FuseRank (Demo): Filtered Vector Search in Multimodal Structured Data

Dimitris Paraschakis; Rasmus Ros; Markus Borg; Per Runeson

doi:10.1007/978-3-031-70371-3_29

FuseRank (Demo): Filtered Vector Search in Multimodal Structured Data

Dimitris Paraschakis, Rasmus Ros, Markus Borg, Per Runeson

Theca Systems

Research output: Chapter in Book/Report/Conference proceeding › Paper in conference proceeding › peer-review

Abstract

We describe and demonstrate our work on multimodal filtered vector search in tabular data. It offers a practical way for businesses to vectorize their product assortments or any other business-critical data and then simultaneously retrieve and filter this information using state-of-the-art similarity search. Our methodology is based on the extended vector space model, with multiple modalities represented as sub-vectors that get concatenated and compared to the query vector via a dot product operation. It is a flexible framework that allows manipulating the influence of each modality on the overall item ranking via modality weights. We share the source code, the demonstration video, and the screenshot of the application. We also provide a brief description of its main building blocks, the supported data types, and modality filters. The application is bundled with two public datasets and pre-computed text embeddings so that it can be easily run without prior preparation.

Original language	English
Title of host publication	Machine Learning and Knowledge Discovery in Databases. Research Track and Demo Track - European Conference, ECML PKDD 2024, Proceedings
Editors	Albert Bifet, Povilas Daniušis, Jesse Davis, Tomas Krilavičius, Meelis Kull, Eirini Ntoutsi, Kai Puolamäki, Indrė Žliobaitė
Publisher	Springer
Pages	404-408
Number of pages	5
ISBN (Electronic)	9783031703713
ISBN (Print)	9783031703706
DOIs	https://doi.org/10.1007/978-3-031-70371-3_29
Publication status	Published - 2024
Event	European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2024 - Vilnius, Lithuania Duration: 2024 Sept 9 → 2024 Sept 13

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14948
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2024
Country/Territory	Lithuania
City	Vilnius
Period	2024/09/09 → 2024/09/13

Subject classification (UKÄ)

Computer Sciences

Free keywords

Information retrieval
Multimodal filtered vector search

Access to Document

10.1007/978-3-031-70371-3_29

Continuous Optimization for Multimodal Data Fusion in Vector Search
Paraschakis, D. (Researcher), Runeson, P. (PI), Borg, M. (PI) & Ros, R. (Supervisor)
2023/09/01 → 2025/08/31
Project: Research

Cite this

Paraschakis, D., Ros, R., Borg, M., & Runeson, P. (2024). FuseRank (Demo): Filtered Vector Search in Multimodal Structured Data. In A. Bifet, P. Daniušis, J. Davis, T. Krilavičius, M. Kull, E. Ntoutsi, K. Puolamäki, & I. Žliobaitė (Eds.), Machine Learning and Knowledge Discovery in Databases. Research Track and Demo Track - European Conference, ECML PKDD 2024, Proceedings (pp. 404-408). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14948). Springer. https://doi.org/10.1007/978-3-031-70371-3_29

Paraschakis, Dimitris ; Ros, Rasmus ; Borg, Markus et al. / FuseRank (Demo) : Filtered Vector Search in Multimodal Structured Data. Machine Learning and Knowledge Discovery in Databases. Research Track and Demo Track - European Conference, ECML PKDD 2024, Proceedings. editor / Albert Bifet ; Povilas Daniušis ; Jesse Davis ; Tomas Krilavičius ; Meelis Kull ; Eirini Ntoutsi ; Kai Puolamäki ; Indrė Žliobaitė. Springer, 2024. pp. 404-408 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{3f87a21e83ae400b956c211f7a4dc6fa,

title = "FuseRank (Demo): Filtered Vector Search in Multimodal Structured Data",

abstract = "We describe and demonstrate our work on multimodal filtered vector search in tabular data. It offers a practical way for businesses to vectorize their product assortments or any other business-critical data and then simultaneously retrieve and filter this information using state-of-the-art similarity search. Our methodology is based on the extended vector space model, with multiple modalities represented as sub-vectors that get concatenated and compared to the query vector via a dot product operation. It is a flexible framework that allows manipulating the influence of each modality on the overall item ranking via modality weights. We share the source code, the demonstration video, and the screenshot of the application. We also provide a brief description of its main building blocks, the supported data types, and modality filters. The application is bundled with two public datasets and pre-computed text embeddings so that it can be easily run without prior preparation.",

keywords = "Information retrieval, Multimodal filtered vector search",

author = "Dimitris Paraschakis and Rasmus Ros and Markus Borg and Per Runeson",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.; European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2024 ; Conference date: 09-09-2024 Through 13-09-2024",

year = "2024",

doi = "10.1007/978-3-031-70371-3_29",

language = "English",

isbn = "9783031703706",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "404--408",

editor = "Albert Bifet and Povilas Daniu{\v s}is and Jesse Davis and Tomas Krilavi{\v c}ius and Meelis Kull and Eirini Ntoutsi and Kai Puolam{\"a}ki and Indrė {\v Z}liobaitė",

booktitle = "Machine Learning and Knowledge Discovery in Databases. Research Track and Demo Track - European Conference, ECML PKDD 2024, Proceedings",

address = "Germany",

}

Paraschakis, D, Ros, R, Borg, M & Runeson, P 2024, FuseRank (Demo): Filtered Vector Search in Multimodal Structured Data. in A Bifet, P Daniušis, J Davis, T Krilavičius, M Kull, E Ntoutsi, K Puolamäki & I Žliobaitė (eds), Machine Learning and Knowledge Discovery in Databases. Research Track and Demo Track - European Conference, ECML PKDD 2024, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14948, Springer, pp. 404-408, European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2024, Vilnius, Lithuania, 2024/09/09. https://doi.org/10.1007/978-3-031-70371-3_29

FuseRank (Demo): Filtered Vector Search in Multimodal Structured Data. / Paraschakis, Dimitris; Ros, Rasmus; Borg, Markus et al.
Machine Learning and Knowledge Discovery in Databases. Research Track and Demo Track - European Conference, ECML PKDD 2024, Proceedings. ed. / Albert Bifet; Povilas Daniušis; Jesse Davis; Tomas Krilavičius; Meelis Kull; Eirini Ntoutsi; Kai Puolamäki; Indrė Žliobaitė. Springer, 2024. p. 404-408 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14948).

Research output: Chapter in Book/Report/Conference proceeding › Paper in conference proceeding › peer-review

TY - GEN

T1 - FuseRank (Demo)

T2 - European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2024

AU - Paraschakis, Dimitris

AU - Ros, Rasmus

AU - Borg, Markus

AU - Runeson, Per

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

PY - 2024

Y1 - 2024

N2 - We describe and demonstrate our work on multimodal filtered vector search in tabular data. It offers a practical way for businesses to vectorize their product assortments or any other business-critical data and then simultaneously retrieve and filter this information using state-of-the-art similarity search. Our methodology is based on the extended vector space model, with multiple modalities represented as sub-vectors that get concatenated and compared to the query vector via a dot product operation. It is a flexible framework that allows manipulating the influence of each modality on the overall item ranking via modality weights. We share the source code, the demonstration video, and the screenshot of the application. We also provide a brief description of its main building blocks, the supported data types, and modality filters. The application is bundled with two public datasets and pre-computed text embeddings so that it can be easily run without prior preparation.

AB - We describe and demonstrate our work on multimodal filtered vector search in tabular data. It offers a practical way for businesses to vectorize their product assortments or any other business-critical data and then simultaneously retrieve and filter this information using state-of-the-art similarity search. Our methodology is based on the extended vector space model, with multiple modalities represented as sub-vectors that get concatenated and compared to the query vector via a dot product operation. It is a flexible framework that allows manipulating the influence of each modality on the overall item ranking via modality weights. We share the source code, the demonstration video, and the screenshot of the application. We also provide a brief description of its main building blocks, the supported data types, and modality filters. The application is bundled with two public datasets and pre-computed text embeddings so that it can be easily run without prior preparation.

KW - Information retrieval

KW - Multimodal filtered vector search

U2 - 10.1007/978-3-031-70371-3_29

DO - 10.1007/978-3-031-70371-3_29

M3 - Paper in conference proceeding

AN - SCOPUS:85203877698

SN - 9783031703706

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 404

EP - 408

BT - Machine Learning and Knowledge Discovery in Databases. Research Track and Demo Track - European Conference, ECML PKDD 2024, Proceedings

A2 - Bifet, Albert

A2 - Daniušis, Povilas

A2 - Davis, Jesse

A2 - Krilavičius, Tomas

A2 - Kull, Meelis

A2 - Ntoutsi, Eirini

A2 - Puolamäki, Kai

A2 - Žliobaitė, Indrė

PB - Springer

Y2 - 9 September 2024 through 13 September 2024

ER -

Paraschakis D, Ros R, Borg M , Runeson P. FuseRank (Demo): Filtered Vector Search in Multimodal Structured Data. In Bifet A, Daniušis P, Davis J, Krilavičius T, Kull M, Ntoutsi E, Puolamäki K, Žliobaitė I, editors, Machine Learning and Knowledge Discovery in Databases. Research Track and Demo Track - European Conference, ECML PKDD 2024, Proceedings. Springer. 2024. p. 404-408. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-70371-3_29