Open-world Semi-supervised Generalized Relation Discovery Aligned in a Real-world Setting
This work addresses the practicality of OpenRE for real-world applications by aligning it with data characteristics, though it is incremental in refining the problem setting.
The paper tackled the problem of Open-world Relation Extraction by proposing a more realistic setting where unlabeled data includes known and novel classes with hard negatives and long-tail relations, and introduced KNoRD to classify explicit and implicit relations, achieving significant performance gains on benchmarks.
Open-world Relation Extraction (OpenRE) has recently garnered significant attention. However, existing approaches tend to oversimplify the problem by assuming that all unlabeled texts belong to novel classes, thereby limiting the practicality of these methods. We argue that the OpenRE setting should be more aligned with the characteristics of real-world data. Specifically, we propose two key improvements: (a) unlabeled data should encompass known and novel classes, including hard-negative instances; and (b) the set of novel classes should represent long-tail relation types. Furthermore, we observe that popular relations such as titles and locations can often be implicitly inferred through specific patterns, while long-tail relations tend to be explicitly expressed in sentences. Motivated by these insights, we present a novel method called KNoRD (Known and Novel Relation Discovery), which effectively classifies explicitly and implicitly expressed relations from known and novel classes within unlabeled data. Experimental evaluations on several Open-world RE benchmarks demonstrate that KNoRD consistently outperforms other existing methods, achieving significant performance gains.