Finding Influential Instances for Distantly Supervised Relation Extraction
This work addresses label noise in relation extraction for NLP researchers, offering an interpretable and efficient solution, though it is incremental as it builds on existing influence function techniques.
The paper tackles the problem of label noise in distantly supervised relation extraction by proposing REIF, a model-agnostic instance sampling method based on influence functions, which reduces computational complexity from O(mn) to O(1) and outperforms baselines with complex architectures.
Distant supervision (DS) is a strong way to expand the datasets for enhancing relation extraction (RE) models but often suffers from high label noise. Current works based on attention, reinforcement learning, or GAN are black-box models so they neither provide meaningful interpretation of sample selection in DS nor stability on different domains. On the contrary, this work proposes a novel model-agnostic instance sampling method for DS by influence function (IF), namely REIF. Our method identifies favorable/unfavorable instances in the bag based on IF, then does dynamic instance sampling. We design a fast influence sampling algorithm that reduces the computational complexity from $\mathcal{O}(mn)$ to $\mathcal{O}(1)$, with analyzing its robustness on the selected sampling function. Experiments show that by simply sampling the favorable instances during training, REIF is able to win over a series of baselines that have complicated architectures. We also demonstrate that REIF can support interpretable instance selection.