MétaCan
Menu
Back to cohort

Colflor: Towards Bert-Size Vision-Language Document Retrieval Models

2025· article· W4415524320 on OpenAlexaff
Ahmed Masry, Enamul Hoque

Bibliographic record

Venuenot available
Typearticle
Language
FieldComputer Science
TopicNatural Language Processing Techniques
Canadian institutionsYork University
Fundersnot available
KeywordsDocument retrievalEncoding (memory)Process (computing)Image retrievalDocument clusteringVisual WordInfographicData retrievalVector space model

Abstract

fetched live from OpenAlex

Traditional document retrieval systems for PDFs, charts, and infographics rely heavily on Optical Character Recognition (OCR) pipelines to extract textual content, a process that is both error-prone and resource-intensive. Recent advancements in multimodal models like ColPali have enabled OCRfree retrieval by processing documents directly as images, but their large size (three billion parameters) makes them computationally expensive and impractical for large-scale applications. To address this limitation, we introduce ColFlor, an efficient OCR-free visual document retrieval model with only 174 million parameters. ColFlor achieves comparable performance to ColPali on text-rich English documents, with only a 1.8% decrease in performance (measured by NDCG@5 metric), while being significantly faster in image encoding (5.25 times faster) and query encoding (9.8 times faster). This makes OCR-free document retrieval systems more cost-effective for large-scale applications and more accessible to users with limited computational resources.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame machine prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.

metaresearch head score (Codex)0.002
metaresearch head score (Gemma)0.009
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: none
GenreCandidate signal: Methods · Consensus signal: Methods
Teacher disagreement score0.017
Threshold uncertainty score0.033

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0020.009
Meta-epidemiology (narrow)0.0020.001
Meta-epidemiology (broad)0.0020.002
Bibliometrics0.0030.002
Science and technology studies0.0010.001
Scholarly communication0.0030.005
Open science0.0040.002
Research integrity0.0030.002
Insufficient payload (model declined to judge)0.0060.007

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.009
GPT teacher head0.310
Teacher spread0.300 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designSimulation or modeling
Domainnot available
GenreMethods

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations2
Published2025
Admission routes1
Has abstractyes

Explore more

Same topicNatural Language Processing TechniquesFrench-language works237,207