Technology Category
Machine Learning: Calibration & Uncertainty QuantificationComputer Vision: Adversarial Attacks & RobustnessNatural Language Processing: Safety and Robustness
Prior Distribution and Model Confidence
🤖 AI Summary
This work investigates how training data distribution affects the generalization of image classification models and proposes a model-agnostic, retraining-free confidence assessment framework. The method fuses multiple embeddings into a unified embedding space to jointly characterize the structure of the training distribution; it then employs an adaptive distance metric to quantify each sample’s deviation from this distribution, enabling confidence calibration, low-confidence prediction filtering, and out-of-distribution (OOD) detection. The framework is architecture-agnostic and domain-transferable, delivering consistent improvements across diverse backbones—including ResNet and Vision Transformers—without architectural modification or fine-tuning. Notably, accuracy gains are especially pronounced after filtering low-confidence predictions. Empirically, it enhances classification robustness and reliability under distribution shifts, offering a lightweight, plug-and-play, distribution-aware inference mechanism for trustworthy AI systems.
Technology Category
Application Category
📝 Abstract
This paper investigates the impact of training data distribution on the performance of image classification models. By analyzing the embeddings of the training set, we propose a framework to understand the confidence of model predictions on unseen data without the need for retraining. Our approach filters out low-confidence predictions based on their distance from the training distribution in the embedding space, significantly improving classification accuracy. We demonstrate this on the example of several classification models, showing consistent performance gains across architectures. Furthermore, we show that using multiple embedding models to represent the training data enables a more robust estimation of confidence, as different embeddings capture complementary aspects of the data. Combining these embeddings allows for better detection and exclusion of out-of-distribution samples, resulting in further accuracy improvements. The proposed method is model-agnostic and generalizable, with potential applications beyond computer vision, including domains such as Natural Language Processing where prediction reliability is critical.
Problem
Research questions and friction points this paper is trying to address.
Impact of training data distribution on model performance
Framework for model confidence without retraining
Filtering low-confidence predictions using embedding distance
Innovation
Methods, ideas, or system contributions that make the work stand out.
Uses embedding distance to filter low-confidence predictions
Combines multiple embedding models for robust confidence estimation
Model-agnostic framework improves accuracy without retraining
🔎 Similar Papers
No similar papers found.
