Call for Papers: Special Issue on Data and Knowledge Empowered Generative Artificial Intelligence (DK-GenAI)

IEEE Transactions on Knowledge and Data Engineering seeks submissions for upcoming issues.

Submission Deadline: 30 June 2026

Publication Date: Early 2027

Generative Artificial Intelligence (GenAI), particularly Large Language Models (LLMs), has achieved unprecedented success in natural language understanding and generation. However, the transition of GenAI from experimental demos to reliable, mission-critical systems is currently hindered by fundamental limitations inherent to purely probabilistic models: hallucinations, lack of factual freshness, poor interpretability, and the inability to access private or domain-specific data.

We argue that the next leap in AI performance will not come solely from more parameters or larger computing power, but from data-centric and knowledge-centric approaches. This includes the rigorous engineering of pre-training data, the integration of structured knowledge graphs to provide factual grounding, and the optimization of retrieval systems (e.g., vector and graph databases) to supply relevant context.

The central thesis of this special issue is that the solution to the limitations of GenAI lies in Data and Knowledge Engineering. This special issue seeks to establish the role of the TKDE community in the GenAI era: providing the structured, high-quality foundations based on data engineering and symbolic knowledge representation that are necessary to make GenAI trustworthy, accurate, and efficient. We invite submissions of papers that address fundamental issues, propose novel methods, or showcase compelling applications that shed light on the data and knowledge empowered GenAI.

Topics and Scopes

Important: This special issue focuses exclusively on how data and knowledge empower GenAI.

In Scope: Methodologies where data engineering, database systems, knowledge graphs, or knowledge representation technologies are used to improve the training, inference, accuracy, or efficiency of GenAI models.
Out of Scope: Applications where GenAI is merely used as a tool to solve data problems (e.g., using LLMs solely for data cleaning or Text-to-SQL generation), without a focus on the underlying data and knowledge infrastructure. Papers that merely fine-tune a model without providing insights into the data engineering or knowledge representation aspects will be also considered out of scope.

We invite original research papers that utilize data and knowledge-based techniques to enhance GenAI. Topics include, but are not limited to:

Data-Centric AI for Model Training and Tuning

Automated data curation, filtering, and deduplication strategies for pre-training datasets.
Impact of data quality and diversity on GenAI scaling laws.
Instruction tuning data engineering: construction, balancing, and quality assessment.
Algorithms for machine unlearning and data influence estimation in generative models.
Efficient data loaders and pipelines for large-scale model training.

Knowledge Graph-Enhanced Generation (Neuro-Symbolic AI)

Grounding GenAI outputs in Knowledge Graphs to mitigate hallucinations.
RAG utilizing structured knowledge (e.g., sub-graph retrieval).
Fact verification and consistency checking of generated content using external knowledge bases.
Injecting symbolic logic and ontological constraints into the generation process.

Retrieval Systems and Infrastructure for GenAI

Optimization of vector databases and approximate nearest neighbor search for RAG.
Hybrid retrieval techniques combining keyword search and dense vector retrieval.
Data freshness and real-time knowledge updates for frozen LLMs.
Efficient context window management and data selection for in-context learning.

Data Management and Infrastructure for AI Agents

Architectures for long-term and short-term agent memory management.
State persistence, checkpointing, and recovery mechanisms for long-running agents.
Semantic indexing and schema design for tool and API registries.
Shared context management and synchronization in multi-agent environments.
Data-centric context engineering for Agentic AI.

Trustworthiness through Data and Knowledge

Bias detection and mitigation in training data and knowledge bases.
Copyright protection and attribution mechanisms in data retrieval pipelines.
Privacy-preserving retrieval and generation (e.g., differential privacy in RAG).

Systems and Evaluation

Benchmarks and evaluation metrics for knowledge-intensive generation.
System architectures for deploying data-empowered GenAI.
Cost-efficient strategies for RAG and fine-tuning.

Submission Instructions:

For author information and guidelines on submission criteria, visit the Author’s Information Page. Please submit papers through the IEEE Author Portal and be sure to select the special issue or special section name. Manuscripts should not be published or currently submitted for publication elsewhere. An extended version of a published conference paper can be submitted to TKDE for consideration if (1) the conference paper does not already appear in a journal and (2) it contains significant new material over 30%. Contact Li Sun (lsun@bupt.edu.cn) and Zhongtian Sun (z.sun-256@kent.ac.uk) for further details on this special issue.

In addition to submitting your paper to TDKE, you are also encouraged to upload the data related to your paper to IEEE DataPort. IEEE DataPort is IEEE's data platform that supports the storage and publishing of datasets while also providing access to thousands of research datasets. Uploading your dataset to IEEE DataPort will strengthen your paper and will support research reproducibility. Your paper and the dataset can be linked, providing a good opportunity for you to increase the number of citations you receive. Data can be uploaded to IEEE DataPort prior to submitting your paper or concurrent with the paper submission.

Guest Editors

Sihem Amer-Yahia, CNRS, France
Diego Calvanese, Free University of Bozen-Bolzano, Italy
Arijit Khan, Bowling Green State University, USA
Xin Wang, Tianjin University, China