RALF: Accuracy-Aware Scheduling for Feature Store Maintenance

Summary: RALF schedules costly feature/embedding recomputation using downstream prediction-error feedback, query access patterns, and a regret objective rather than fixed refresh policies. Experiments show up to 32.7% lower error or 1.6× lower compute cost. (summarized by gpt-5.6-luna on Jul 24 2026)

Paper ID: 13923
Venue: VLDB
Year: 2024
Pagerank: 5.1924403e-05
Overall Rank: 9,940 | 31.81%
DOI: 10.14778/3632093.3632116

Incoming Non-self Citations Over Time

Authors

1. Sarah Wooders (University of California Berkeley)
2. Xiangxi Mo (University of California Berkeley)
3. Amit Narang (University of California Berkeley)
4. Kevin Lin (University of California Berkeley)
5. Ion Stoica (University of California Berkeley)
6. Joseph M. Hellerstein (University of California Berkeley)
7. Natacha Crooks (University of California Berkeley)
8. Joseph E. Gonzalez (University of California Berkeley)

BibTeX Citation

@article{wooders_vldb24,
        title = {{RALF: Accuracy-Aware Scheduling for Feature Store Maintenance}},
        author = {Wooders, Sarah and Mo, Xiangxi and Narang, Amit and Lin, Kevin and Stoica, Ion and Hellerstein, Joseph M. and Crooks, Natacha and Gonzalez, Joseph E.},
        journal = {PVLDB},
        series = {{VLDB} '24},
        volume = {17},
        number = {3},
        pages = {563--576},
        doi = {10.14778/3632093.3632116},
        url = {https://doi.org/10.14778/3632093.3632116},
        year = {2024}
}

Incoming Citations (Sorted by Pagerank)

Showing 1 of 1 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
7,847	Biathlon: Harnessing Model Resilience for Accelerating ML Inference Pipelines	2024	VLDB	5.5330423e-05

Outgoing Citations (Sorted by Pagerank)

Showing 12 of 12 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
47	PNUTS: Yahoo!'s Hosted Data Serving Platform	2008	VLDB	0.00044503718
284	NoScope: Optimizing Neural Network Queries over Video at Scale	2017	VLDB	0.00022370521
867	Coordination Avoidance in Database Systems	2015	VLDB	0.00013511197
1,108	Approximate Query Processing: No Silver Bullet	2017	SIGMOD	0.00012145154
1,667	Amazon DynamoDB: A Seamlessly Scalable Non-relational Datastore	2012	SIGMOD	0.00010066726
2,665	Probabilistically Bounded Staleness for Practical Partial Quorums	2012	VLDB	8.280143e-05
2,898	Approximate Selection with Guarantees using Proxies	2020	VLDB	7.978725e-05
3,438	A Demonstration of Willump: A Statistically-Aware End-to-end Optimizer for Machine Learning Inference	2020	VLDB	7.415647e-05
4,016	TARDiS: A Branch-and-Merge Approach To Weak Consistency	2016	SIGMOD	6.9530596e-05
4,534	Shared Arrangements: practical inter-query sharing for streaming dataflows	2020	VLDB	6.6420049e-05
6,073	OnlineSTL: Scaling Time Series Decomposition by 100x	2022	VLDB	5.9862575e-05
9,025	Tolerating Bounded Inconsistency for Increasing Concurrency in Database Systems	1992	PODS	5.329656e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	6,038	Efficient Construction of Approximate Ad-Hoc ML models Through Materialization and Reuse	2018	VLDB
2	5,925	Managing ML Pipelines: Feature Stores and the Coming Wave of Embedding Ecosystems	2021	VLDB
3	4,053	Database-Agnostic Workload Management	2019	CIDR
4	6,954	DLRover-RM: Resource Optimization for Deep Recommendation Models Training in the Cloud	2024	VLDB
5	6,843	Learning to Optimize LSM-trees: Towards A Reinforcement Learning based Key-Value Store for Dynamic Workloads	2023	SIGMOD
6	5,704	Optimizing Data Pipelines for Machine Learning in Feature Stores	2023	VLDB
7	8,040	PerfGuard: Deploying ML-for-Systems without Performance Regressions, Almost!	2021	VLDB
8	4,118	Rafiki: Machine Learning as an Analytics Service System	2019	VLDB
9	9,385	The Hopsworks Feature Store for Machine Learning	2024	SIGMOD
10	6,708	Serving Deep Learning Models with Deduplication from Relational Databases	2022	VLDB