Key research themes
1. How can artificial systems model and exploit subjective perception and emotion to enhance autonomous adaptive behavior?
This research theme investigates the modeling of subjective experiences — such as emotions and perception — within artificial agents to improve autonomous decision-making, adaptability, and cognitive capabilities. Aligning computational models with subjective internal states like emotions enables systems to generate meaningful feedback loops, self-evaluate, and adjust behaviors contextually rather than relying on externally imposed parameters. This theme is crucial for developing artificial subjective systems that exhibit autonomous, self-aware, and context-sensitive intelligence beyond purely objective data processing.
2. What are effective computational frameworks and methodologies to handle subjectivity and uncertainty in data labeling and machine learning?
This theme explores mechanisms to address subjectivity and label uncertainty inherent in user-generated or subjective datasets, which challenge traditional supervised learning assumptions. Differentiating between subjective and objective classes, quantifying label noise, incorporating annotator reliability, and leveraging active or cooperative learning strategies are investigated to improve model robustness and reliability. This research area is fundamental to ensure systems learn valid representations when ground truth is ambiguous or influenced by human biases and perspectives.
3. How can theoretical and computational models quantitatively capture consciousness, intelligence, and qualia in artificial and natural agents?
This theme focuses on formal and computational theories aimed at defining, measuring, and detecting subjective experiences like consciousness and qualia, and intelligence across biological and artificial systems. It includes approaches applying theoretical computer science, algorithmic measures of prediction-based intelligence, and testing frameworks to detect subjective experience. Developing quantitative and substrate-independent models of consciousness and intelligence is pivotal to advancing artificial subjective systems capable of phenomenal experience and sophisticated reasoning.