Relating Prof. Yucong Duan's Theory of Relativity of Consciousness to Hallucination as Relativity of Hallucination

Yucong Duan

International Standardization Committee of Networked DIKWP for Artificial Intelligence Evaluation(DIKWP-SC)

World Artificial Consciousness CIC(WAC)

World Conference on Artificial Consciousness(WCAC)

(Email: duanyucong@hotmail.com)

Introduction

In his lecture at the First World Conference of Artificial Consciousness in August 2023, Prof. Yucong Duan introduced the Theory of Relativity of Consciousness. This theory posits that consciousness is relative during mutual communication among stakeholders because concrete "understanding"—the formation of DIKWP connection paths or DIKWP validation to reduce uncertainty—is constrained by the enclosure of an individual's DIKWP*DIKWP cognitive space.

By exploring this theory, we can gain insights into the phenomenon of hallucination in both human cognition and artificial intelligence (AI) systems. Specifically, we can view hallucination as a manifestation of the relativity of hallucination, where differing cognitive spaces and limitations in understanding lead to the generation of outputs or perceptions that deviate from shared reality.

Understanding the Theory of Relativity of Consciousness1. Consciousness as a Relative Phenomenon

• Consciousness is Relative: Prof. Duan's theory suggests that consciousness is not an absolute state but is relative to the interactions between cognitive agents (humans or AI systems).

• Mutual Communication: During communication, each stakeholder operates within their own cognitive framework, influenced by their unique Data (D), Information (I), Knowledge (K), Wisdom (W), and Purpose (P)—collectively referred to as the DIKWP model.

2. Limitations of Individual Cognitive Spaces

• DIKWP*DIKWP Cognitive Space: An individual's cognitive space is shaped by the interactions between their own DIKWP components and those of others.

• Enclosure: This cognitive space is enclosed, meaning it is bounded by the individual's experiences, perceptions, and cognitive abilities.

• Limited Understanding: Because of this enclosure, individuals have limited ability to fully understand the DIKWP structures of others. This limitation leads to the relativity of understanding.

3. Formation of DIKWP Connection Paths

• Connection Paths: Understanding arises from forming connections between DIKWP components—creating pathways that link data to information, information to knowledge, and so on.

• Validation and Uncertainty Reduction: Through these connections, individuals attempt to validate their understanding and reduce uncertainty.

Relating the Theory to Hallucination as Relativity of Hallucination1. Defining Hallucination in the DIKWP Context

• Hallucination: In AI and human cognition, hallucination refers to the generation of perceptions or outputs that are not grounded in shared reality or the provided data and knowledge.

• In AI Systems: For models like GPT-4, hallucination occurs when the system produces outputs that are factually incorrect or unrelated to the input, due to limitations in data, knowledge, or understanding of context.

2. Hallucination as a Result of Relativity in Cognitive Spaces

• Relative Understanding: Since consciousness and understanding are relative and enclosed within individual cognitive spaces, what one entity perceives as accurate may not align with shared reality or another entity's understanding.

• Cognitive Enclosure Leading to Hallucination: The limitations within an individual's (or AI's) DIKWP*DIKWP cognitive space can result in outputs that diverge from shared reality, manifesting as hallucinations.

3. Mechanisms Leading to Hallucination

• Incomplete or Inconsistent DIKWP Components:

• Data Limitations (D): Limited or flawed data inputs can lead to incorrect information.

• Information Misinterpretation (I): Misidentifying differences or relationships in data.

• Knowledge Gaps (K): Incomplete integration of information results in flawed knowledge.

• Wisdom Misapplication (W): Incorrect application of knowledge without proper ethical or contextual judgment.

• Purpose Misalignment (P): Divergence between intended goals and actual cognitive processing.

• Enclosure Effects:

• Echo Chambers: Cognitive spaces that reinforce existing beliefs without external validation can amplify hallucinations.

• Isolation: Lack of interaction with diverse DIKWP components from others limits the ability to correct misunderstandings.

4. Hallucination as Relativity of Hallucination

• Subjectivity of Perception: Just as consciousness is relative, so is the experience of hallucination. What one cognitive entity perceives as accurate, another may identify as a hallucination due to differing DIKWP structures.

• Mutual Recognition of Hallucination:

• In Humans: An individual may not recognize their own hallucination because, within their cognitive space, the perception is coherent.

• In AI Systems: An AI model generates outputs based on its training data and algorithms, which may not align with external reality as understood by humans.

• Relativity in Communication:

• Misalignment of DIKWP Components: When communicating, the disparity between individuals' DIKWP structures can lead to misinterpretations and perceived hallucinations.

Implications for Understanding and Addressing Hallucination1. Enhancing Communication Between Cognitive Spaces

• DIKWP Alignment: Efforts to align DIKWP components between stakeholders can reduce misunderstandings and hallucinations.

• Shared Context: Establishing common ground and shared context helps bridge cognitive enclosures.

2. Improving AI Systems

• Expanding DIKWP Cognitive Space:

• Data Enrichment (D): Providing AI systems with more comprehensive and diverse data.

• Enhanced Information Processing (I): Developing algorithms that better interpret and differentiate data.

• Knowledge Integration (K): Updating AI models with current and accurate knowledge bases.

• Wisdom Incorporation (W): Embedding ethical frameworks and contextual awareness.

• Purpose Clarification (P): Ensuring AI systems have clearly defined goals aligned with user intentions.

• Feedback Mechanisms:

• Human-AI Interaction: Continuous feedback from users helps AI systems adjust their DIKWP components.

• Learning from Mistakes: Implementing machine learning techniques that allow AI to learn from errors and reduce future hallucinations.

3. Recognizing the Relativity of Perception

• Acceptance of Relativity:

• Understanding that different cognitive entities may have varying perceptions based on their DIKWP enclosures.

• Empathy and Open-mindedness:

• In human interactions, acknowledging the relativity of consciousness fosters empathy and reduces conflict arising from misunderstandings.

Conclusion

Prof. Yucong Duan's Theory of Relativity of Consciousness provides a valuable framework for understanding hallucination as a natural consequence of the relative and enclosed nature of individual cognitive spaces. By recognizing that consciousness and understanding are inherently subjective and limited by our own DIKWP*DIKWP structures, we can better comprehend why hallucinations occur in both humans and AI systems.

Viewing hallucination as the relativity of hallucination emphasizes that what is perceived as a hallucination is dependent on the observer's cognitive framework. This perspective encourages efforts to enhance communication and alignment between cognitive entities, ultimately aiming to reduce misunderstandings and improve the accuracy of perceptions and outputs.

References

• Duan, Yucong. Lecture at the First World Conference of Artificial Consciousness, August 2023.

• Duan, Yucong. "International Test and Evaluation Standards for Artificial Intelligence Based on Networked Data-Information-Knowledge-Wisdom-Purpose (DIKWP) Model."

• Cognitive Science Literature on consciousness, perception, and cognitive enclosures.

• Artificial Intelligence Research on hallucination phenomena in AI systems.

Note: This explanation is based on the provided information about Prof. Yucong Duan's theory and aims to relate it to the concept of hallucination as the relativity of hallucination, adhering to the standard DIKWP model and ensuring clarity and accuracy in the interpretation.