Transformer-Based Astronomical Time Series Model with Uncertainty Estimation for Detecting Misclassified Instances
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866909376157057024 |
|---|---|
| author | Cádiz-Leyton, Martina Cabrera-Vives, Guillermo Protopapas, Pavlos Moreno-Cartagena, Daniel Donoso-Oliva, Cristobal |
| author_facet | Cádiz-Leyton, Martina Cabrera-Vives, Guillermo Protopapas, Pavlos Moreno-Cartagena, Daniel Donoso-Oliva, Cristobal |
| contents | In this work, we present a framework for estimating and evaluating uncertainty in deep-attention-based classifiers for light curves for variable stars. We implemented three techniques, Deep Ensembles (DEs), Monte Carlo Dropout (MCD) and Hierarchical Stochastic Attention (HSA) and evaluated models trained on three astronomical surveys. Our results demonstrate that MCD and HSA offers a competitive and computationally less expensive alternative to DE, allowing the training of transformers with the ability to estimate uncertainties for large-scale light curve datasets. We conclude that the quality of the uncertainty estimation is evaluated using the ROC AUC metric. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2411_01363 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Transformer-Based Astronomical Time Series Model with Uncertainty Estimation for Detecting Misclassified Instances Cádiz-Leyton, Martina Cabrera-Vives, Guillermo Protopapas, Pavlos Moreno-Cartagena, Daniel Donoso-Oliva, Cristobal Instrumentation and Methods for Astrophysics In this work, we present a framework for estimating and evaluating uncertainty in deep-attention-based classifiers for light curves for variable stars. We implemented three techniques, Deep Ensembles (DEs), Monte Carlo Dropout (MCD) and Hierarchical Stochastic Attention (HSA) and evaluated models trained on three astronomical surveys. Our results demonstrate that MCD and HSA offers a competitive and computationally less expensive alternative to DE, allowing the training of transformers with the ability to estimate uncertainties for large-scale light curve datasets. We conclude that the quality of the uncertainty estimation is evaluated using the ROC AUC metric. |
| title | Transformer-Based Astronomical Time Series Model with Uncertainty Estimation for Detecting Misclassified Instances |
| topic | Instrumentation and Methods for Astrophysics |
| url | https://arxiv.org/abs/2411.01363 |