Abstract:Loss functions and sample mining strategies are essential components in deep metric learning algorithms. However, the existing loss function or mining strategy often necessitate the incorporation of additional hyperparameters, notably the threshold, which defines whether the sample pair is informative. The threshold provides a stable numerical standard for determining whether to retain the pairs. It is a vital parameter to reduce the redundant sample pairs participating in training. Nonetheless, finding the optimal threshold can be a time-consuming endeavor, often requiring extensive grid searches. Because the threshold cannot be dynamically adjusted in the training stage, we should conduct plenty of repeated experiments to determine the threshold. Therefore, we introduce a novel approach for adjusting the thresholds associated with both the loss function and the sample mining strategy. We design a static Asymmetric Sample Mining Strategy (ASMS) and its dynamic version Adaptive Tolerance ASMS (AT-ASMS), tailored for sample mining methods. ASMS utilizes differentiated thresholds to address the problems (too few positive pairs and too many redundant negative pairs) caused by only applying a single threshold to filter samples. AT-ASMS can adaptively regulate the ratio of positive and negative pairs during training according to the ratio of the currently mined positive and negative pairs. This meta-learning-based threshold generation algorithm utilizes a single-step gradient descent to obtain new thresholds. We combine these two threshold adjustment algorithms to form the Dual Dynamic Threshold Adjustment Strategy (DDTAS). Experimental results show that our algorithm achieves competitive performance on CUB200, Cars196, and SOP datasets. Our codes are available at https://github.com/NUST-Machine-Intelligence-Laboratory/DDTAS.

A THRESHOLD ADJUSTING ALGORITHM FOR TEXT FILTERING

Threshold Optimization with a Small Number of Samples in Adaptive Information Filtering

Threshold-Based Resampling for High-Speed Particle Phd Filter

Particle Filter Algorithm Based on Adaptive Resampling Strategy

A Video Text Detection Approach Based on Adaptive Thresholds

Answering unique topic queries with dynamic threshold

Adaptive Weight Mirrored Threshold Stack Filters

A dual-filters echo canceller using adaptive threshold to handle double talk

Threshold Selection in Feature Screening for Error Rate Control*

Automatic Determination of Optimal Fault Detection Filter

A Text Filter Method against Chinese Vicious-topic Information

Improved Boosting Algorithm With Adaptive Filtration

A Non-Contact Speech Enhancement Algorithm Based on Wavelet Packet Adaptive Threshold

Optimization of Stack Filters Based on Mirrored Threshold Decomposition

A Pulse Wave Filter Method Based On Wavelet Transform Soft-Threshold And Adaptive Algorithm

New Adaptive Threshold SSDA Algorithm

Study on Filtering Algorithm of Image Using Matlab

A New Data Filter Algorithm

Text Stream Clustering Algorithm Based on Adaptive Feature Selection.

Dual Dynamic Threshold Adjustment Strategy

An Accurate Threshold Insensitive Kernel Detector for Arbitrary Shaped Text.