Knowledge Graphs (KGs) have found many applications in industrial and in academic settings, which in turn, have motivated considerable research efforts towards large-scale information extraction from a variety of sources. Despite such efforts, it is well known that even the largest KGs suffer from incompleteness; Link Prediction (LP) techniques address this issue by identifying missing facts among entities already in the KG. Among the recent LP techniques, those based on KG embeddings have achieved very promising performance in some benchmarks. Despite the fast-growing literature on the subject, insufficient attention has been paid to the effect of the design choices in those methods. Moreover, the standard practice in this area is to report accuracy by aggregating over a large number of test facts in which some entities are vastly more represented than others; this allows LP methods to exhibit good results by just attending to structural properties that include such entities, while ignoring the remaining majority of the KG. This analysis provides a comprehensive comparison of embedding-based LP methods, extending the dimensions of analysis beyond what is commonly available in the literature. We experimentally compare the effectiveness and efficiency of 18 state-of-the-art methods, consider a rule-based baseline, and report detailed analysis over the most popular benchmarks in the literature.

No abstract

The task of Fine-grained Entity Type Classification (FETC) consists of assigning types from a hierarchy to entity mentions in text. Existing methods rely on distant supervision and are thus susceptible to noisy labels that can be out-of-context or overly-specific for the training sentence. Previous methods that attempt to address these issues do so with heuristics or with the help of hand-crafted features. Instead, we propose an end-to-end solution with a neural network model that uses a variant of crossentropy loss function to handle out-of-context labels, and hierarchical loss normalization to cope with overly-specific ones. Also, previous work solve FETC a multi-label classification followed by ad-hoc post-processing. In contrast, our solution is more elegant: we use public word embeddings to train a single-label that jointly learns representations for entity mentions and their context. We show experimentally that our approach is robust against noise and consistently outperforms the state-of-theart on established benchmarks for the task.

We

Entity Linking is the task of assigning entities from a Knowledge Base to textual mentions of such entities in a document. State-of-the-art approaches rely on lexical and statistical features which are abundant for popular entities but sparse for unpopular ones, resulting in a clear bias towards popular entities and poor accuracy for less popular ones. In this work, we present a novel approach that is guided by a natural notion of semantic similarity which is less amenable to such bias. We adopt a unified semantic representation for entities and documents-the probability distribution obtained from a random walk on a subgraph of the knowledge base-which can overcome the feature sparsity issue that affects previous work. Our algorithm continuously updates the semantic signature of the document as mentions are disambiguated, thus focusing the search based on context. Our experimental evaluation uses well-known benchmarks and different samples of a Wikipedia-based benchmark with varying entity popularity; the results illustrate well the bias of previous methods and the superiority of our approach, especially for the less popular entities.