Draft:DIKWP Model

• Comment: zero reliable independent sources? Theroadislong (talk) 11:17, 27 July 2025 (UTC)

---

Related papers in IEEEDIKWAbout DIKWP in mdpi

DIKWP Model
Title	DIKWP Model

overview

The DIKWP Model (short for Data–Information–Knowledge–Wisdom–Purpose) is a cognitive architecture for artificial intelligence proposed by Professor Duan Yucong. It expands upon the traditional DIKW hierarchy by introducing a fifth top-level element, Purpose, above Wisdom. This results in a five-layer structure: Data → Information → Knowledge → Wisdom → Purpose.

By adding the layer of Purpose, the DIKWP model emphasizes goal-driven cognition and aims to form a closed-loop system from perception (data) to decision-making (action). It is considered a new paradigm for artificial consciousness, designed to provide AI systems with higher-level cognitive capabilities that are interpretable and controllable.

Core Theory

The core concept of the DIKWP model involves incorporating subjective goals into AI cognitive architecture, addressing the traditional AI's lack of intrinsic motivation and self-awareness. In this model, the base "Data" layer covers raw sensory inputs, ascending through refined "Information," integrated "Knowledge," experiential "Wisdom," and culminating in the highest action-oriented "Purpose" layer. Unlike a simple unidirectional accumulation, interactions across layers are bidirectional, forming networked feedback loops. Upper-level intentions influence data processing and knowledge acquisition at lower layers, while lower-level knowledge and wisdom adjust and inform higher-level purposes. This dynamic bidirectional flow creates a closed cognitive loop, enhancing AI's self-adjustment and learning capabilities. In contrast to traditional passive black-box models, DIKWP's multilayer semantic parsing ensures traceability in AI reasoning processes, aligning AI actions with human value-oriented expectations. In short, the DIKWP model explicitly layers cognition and integrates autonomy and interpretability through purpose-driven intentions.

Model Structure

The five-layer structure of the DIKWP model is a formal extension of the classic DIKW hierarchy, traditionally visualized as a pyramid. DIKWP adds "Purpose" at the top and emphasizes a networked rather than strictly linear hierarchy. Each pair of adjacent layers has potential interactions, totaling 25 interactive modules (e.g., Data-Information, Information-Knowledge, up to Wisdom-Purpose interactions). These modules ensure a bidirectional information flow from foundational data to high-level objectives. For instance, the "Purpose" layer provides targets to guide the lower Wisdom and Knowledge layers, influencing data collection and information filtering, while new developments at the Data and Information layers can prompt purpose recalibration. This interconnected feedback loop structure mirrors human consciousness more closely. To operationalize the model, Duan's team proposed concepts such as Artificial Consciousness Processing Units (ACPU) and Artificial Consciousness Operating Systems (ACOS), facilitating computational and interactive implementations across the model's layers.

Applications

The DIKWP cognitive framework enhances AI systems' intelligence and reliability across several cutting-edge fields. In Active AI, it aids the development of explainable cognitive engines capable of self-reflection and continuous learning. Specific applications include multimodal large language models (LLMs), where DIKWP structure helps identify and address cognitive shortcomings within AI systems. In proactive medicine, DIKWP supports medical AI by facilitating semantic integration of complex medical data and knowledge, enabling personalized patient care through diagnostic knowledge graphs and semantic inference frameworks. Additionally, DIKWP principles enhance educational technology (such as learning pathway recommendations) and intelligent decision-support systems by embedding intent-driven interpretability and boosting overall AI credibility and effectiveness.

References

1. Duan, Yucong et al. “The Proposal of the DIKWP Five-layer Cognitive Model.” Artificial Intelligence Technology Report, 2025.
2. Hainan Daily. “Interview: DIKWP Model and Multimodal Artificial Intelligence.” 2023.
3. World Artificial Consciousness Conference Report. “Relativity of Consciousness and DIKWP.” 2023.
4. ScienceNet Blog. “Analysis of Artificial Consciousness Framework Based on DIKWP Model.” 2025.
5. 2023 IEEE International Conference on High Performance Computing & Communications, Data Science & Systems, Smart City & Dependability in Sensor, Cloud & Big Data Systems & Application (HPCC/DSS/SmartCity/DependSys)

External links

• The DIKWP Revolution: A New Horizon in Medical Dispute ResolutionBias in artificial intelligence - EQ and IQ bias testing based on the global language model
• DIKWP Model
• Related papers in IEEE

Category:Artificial intelligence Category:Cognitive science Category:Artificial consciousness