IMLane: Composable Framework for Efficient AI Function Execution in Database Engine
Summary: IMLane accelerates in-database AI functions by replacing CPython-limited thread parallelism with low-overhead process execution and decoupling scheduling from the engine. Integrated into OceanBase/DuckDB, it delivers 7.48×/5.04× average speedups. (summarized by gpt-5.6-luna on Aug 28 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Chenyang Zhang (East China Normal University)
- 2. Linjun Lu (East China Normal University)
- 3. Qingfeng Pan (East China Normal University)
- 4. Chen Xu (East China Normal University)
- 5. Xianzhong Cao (East China Normal University)
- 6. Quanqing Xu (Ant Financial)
- 7. Chuanhui Yang (Ant Financial)
BibTeX Citation
@article{zhang_vldb26,
title = {{IMLane: Composable Framework for Efficient AI Function Execution in Database Engine}},
author = {Zhang, Chenyang and Lu, Linjun and Pan, Qingfeng and Xu, Chen and Cao, Xianzhong and Xu, Quanqing and Yang, Chuanhui},
journal = {PVLDB},
series = {{VLDB} '26},
volume = {19},
number = {12},
pages = {4223--4236},
doi = {10.14778/3827998.3828028},
url = {https://doi.org/10.14778/3827998.3828028},
year = {2026}
}Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 43 of 43 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 10,653 | MCTuner: Spatial Decomposition-Enhanced Database Tuning via LLM-Guided Exploration | 2026 | SIGMOD |
| 2 | 2,913 | AI Meets Database: AI4DB and DB4AI | 2021 | SIGMOD |
| 3 | 5,211 | Database Workload Characterization with Query Plan Encoders | 2022 | VLDB |
| 4 | 6,455 | AlayaDB: The Data Foundation for Efficient and Effective Long-context LLM Inference | 2025 | SIGMOD |
| 5 | 11,198 | OpenMLDB: A Real-Time Relational Data Feature Computation System for Online ML | 2025 | SIGMOD |
| 6 | 5,723 | Accelerating Python UDFs in Vectorized Query Execution | 2022 | CIDR |
| 7 | 10,494 | Automating Database-Native Function Code Synthesis with LLMs | 2026 | SIGMOD |
| 8 | 7,844 | Bespoke OLAP: Synthesizing Workload-Specific One-size-fits-one Database Engines | 2026 | VLDB |
| 9 | 9,760 | Database as Runtime: Compiling LLMs to SQL for In-database Model Serving | 2025 | SIGMOD |
| 10 | 6,765 | Mitigating the Impedance Mismatch between Prediction Query Execution and Database Engine | 2025 | SIGMOD |