Aggregate Query Prediction under Dynamic Workloads

Savva, Fotis; Anagnostopoulos, Christos; Triantafillou, Peter

doi:10.1109/bigdata47090.2019.9006267

Cited by 4 publications

(3 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recent ML models usually use the predictive probabilities of the classifiers as a confidence score to identify changes [23,53,58,72]. Others may monitor the error of the underlying models and trigger an OOD signal when a significant change is captured [2,13,37,50,60]. While these approaches are very efficient in time, they typically come with limiting assumptions depending on the underlying model or application.…”

Section: Ood Detectionmentioning

confidence: 99%

Detect, Distill and Update: Learned DB Systems Facing Out of Distribution Data

Kurmanji,

Triantafillou

2023

Proc. ACM Manag. Data

Self Cite

View full text Add to dashboard Cite

Machine Learning (ML) is changing DBs as many DB components are being replaced by ML models. One open problem in this setting is how to update such ML models in the presence of data updates. We start this investigation focusing on data insertions (dominating updates in analytical DBs). We study how to update neural network (NN) models when new data follows a different distribution (a.k.a. it is "out-of-distribution" -- OOD), rendering previously-trained NNs inaccurate. A requirement in our problem setting is that learned DB components should ensure high accuracy for tasks on old and new data (e.g., for approximate query processing (AQP), cardinality estimation (CE), synthetic data generation (DG), etc.). This paper proposes a novel updatability framework (DDUp). DDUp can provide updatability for different learned DB system components, even based on different NNs, without the high costs to retrain the NNs from scratch. DDUp entails two components: First, a novel, efficient, and principled statistical-testing approach to detect OOD data. Second, a novel model updating approach, grounded on the principles of transfer learning with knowledge distillation, to update learned models efficiently, while still ensuring high accuracy. We develop and showcase DDUp's applicability for three different learned DB components, AQP, CE, and DG, each employing a different type of NN. Detailed experimental evaluation using real and benchmark datasets for AQP, CE, and DG detail DDUp's performance advantages.

show abstract

Section: Ood Detectionmentioning

confidence: 99%

Detect, Distill and Update: Learned DB Systems Facing Out of Distribution Data

Kurmanji,

Triantafillou

2023

Proc. ACM Manag. Data

Self Cite

View full text Add to dashboard Cite

show abstract

“…A new distribution might signify new user interests and application requirements. Studying the full effects of this has been studied in the context of Approximate Query Processing [44][45][46], but not in the context of explanation functions, which remains part of our future work.…”

Section: Off-line Adjustmentmentioning

confidence: 99%

Large-scale Data Exploration Using Explanatory Regression Functions

Savva

Anagnostopoulos

Triantafillou

et al. 2020

ACM Trans. Knowl. Discov. Data

Self Cite

View full text Add to dashboard Cite

Analysts wishing to explore multivariate data spaces, typically issue queries involving selection operators, i.e., range or equality predicates, which define data subspaces of potential interest. Then, they use aggregation functions, the results of which determine a subspace's interestingness for further exploration and deeper analysis. However, Aggregate Query (AQ) results are scalars and convey limited information and explainability about the queried subspaces for enhanced exploratory analysis. Analysts have no way of identifying how these results are derived or how they change w.r.t query (input) parameter values. We address this shortcoming by aiding analysts to explore and understand data subspaces by contributing a novel explanation mechanism based on Machine Learning. We explain AQ results using functions obtained by a threefold joint optimization problem which assume the form of explainable piecewise-linear regression functions. A key feature of the proposed solution is that the explanation functions are estimated using past executed queries. These queries provide a coarse grained overview of the underlying aggregate function (generating the AQ results) to be learned. Explanations for future, previously unseen AQs can be computed without accessing the underlying data and can be used to further explore the queried data subspaces, without issuing more queries to the backend analytics engine. We evaluate the explanation accuracy and efficiency through theoretically grounded metrics over real-world and synthetic datasets and query workloads. CCS Concepts: • Mathematics of computing → Exploratory data analysis; • Information systems → Data analytics; • Computing methodologies → Supervised learning by regression.

show abstract

“…In our experiments, we used the Keras library to train models incrementally within each node over the supporting clusters' data for each of the 200 queries issued. Each query has been randomly created over the whole data space based on the dynamic query workload method described in[18]. Accordingly, each query needs a specific range of numerical data to build a prediction model.…”

mentioning

confidence: 99%