MétaCan
Menu
Back to cohort
Record W4379521537 · doi:10.21428/594757db.62395a61

Towards Improving Text Classification Tasks Based on Knowledge Graphs for Limited Labeled Data

2023· article· en· W4379521537 on OpenAlexaff
Hongzhi Zhang, Omair Shafiq

Bibliographic record

Venuenot available
Typearticle
Languageen
FieldComputer Science
TopicTopic Modeling
Canadian institutionsCarleton University
Fundersnot available
KeywordsComputer scienceKnowledge graphTransformerArtificial intelligenceDomain knowledgeMachine learningLanguage modelGraphTraining setLabeled dataNatural language processingProcess (computing)Data miningTheoretical computer science

Abstract

fetched live from OpenAlex

Pre-trained Transformer models have become popular in various Natural Language Processing (NLP) tasks, following a two-step process of ’pre-training’ and ’fine-tuning’. However, with the abundance of information on the web, some domain knowledge may be lacking. This can result in poor performance during the fine-tuning step when there is limited training data available. To address this issue in the case of limited data, we propose a knowledge graph-based data expansion method that enables the model to achieve good results even when there is limited data in the fine-tuning step. We extract entities in the text through Named Entity Recognition and then search for related information in the knowledge graph to expand the text’s content. This allows the pretrained model to acquire more external knowledge and enhance its training. We used our data expansion method to conduct experiments on the following well-known models, i.e., BERT, RoBERTa, and GPT-3. Our experiments show that our approach can improve the accuracy of language models on text classification tasks when training data is limited.

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)0.003
metaresearch head score (Gemma)0.011
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: Empirical · Consensus signal: none
Teacher disagreement score0.011
Threshold uncertainty score0.023

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0030.011
Meta-epidemiology (narrow)0.0020.001
Meta-epidemiology (broad)0.0010.002
Bibliometrics0.0040.003
Science and technology studies0.0010.001
Scholarly communication0.0020.008
Open science0.0030.002
Research integrity0.0020.003
Insufficient payload (model declined to judge)0.0020.002

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 head0.139
GPT teacher head0.331
Teacher spread0.192 · 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
GenreEmpirical

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

Citations1
Published2023
Admission routes1
Has abstractyes

Explore more

Same topicTopic ModelingFrench-language works237,207