Semi-Supervised Learning using Differentiable Reasoning
This addresses the problem of leveraging unlabeled data for semantic image interpretation, but it appears incremental as it builds on existing semi-supervised and reasoning methods.
The paper tackled the Semantic Image Interpretation task by introducing Differentiable Reasoning, a semi-supervised learning technique that uses relational background knowledge to improve performance with unlabeled data, showing significant improvement.
We introduce Differentiable Reasoning (DR), a novel semi-supervised learning technique which uses relational background knowledge to benefit from unlabeled data. We apply it to the Semantic Image Interpretation (SII) task and show that background knowledge provides significant improvement. We find that there is a strong but interesting imbalance between the contributions of updates from Modus Ponens (MP) and its logical equivalent Modus Tollens (MT) to the learning process, suggesting that our approach is very sensitive to a phenomenon called the Raven Paradox. We propose a solution to overcome this situation.