Guardado en:
Detalles Bibliográficos
Autores principales: Rabanser, Stephan, Rauschmayr, Nathalie, Kulshrestha, Achin, Poklukar, Petra, Jitkrittum, Wittawat, Augenstein, Sean, Wang, Congchao, Tombari, Federico
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2502.19335
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917034588110848
author Rabanser, Stephan
Rauschmayr, Nathalie
Kulshrestha, Achin
Poklukar, Petra
Jitkrittum, Wittawat
Augenstein, Sean
Wang, Congchao
Tombari, Federico
author_facet Rabanser, Stephan
Rauschmayr, Nathalie
Kulshrestha, Achin
Poklukar, Petra
Jitkrittum, Wittawat
Augenstein, Sean
Wang, Congchao
Tombari, Federico
contents Large-scale machine learning models deliver strong performance across a wide range of tasks but come with significant computational and resource constraints. To mitigate these challenges, local smaller models are often deployed alongside larger models, relying on routing and deferral mechanisms to offload complex tasks. However, existing approaches inadequately balance the capabilities of these models, often resulting in unnecessary deferrals or sub-optimal resource usage. In this work we introduce a novel loss function called Gatekeeper for calibrating smaller models in cascade setups. Our approach fine-tunes the smaller model to confidently handle tasks it can perform correctly while deferring complex tasks to the larger model. Moreover, it incorporates a mechanism for managing the trade-off between model performance and deferral accuracy, and is broadly applicable across various tasks and domains without any architectural changes. We evaluate our method on encoder-only, decoder-only, and encoder-decoder architectures. Experiments across image classification, language modeling, and vision-language tasks show that our approach substantially improves deferral performance.
format Preprint
id arxiv_https___arxiv_org_abs_2502_19335
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Gatekeeper: Improving Model Cascades Through Confidence Tuning
Rabanser, Stephan
Rauschmayr, Nathalie
Kulshrestha, Achin
Poklukar, Petra
Jitkrittum, Wittawat
Augenstein, Sean
Wang, Congchao
Tombari, Federico
Machine Learning
Large-scale machine learning models deliver strong performance across a wide range of tasks but come with significant computational and resource constraints. To mitigate these challenges, local smaller models are often deployed alongside larger models, relying on routing and deferral mechanisms to offload complex tasks. However, existing approaches inadequately balance the capabilities of these models, often resulting in unnecessary deferrals or sub-optimal resource usage. In this work we introduce a novel loss function called Gatekeeper for calibrating smaller models in cascade setups. Our approach fine-tunes the smaller model to confidently handle tasks it can perform correctly while deferring complex tasks to the larger model. Moreover, it incorporates a mechanism for managing the trade-off between model performance and deferral accuracy, and is broadly applicable across various tasks and domains without any architectural changes. We evaluate our method on encoder-only, decoder-only, and encoder-decoder architectures. Experiments across image classification, language modeling, and vision-language tasks show that our approach substantially improves deferral performance.
title Gatekeeper: Improving Model Cascades Through Confidence Tuning
topic Machine Learning
url https://arxiv.org/abs/2502.19335