system that models both control systems together. This allows agents to quickly automate routine tasks and protect flexibility to cope with new situations simultaneously. For example, a robot can be used to certain environment and can also give flexible reactions to unexpected obstacles, when doing repetitive movements there.

This progress opens the way to create close, adaptable and natural-looking artificial systems to human behavior. It is expected to be applied especially in robotics, game design and interactive simulation areas.