Challenge: Annotator disagreements are resolved before learning takes place, but researchers question the performance of a system when annotators disagree.
Approach: They propose a method that uses language features and label distributions to pool similar items into larger labels.
Outcome: The proposed method is based on five publicly available datasets with varying levels of disagreements on social media and in the wild using a dataset from Facebook.

Similar Papers

Capturing Perspectives of Crowdsourced Annotators in Subjective Learning Tasks (2024.naacl-long)

Copied to clipboard

Challenge: Existing approaches to label aggregation fail to capture subjective annotations and can lead to biases.
Approach: They propose annotator-aware representations for text for subjective classification tasks that involve learning representations of annotators.
Outcome: The proposed model improves on metrics that assess the performance on capturing individual annotators’ perspectives.
Agreeing to Disagree: Annotating Offensive Language Datasets with Annotators’ Disagreement (2021.emnlp-main)

Copied to clipboard

Challenge: supervised learning is a key component of offensive language detection, but there is little attention given to the quality of annotated data.
Approach: They propose to examine the level of agreement among annotators while selecting data to create offensive language datasets, a task involving a high level of subjectivity.
Outcome: The proposed datasets show that annotators' agreement has a strong effect on classifiers performance and robustness.
When the Majority is Wrong: Modeling Annotator Disagreement for Subjective Tasks (2023.emnlp-main)

Copied to clipboard

Challenge: a number of studies have questioned assumptions of majority vote aggregated labels.
Approach: They construct a model that predicts individual annotator ratings on potentially offensive text and combines this information with the predicted target group of the text to predict the ratings of target group members.
Outcome: The proposed model raises performance over baseline by 22% and 33% at predicting variance among annotators.
Voices in a Crowd: Searching for clusters of unique perspectives (2024.emnlp-main)

Copied to clipboard

Challenge: Proposed solutions aim to capture minority perspectives by either modelling annotator disagreements or grouping annotators based on shared metadata.
Approach: They propose a framework that trains models without encoding annotator metadata and creates clusters of similar opinions, that are called voices.
Outcome: The proposed framework captures minority perspectives based on demographic factors in two distinct datasets while also capturing majority perspectives.
Learning from Measurements in Crowdsourcing Models: Inferring Ground Truth from Diverse Annotation Types (C18-1)

Copied to clipboard

Challenge: Annotated corpora are often assigned to internet workers whose judgments are reconciled by crowdsourcing models.
Approach: They propose a framework for learning from rich prior knowledge to combine annotations with different structures.
Outcome: The proposed model compares favorably with previous work and enables active sample selection to reduce annotation effort.
Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotations (2022.tacl-1)

Copied to clipboard

Challenge: Annotators may systematically disagree with one another, reflecting their individual biases and values, especially in the case of subjective tasks such as detecting affect, aggression, and hate speech.
Approach: They propose to combine multi-annotator models with multi-task based approaches to resolve disagreements between annotations and derive single ground truth labels.
Outcome: The proposed model outperforms majority voting and averaging methods and estimates uncertainty in predictions.
Disagreement Matters: Preserving Label Diversity by Jointly Modeling Item and Annotator Label Distributions with DisCo (2023.findings-acl)

Copied to clipboard

Challenge: a recent study shows that annotator disagreement is common in supervised learning . a simple neural model that learns to predict annotators' labels is competitive with other models that do not model specific annotations.
Approach: They propose a neural model that learns to predict annotator distributions by aggregating over all annotators.
Outcome: The proposed model outperforms models that do not model specific annotators or do not learn label distribution learning.
You Are What You Annotate: Towards Better Models through Annotator Representations (2023.findings-emnlp)

Copied to clipboard

Challenge: Annotator disagreement is ubiquitous in natural language processing tasks.
Approach: They propose to model annotators' idiosyncrasies and account for their idioms by creating representations for each annotator and their annotations.
Outcome: The proposed model improves on an existing dataset with eight annotators with inherent disagreements while increasing model size by 1%.
Are We Modeling the Task or the Annotator? An Investigation of Annotator Bias in Natural Language Understanding Datasets (D19-1)

Copied to clipboard

Challenge: Having only a few workers generate the majority of dataset examples raises concerns about data diversity .
Approach: They perform a series of experiments to investigate annotator biases in recent NLU datasets . they find that models are able to recognize the most productive annotators .
Outcome: The results show that models can recognize the most productive annotators and do not generalize well to examples from annotator that did not contribute to the training set.
NUTMEG: Separating Signal From Noise in Annotator Disagreement (2025.emnlp-main)

Copied to clipboard

Challenge: Recent work suggests that annotators may have genuine disagreements, but few models separate signal from noise in annotator disagreement.
Approach: They propose a Bayesian model that removes noisy annotations from training data while preserving systematic disagreements.
Outcome: The proposed model outperforms models trained on NUTMEG-aggregated data.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations