MétaCan
Menu
Back to cohort

Analyzing and Mitigating Dataset Artifacts in Natural Language Inference Models Using ELECTRA

2024· article· en· W4405702407 on OpenAlexaff
Himanshu Joshi

Bibliographic record

VenueInternational Journal of Multidisciplinary Research and Growth Evaluation · 2024
Typearticle
Languageen
FieldComputer Science
TopicTopic Modeling
Canadian institutionsArtificial Intelligence in Medicine (Canada)
Fundersnot available
KeywordsInferenceComputer scienceArtificial intelligenceRobustness (evolution)WeightingRegularization (linguistics)Adversarial systemGeneralizationNatural language understandingArtifact (error)Machine learningNatural language processingNatural languageMathematics

Abstract

fetched live from OpenAlex

This paper investigates the challenges posed by dataset artifacts in Natural Language Inference (NLI) models, focusing on ELECTRA, a state-of-the-art transformer model. Dataset artifacts such as hypothesis-only biases, lexical overlap issues, and frequent label imbalances significantly impact model generalization, leading to erroneous predictions. We propose and evaluate a range of strategies, including adversarial training, data augmentation, instance weighting, and artifact-aware regularization, to mitigate these issues. Extensive experimental results demonstrate up to a 6% improvement in robustness and generalization, providing valuable insights for creating artifact-resistant NLP models.

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 distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.004
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation 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: Empirical
Teacher disagreement score0.916
Threshold uncertainty score0.429

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0040.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0010.000
Science and technology studies0.0000.000
Scholarly communication0.0000.002
Open science0.0000.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

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 head0.148
GPT teacher head0.469
Teacher spread0.321 · 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 teacher head, 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

Citations0
Published2024
Admission routes1
Has abstractyes

Explore more

Same venueInternational Journal of Multidisciplinary Research and Growth EvaluationSame topicTopic ModelingFrench-language works237,207