LPI-RIT at LeWiDi-2025: Improving Distributional Predictions via Metadata and Loss Reweighting with DisCo

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sawkar, Mandira, Shetty, Samay U., Pandita, Deepak, Weerasooriya, Tharindu Cyril, Homan, Christopher M.
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914074718109696
author Sawkar, Mandira
Shetty, Samay U.
Pandita, Deepak
Weerasooriya, Tharindu Cyril
Homan, Christopher M.
author_facet Sawkar, Mandira
Shetty, Samay U.
Pandita, Deepak
Weerasooriya, Tharindu Cyril
Homan, Christopher M.
contents The Learning With Disagreements (LeWiDi) 2025 shared task aims to model annotator disagreement through soft label distribution prediction and perspectivist evaluation, which focuses on modeling individual annotators. We adapt DisCo (Distribution from Context), a neural architecture that jointly models item-level and annotator-level label distributions, and present detailed analysis and improvements. In this paper, we extend DisCo by introducing annotator metadata embeddings, enhancing input representations, and multi-objective training losses to capture disagreement patterns better. Through extensive experiments, we demonstrate substantial improvements in both soft and perspectivist evaluation metrics across three datasets. We also conduct in-depth calibration and error analyses that reveal when and why disagreement-aware modeling improves. Our findings show that disagreement can be better captured by conditioning on annotator demographics and by optimizing directly for distributional metrics, yielding consistent improvements across datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2508_08163
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LPI-RIT at LeWiDi-2025: Improving Distributional Predictions via Metadata and Loss Reweighting with DisCo
Sawkar, Mandira
Shetty, Samay U.
Pandita, Deepak
Weerasooriya, Tharindu Cyril
Homan, Christopher M.
Computation and Language
Artificial Intelligence
Machine Learning
The Learning With Disagreements (LeWiDi) 2025 shared task aims to model annotator disagreement through soft label distribution prediction and perspectivist evaluation, which focuses on modeling individual annotators. We adapt DisCo (Distribution from Context), a neural architecture that jointly models item-level and annotator-level label distributions, and present detailed analysis and improvements. In this paper, we extend DisCo by introducing annotator metadata embeddings, enhancing input representations, and multi-objective training losses to capture disagreement patterns better. Through extensive experiments, we demonstrate substantial improvements in both soft and perspectivist evaluation metrics across three datasets. We also conduct in-depth calibration and error analyses that reveal when and why disagreement-aware modeling improves. Our findings show that disagreement can be better captured by conditioning on annotator demographics and by optimizing directly for distributional metrics, yielding consistent improvements across datasets.
title LPI-RIT at LeWiDi-2025: Improving Distributional Predictions via Metadata and Loss Reweighting with DisCo
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2508.08163