Imagen de Portada

Gatekeeper: Improving Model Cascades Through Confidence Tuning

Fuente: arXiv

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:	Machine Learning
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Cascade-Aware Training of Language Models
por: Wang, Congchao, et al.
Publicado: (2024)

BRAVE: Broadening the visual encoding of vision-language models
por: Kar, Oğuzhan Fatih, et al.
Publicado: (2024)

When Does Confidence-Based Cascade Deferral Suffice?
por: Jitkrittum, Wittawat, et al.
Publicado: (2023)

Test-Time Visual In-Context Tuning
por: Xie, Jiahao, et al.
Publicado: (2025)

Reading Between the Lines: Abstaining from VLM-Generated OCR Errors via Latent Representation Probes
por: Yao, Jihan, et al.
Publicado: (2025)

Language Model Cascades: Token-level uncertainty and beyond
por: Gupta, Neha, et al.
Publicado: (2024)

Extracting Training Data from Document-Based VQA Models
por: Pinto, Francesco, et al.
Publicado: (2024)

Uncertainty-Driven Reliability: Selective Prediction and Trustworthy Deployment in Modern Machine Learning
por: Rabanser, Stephan
Publicado: (2025)

Faster Cascades via Speculative Decoding
por: Narasimhan, Harikrishna, et al.
Publicado: (2024)

What Does It Take to Build a Performant Selective Classifier?
por: Rabanser, Stephan, et al.
Publicado: (2025)

Universal Model Routing for Efficient LLM Inference
por: Jitkrittum, Wittawat, et al.
Publicado: (2025)

Cost-Aware Routing for Efficient Text-To-Image Generation
por: Li, Qinchan, et al.
Publicado: (2025)

Omnia de EgoTempo: Benchmarking Temporal Understanding of Multi-Modal LLMs in Egocentric Videos
por: Plizzari, Chiara, et al.
Publicado: (2025)

Suitability Filter: A Statistical Framework for Classifier Evaluation in Real-World Deployment Settings
por: Pouget, Angéline, et al.
Publicado: (2025)

Bi-directional Model Cascading with Proxy Confidence
por: Warren, David, et al.
Publicado: (2025)

SSL-R1: Self-Supervised Visual Reinforcement Post-Training for Multimodal Large Language Models
por: Xie, Jiahao, et al.
Publicado: (2026)

R-CoV: Region-Aware Chain-of-Verification for Alleviating Object Hallucinations in LVLMs
por: Xie, Jiahao, et al.
Publicado: (2026)

Hyperbolic Delaunay Geometric Alignment
por: Medbouhi, Aniss Aiman, et al.
Publicado: (2024)

Confidential Guardian: Cryptographically Prohibiting the Abuse of Model Abstention
por: Rabanser, Stephan, et al.
Publicado: (2025)

Re-Nerfing: Improving Novel View Synthesis through Novel View Synthesis
por: Tristram, Felix, et al.
Publicado: (2023)

HybridServe: Efficient Serving of Large AI Models with Confidence-Based Cascade Routing
por: Xue, Leyang, et al.
Publicado: (2025)

Bipartite Ranking From Multiple Labels: On Loss Versus Label Aggregation
por: Lukasik, Michal, et al.
Publicado: (2025)

Towards a Science of AI Agent Reliability
por: Rabanser, Stephan, et al.
Publicado: (2026)

Show Your Work with Confidence: Confidence Bands for Tuning Curves
por: Lourie, Nicholas, et al.
Publicado: (2023)

Self-supervised Latent Space Optimization with Nebula Variational Coding
por: Wang, Yida, et al.
Publicado: (2025)

Selective Prediction via Training Dynamics
por: Rabanser, Stephan, et al.
Publicado: (2022)

CO-MAP: A Reinforcement Learning Approach to the Qubit Allocation Problem
por: Kulshrestha, Ankit, et al.
Publicado: (2026)

Rational Tuning of LLM Cascades via Probabilistic Modeling
por: Zellinger, Michael J., et al.
Publicado: (2025)

Investigating the Impact of Model Instability on Explanations and Uncertainty
por: Marjanović, Sara Vera, et al.
Publicado: (2024)

Rethinking Fine-Tuning when Scaling Test-Time Compute: Limiting Confidence Improves Mathematical Reasoning
por: Chen, Feng, et al.
Publicado: (2025)

Contrastive Test-Time Composition of Multiple LoRA Models for Image Generation
por: Meral, Tuna Han Salih, et al.
Publicado: (2024)

TokenFormer: Rethinking Transformer Scaling with Tokenized Model Parameters
por: Wang, Haiyang, et al.
Publicado: (2024)

DrJAX: Scalable and Differentiable MapReduce Primitives in JAX
por: Rush, Keith, et al.
Publicado: (2024)

Reducing Reward Dependence in RL Through Adaptive Confidence Discounting
por: Satici, Muhammed Yusuf, et al.
Publicado: (2025)

Improved Online Confidence Bounds for Multinomial Logistic Bandits
por: Lee, Joongkyu, et al.
Publicado: (2025)

Maximizing Confidence Alone Improves Reasoning
por: Prabhudesai, Mihir, et al.
Publicado: (2025)

Denoising Diffusion via Image-Based Rendering
por: Anciukevičius, Titas, et al.
Publicado: (2024)

A Little Help Goes a Long Way: Efficient LLM Training by Leveraging Small LMs
por: Rawat, Ankit Singh, et al.
Publicado: (2024)

Confidence Is All You Need: Few-Shot RL Fine-Tuning of Language Models
por: Li, Pengyi, et al.
Publicado: (2025)

VLAD-Grasp: Zero-shot Grasp Detection via Vision-Language Models
por: Kulshrestha, Manav, et al.
Publicado: (2025)