Lab Overview

Welcome to the Candy Lab of the School of Data Science and Engineering (DaSE) at East China Normal University (ECNU). As a prominent research group, we thrive on exploring and pushing the boundaries of how machine learning can redefine and revolutionize the world of database management.

Our Vision

Our main vision is to create an intelligent and self-adaptive data management system that seamlessly integrates cutting-edge machine learning techniques. By doing so, we aim to unlock unprecedented levels of efficiency, reliability, and scalability in data management tasks.

Core Research Areas

We are deeply passionate about a multitude of research topics within the broad spectrum of database management. Currently, our primary research interests encompass:

• AI4Storage: Here, we delve into the integration of AI with storage technologies. Key areas of focus include:

  • Non-volatile memory: Exploring how machine learning can optimize and enhance the functionalities of persistent memory structures.
  • LSM-tree (Log-Structured Merge-tree): Investigating the fusion of AI techniques to improve the performance and efficiency of LSM-tree operations.

• AI4Optimizer: Optimization lies at the heart of efficient data management. Our interests in this domain include:

  • Cardinality estimation: Employing AI to refine the predictability of the number of rows returned by a query, which is essential for optimizing query performance.

• AI4Operations: Elevating database operations through AI-driven strategies is one of our core pursuits. Key topics under this umbrella are:

  • Index recommendation: Harnessing machine learning algorithms to suggest optimal indexing strategies for databases, ensuring efficient data retrieval and management.

• Vector DB: In the realm of databases, vector databases are becoming increasingly crucial for tasks such as similarity search. Our focus areas include:

  • ANNS (Approximate Nearest Neighbor Search): Leveraging machine learning techniques to enhance the speed and accuracy of searches within large-scale datasets.

Join us in our journey as we blend traditional database management principles with state-of-the-art machine learning advancements to shape the future of data science and engineering.