aiqbench analysis
This research addresses the accuracy-performance trade-off in out-of-distribution detection by optimizing model retention, data sampling, and representation learning. The proposed framework successfully mitigates classification degradation while enhancing robustness across diverse benchmarks, including long-tailed scenarios. It establishes a practical baseline for deploying reliable AI systems that can effectively distinguish between known and unknown data distributions.
Optimizing Three Critical Factors for Practical and Effective OOD Detection Fine-Tuning
arXiv:2308.01030v2 Announce Type: replace-cross Abstract: In out-of-distribution (OOD) detection, fine-tuning with auxiliary outlier data often improves detection performance at the cost of classification accuracy. This trade-off stems from the loss of the original in-distribution (ID) distribution during fine-tuning. To establish a more practical and effective paradigm, we optimize three critical factors: model reminder, data sampling, and representation learning. We propose: (1) Self-Knowledge Distillation (SKD) to mitigate accuracy reduction; (2) Semi-hard Outlier Sampling (SOS) to improve detection efficiency with minimal data; and (3) Outlier-aware Supervised Contrastive Learning (OSCL) to promote ID-OOD separability. Optimizing these factors produces cumulative gains, boosting both OOD detection performance and classification accuracy. Our framework outperforms existing methods across diverse benchmarks, particularly in long-tailed scenarios, providing a robust baseline for real-world OOD detection.
Abstract:In out-of-distribution (OOD) detection, fine-tuning with auxiliary outlier data often improves detection performance at the cost of classification accuracy. This trade-off stems from the loss of the original in-distribution (ID) distribution during fine-tuning. To establish a more practical and effective paradigm, we optimize three critical factors: model reminder, data sampling, and representation learning. We propose: (1) Self-Knowledge Distillation (SKD) to mitigate accuracy reduction; (2) Semi-hard Outlier Sampling (SOS) to improve detection efficiency with minimal data; and (3) Outlier-aware Supervised Contrastive Learning (OSCL) to promote ID-OOD separability. Optimizing these factors produces cumulative gains, boosting both OOD detection performance and classification accuracy. Our framework outperforms existing methods across diverse benchmarks, particularly in long-tailed scenarios, providing a robust baseline for real-world OOD detection.
Accepted at ICPR 2026. Code: this https URL
Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
arXiv:2308.01030 [cs.LG]
(or arXiv:2308.01030v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2308.01030 Focus to learn more arXiv-issued DOI via DataCite
ICPR 2026, LNCS 16825, pp. 429-446, Springer, 2026
https://doi.org/10.1007/978-3-032-31930-2_29 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 247 words
Related coverage
State of Pandemic Early Warning
LessWrong · Sep 24, 2026
Improving CoT Monitorability of Evaluation Awareness via Verbalization Training
LessWrong · Oct 2, 2026
Concrete Evaluations to Investigate the OpenAI Model That Hacked Hugging Face
AI Alignment Forum · Aug 3, 2026
Building Blocks for Foundation Model Training and Inference on AWS
Hugging Face · May 11, 2026