Vendi Information Gain for Active Learning and its Application to Ecology

2025-09-12

AI Summary

To address the scarcity of labeled data in camera-trap image species identification, this paper proposes a dataset-level active learning method leveraging predictive uncertainty. Unlike conventional approaches that rely solely on per-sample uncertainty, our method introduces Vendi information gain—a global distribution-aware metric—to quantify how candidate samples collectively reduce overall dataset uncertainty. This enables joint optimization of informativeness and diversity during batch sampling. By integrating deep feature representations with Vendi score evaluation, the method achieves efficient and robust annotation selection. On the Snapshot Serengeti benchmark, it attains near fully supervised accuracy using less than 10% of the labels, consistently outperforming state-of-the-art baselines across multiple evaluation metrics and batch-size configurations. The approach establishes a scalable new paradigm for biodiversity monitoring under extreme label scarcity.

Technology Category

• Machine Learning: Active Learning
• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Data transparency and provenance
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

While monitoring biodiversity through camera traps has become an important endeavor for ecological research, identifying species in the captured image data remains a major bottleneck due to limited labeling resources. Active learning -- a machine learning paradigm that selects the most informative data to label and train a predictive model -- offers a promising solution, but typically focuses on uncertainty in the individual predictions without considering uncertainty across the entire dataset. We introduce a new active learning policy, Vendi information gain (VIG), that selects images based on their impact on dataset-wide prediction uncertainty, capturing both informativeness and diversity. Applied to the Snapshot Serengeti dataset, VIG achieves impressive predictive accuracy close to full supervision using less than 10% of the labels. It consistently outperforms standard baselines across metrics and batch sizes, collecting more diverse data in the feature space. VIG has broad applicability beyond ecology, and our results highlight its value for biodiversity monitoring in data-limited environments.

Problem

• Active learning for species identification in camera trap images
• Addressing dataset-wide uncertainty beyond individual prediction uncertainty
• Improving biodiversity monitoring with limited labeling resources

Innovation

• Vendi information gain for dataset-wide uncertainty
• Selects images based on informativeness and diversity
• Achieves high accuracy with minimal labeled data