100x Cost & Latency Reduction: Performance Analysis of AI Query Approximation using Lightweight Proxy Models: [Experiments & Analysis]
Summary: Evaluates embedding-based lightweight proxy models for SQL AI.IF/AI.RANK, achieving >100× lower cost and latency while preserving or improving accuracy on datasets up to 10M rows. Demonstrates OLAP and HTAP architectures plus faster proxy training. (summarized by gpt-5.6-luna on Jul 26 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Yeounoh Chung (Google)
- 2. Rushabh Desai (Google)
- 3. Jian He (Google)
- 4. Yu Xiao (Google)
- 5. Thibaud Hottelier (Google)
- 6. Yves-Laurent Kom Samo (Google)
- 7. Pushkar Khadilkar (Google)
- 8. Xianshun Chen (Google)
- 9. Sam Idicula (Google)
- 10. Fatma Özcan (Google)
- 11. Alon Halevy (Google)
- 12. Yannis Papakonstantinou (Google)
BibTeX Citation
@inproceedings{chung_sigmod26,
title = {{100x Cost \& Latency Reduction: Performance Analysis of AI Query Approximation using Lightweight Proxy Models: [Experiments \& Analysis]}},
author = {Chung, Yeounoh and Desai, Rushabh and He, Jian and Xiao, Yu and Hottelier, Thibaud and Samo, Yves-Laurent Kom and Khadilkar, Pushkar and Chen, Xianshun and Idicula, Sam and Özcan, Fatma and Halevy, Alon and Papakonstantinou, Yannis},
series = {{SIGMOD} '26},
booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
publisher = {Association for Computing Machinery},
doi = {10.1145/3802002},
url = {https://dl.acm.org/doi/10.1145/3802002},
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 7 of 7 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 1,245 | Palimpzest: Optimizing AI-Powered Analytics with Declarative Query Processing | 2025 | CIDR | 0.00011507415 |
| 1,343 | DocETL: Agentic Query Rewriting and Evaluation for Complex Document Processing | 2025 | VLDB | 0.00011095866 |
| 3,684 | ThalamusDB: Approximate Query Processing on Multi-Modal Data | 2024 | SIGMOD | 7.2033959e-05 |
| 4,045 | Logical and Physical Optimizations for SQL Query Execution over Large Language Models | 2025 | SIGMOD | 6.9394654e-05 |
| 4,081 | Abacus: A Cost-Based Optimizer for Semantic Operator Systems | 2026 | VLDB | 6.9165634e-05 |
| 6,118 | ELEET: Efficient Learned Query Execution over Text and Tables | 2024 | VLDB | 5.9698795e-05 |
| 7,685 | Beyond Quacking: Deep Integration of Language Models and RAG into DuckDB | 2025 | VLDB | 5.5678545e-05 |
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 4,114 | Optimizing Machine Learning Inference Queries with Correlative Proxy Models | 2022 | VLDB |
| 2 | 11,624 | Accelerating Queries over Unstructured Data with ML | 2021 | CIDR |
| 3 | 9,882 | Demonstration of Accelerating Machine Learning Inference Queries with Correlative Proxy Models | 2022 | VLDB |
| 4 | 295 | Accelerating Machine Learning Inference with Probabilistic Predicates | 2018 | SIGMOD |
| 5 | 1,279 | AI Meets AI: Leveraging Query Executions to Improve Index Recommendations | 2019 | SIGMOD |
| 6 | 5,073 | Database Workload Characterization with Query Plan Encoders | 2022 | VLDB |
| 7 | 4,045 | Logical and Physical Optimizations for SQL Query Execution over Large Language Models | 2025 | SIGMOD |
| 8 | 10,139 | Leveraging Query Optimizers to Verify the Soundness of LLM-based Query Rewrites for Real-World Workloads, and More! | 2026 | CIDR |
| 9 | 3,907 | Accelerating Approximate Aggregation Queries with Expensive Predicates | 2021 | VLDB |
| 10 | 9,353 | On Efficient Approximate Queries over Machine Learning Models | 2023 | VLDB |