Knowledge-augmented Column Networks: Guiding Deep Learning with Advice
This addresses the problem of data sparsity and noise in deep learning for domains with structured representations, offering a potential solution for practitioners in such fields.
The paper tackles the challenge of learning from sparse, noisy data in deep models by introducing Knowledge-augmented Column Networks, a relational framework that incorporates human advice to improve model performance.
Recently, deep models have had considerable success in several tasks, especially with low-level representations. However, effective learning from sparse noisy samples is a major challenge in most deep models, especially in domains with structured representations. Inspired by the proven success of human guided machine learning, we propose Knowledge-augmented Column Networks, a relational deep learning framework that leverages human advice/knowledge to learn better models in presence of sparsity and systematic noise.