Do We Need More Training Data?

Abstract

Datasets for training object recognition systems are steadily increasing in size. This paper investigates the question of whether existing detectors will continue to improve as data grows, or saturate in performance due to limited model complexity and the Bayes risk associated with the feature spaces in which they operate. We focus on the popular paradigm of discriminatively trained templates defined on oriented gradient features. We investigate the performance of mixtures of templates as the number of mixture components and the amount of training data grows. Surprisingly, even with proper treatment of regularization and "outliers", the performance of classic mixture models appears to saturate quickly ($\sim$10 templates and $\sim$100 positive training examples per template). This is not a limitation of the feature space as compositional mixtures that share template parameters via parts and that can synthesize new templates not encountered during training yield significantly better performance. Based on our analysis, we conjecture that the greatest gains in detection performance will continue to derive from improved representations and learning algorithms that can make efficient use of large datasets.

Cited by in corpus (16)

• A Survey on Instance Segmentation: State of the art
• Predicting and Mapping of Soil Organic Carbon Using Machine Learning Algorithms in Northern Iran
• Multi-level Semantic Feature Augmentation for One-shot Learning
• BVI-DVC: A Training Database for Deep Video Compression
• A Two-Stage Approach to Few-Shot Learning for Image Recognition
• A Machine Learning Based Morphological Classification of 14,245 Radio AGNs Selected From The Best-Heckman Sample
• Challenging Machine Learning-based Clone Detectors via Semantic-preserving Code Transformations
• A Pilot Study For Fragment Identification Using 2D NMR and Deep Learning
• Vulnerabilities of Connectionist AI Applications: Evaluation and Defence
• Structured LISTA for Multidimensional Harmonic Retrieval
• Defectors: A Large, Diverse Python Dataset for Defect Prediction
• Not Hydro: Using Neural Networks to estimate galaxy properties on a Dark-Matter-Only simulation
• Designing deep neural networks for driver intention recognition
• SATDAUG -- A Balanced and Augmented Dataset for Detecting Self-Admitted Technical Debt
• Demystifying Ten Big Ideas and Rules Every Fire Scientist & Engineer Should Know About Blackbox, Whitebox & Causal Artificial Intelligence
• Collate: Collaborative Neural Network Learning for Latency-Critical Edge Systems