kth.sePublications KTH
System disruptions
There are recurring service disruptions on some of the search portals due to high traffic. We are working to resolve the issue, you may temporarily encounter an error message.
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Query Optimization for Inference-Based Graph Databases
KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science, Software and Computer systems, SCS.ORCID iD: 0000-0002-8573-0090
2023 (English)In: VLDB-PhD 2023 - Proceedings of the VLDB 2023 PhD Workshop, co-located with the 49th International Conference on Very Large Data Bases, VLDB 2023, CEUR-WS , 2023, p. 33-36Conference paper, Published paper (Refereed)
Abstract [en]

Knowledge Graphs are commonly characterized by two challenges: massive scale and sparsity. The former leads to slow response times for complex queries with random data accesses, especially when they require deep graph traversals. The latter, which is caused by missing connections and characteristics in graphs modeling real information, implies that any analysis based solely on explicitly stored data is bound to yield incomplete results. This work aims to develop a novel graph database architecture that leverages the power of Graph Machine Learning to equip graph queries with prediction capabilities while offering approximate but timely results to complex queries. We discuss challenges, design decisions, and research avenues required in materializing this prototype alongside the outline of the actively-pursued research plan.

Place, publisher, year, edition, pages
CEUR-WS , 2023. p. 33-36
Keywords [en]
Graph Databases, Graph Representation Learning, Query Optimization, Uncertainty Estimation
National Category
Computer Sciences
Identifiers
URN: urn:nbn:se:kth:diva-336736Scopus ID: 2-s2.0-85169435702OAI: oai:DiVA.org:kth-336736DiVA, id: diva2:1798614
Conference
49th International Conference on Very Large Data Bases PhD Workshop, VLDB-PhD Workshop 2023, Vancouver, Canada, Aug 28 2023
Note

QC 20230919

Available from: 2023-09-19 Created: 2023-09-19 Last updated: 2026-05-22Bibliographically approved
In thesis
1. Foundations of Trustworthy AI-Native Data Systems
Open this publication in new window or tab >>Foundations of Trustworthy AI-Native Data Systems
2026 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

In traditional data management systems, queries have well-defined semantics and produce exact results. Integrating Machine Learning inference into data processing pipelines disrupts both properties by introducing operators whose outputs are approximate rather than exact. This thesis establishes two foundations for trustworthy AI-native data systems: empirical characterization of ML operator execution cost, and formal, declarative correctness guarantees that the system enforces on behalf of the user. We develop these foundations across three levels of abstraction, from single-operator cost, to single-operator correctness, to their joint optimization at the pipeline level, and establish Conformal Prediction as a practical statistical foundation for this approach. We introduce Crayfish, a benchmarking framework for ML inference within dataflow engines that reveals how interactions between serving tools, stream processors, and pipeline configurations shape inference costs in ways that are difficult to anticipate from component-level behavior alone. We propose ConANN, the first framework to provide distribution-free recall guarantees for Inverted File-based Approximate Nearest Neighbor search, using conformal methods to replace heuristic index tuning with formal statistical guarantees. At the pipeline level, we study joint cost and correctness optimization in the context of Neural Graph Databases, where multi-hop queries over Knowledge Graphs interleave retrieval and neural execution. We formalize a hybrid query optimization architecture for this setting, then introduce ConRAD, which enforces end-to-end recall guarantees for multi-hop queries while dynamically bypassing expensive neural inference when recall targets can be met with local graph evidence alone. Taken together, these contributions show that the rigor users expect from traditional data systems need not be abandoned as those systems become increasingly driven by Machine Learning.

Abstract [sv]

I traditionella datahanteringssystem har frågor väldefinierad semantik och producerar exakta resultat. Att integrera maskininlärningsinferens i databehandlingspipelines stör båda dessa egenskaper genom att introducera operatorer vars utdata är approximativa snarare än exakta. Denna avhandling etablerar två grundpelare för tillförlitliga AI-nativa datasystem: empirisk karaktärisering av exekveringskostnaden för ML-operatorer, samt formella, deklarativa korrekthetsgarantier som systemet upprätthåller å användarens vägnar. Vi utvecklar dessa grundpelare över tre abstraktionsnivåer, från enskild operatorkostnad till enskild operatorkorrekthet, och slutligen till deras gemensamma optimering på pipelinenivå. Vi etablerar Conformal Prediction som en praktisk statistisk grund för detta tillvägagångssätt. Vi introducerar Crayfish, ett benchmarkingramverk för ML-inferens inom dataflödesmotorer som synliggör hur interaktioner mellan serving-verktyg, strömprocessorer och pipelinekonfigurationer formar inferenskostnaden på sätt som är svåra att förutse enbart utifrån enskilda komponenters beteende. Vi föreslår ConANN, det första ramverket som erbjuder distributionsfria recall-garantier för Inverted File-baserad approximativ närmaste-granne-sökning, genom att använda konforma metoder för att ersätta heuristisk indexjustering med formella statistiska garantier. På pipelinenivå studerar vi gemensam optimering av kostnad och korrekthet i kontexten av neurala grafdatabaser, där flerstegsfrågor över kunskapsgrafer varvar hämtning med neural exekvering. Vi formaliserar en hybrid frågeoptimeringsarkitektur för detta scenario och introducerar sedan ConRAD, som upprätthåller end-to-end recall-garantier för flerstegsfrågor och samtidigt dynamiskt kringgår kostsam neural inferens när recall-målen kan uppnås med enbart lokal grafevidens. Sammantaget visar dessa bidrag att den stringens som användare förväntar sig av traditionella datasystem inte behöver överges i takt med att dessa system i allt högre grad drivs av maskininlärning.

Place, publisher, year, edition, pages
Stockholm, Sweden: KTH Royal Institute of Technology, 2026. p. xxiv, 73
Series
TRITA-EECS-AVL ; 2026:56
Keywords
AI-Native Data Systems, Model Serving, Conformal Risk Control, Approximate Nearest Neighbor Search, Neural Graph Databases, Query Optimization, AI-nativa datasystem, Modellservering, Konform riskkontroll, Approximativ närmaste-granne-sökning, Neurala grafdatabaser, Frågeoptimering
National Category
Computer Systems
Identifiers
urn:nbn:se:kth:diva-382126 (URN)978-91-8106-628-9 (ISBN)
Public defence
2026-06-15, https://kth-se.zoom.us/j/65502477126, F3 Flodis, Lindstedtvägen 26, Stockholm, 14:00 (English)
Opponent
Supervisors
Note

QC 20260522

Available from: 2026-05-22 Created: 2026-05-22 Last updated: 2026-08-24Bibliographically approved

Open Access in DiVA

No full text in DiVA

Scopus

Authority records

Horchidan, Sonia

Search in DiVA

By author/editor
Horchidan, Sonia
By organisation
Software and Computer systems, SCS
Computer Sciences

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 263 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf