Fine-grained entity typing results can serve as important information for entities while constructing knowledge bases. It is a challenging task due to the use of large tag sets and the requirement of understanding the context.
We find that, in some cases, existing neural fine-grained entity typing models may ignore the semantic information in the context that is important for typing. 
To address this problem, we propose to exploit semantic relations extracted from the sentence to improve the use of context. The used semantic relations are mainly those that are between the mention and the other words or phrases in the sentence. 
We investigate the use of two types of semantic relations: hypernym relation, and verb-argument relation. Our approach combine the predictions made based on different semantic relations and the predictions of a base neural model to produce the final results. We conduct experiments on two commonly used datasets: FIGER (GOLD) and BBN. Our approach achieves at least 2\% absolute strict accuracy improvement on both datasets compared with a strong BERT based model.

Exploiting Semantic Relations for Fine-grained Entity Typing

Representing knowledge graphs (KGs) by learning embeddings for entities and relations has led to accurate models for existing KG completion benchmarks. However, due to the open-world assumption of existing KGs, evaluation of KG completion uses ranking metrics and triple classification with negative samples, and is thus unable to directly assess models on the goals of the task: completion. In this paper, we first study the shortcomings of these evaluation metrics. Specifically, we demonstrate that these metrics (1) are unreliable for estimating how calibrated the models are, (2) make strong assumptions that are often violated, and 3) do not sufficiently, and consistently, differentiate embedding methods from each other, or from simpler approaches. To address these issues, we gather a semi-complete KG referred as YAGO3-TC, using a random subgraph from the test and validation data of YAGO3-10, which enables us to compute accurate triple classification accuracy on this data.  Conducting thorough experiments on existing models, we provide new insights and directions for the KG completion research. Along with the dataset and the open source implementation of the models, we also provide a leaderboard for knowledge graph completion that consists of a hidden, and growing, test set, available at https://pouyapez.github.io/yago3-tc/.

Revisiting Evaluation of Knowledge Base Completion Models

Knowledge bases (KBs), pragmatic collections of knowledge about notable entities, are an important asset in applications such as search, question answering and dialogue. Rooted in a long tradition in knowledge representation, all popular KBs only store positive information, but abstain from taking any stance towards statements not contained in them.

In this paper, we make the case for explicitly stating interesting statements which are not true. Negative statements would be important to overcome current limitations of question answering, yet due to their potential abundance, any effort towards compiling them needs a tight coupling with ranking. We introduce two approaches towards automatically compiling negative statements. (i) In peer-based statistical inferences, we compare entities with highly related entities in order to derive potential negative statements, which we then rank using supervised and unsupervised features. (ii) In pattern-based query log extraction, we use a pattern-based approach for harvesting search engine query logs. Experimental results show that both approaches hold promising and complementary potential. Along with this paper, we publish the first datasets on interesting negative information, containing over 1.4M statements for 130K popular Wikidata entities.

Enriching Knowledge Bases with Interesting Negative Statements

Commonsense knowledge (CSK) supports a variety of AI applications, from visual understanding to chatbots. Prior works on acquiring CSK, such as ConceptNet, have compiled statements that associate concepts with properties that hold for most or some of their instances. Each concept and statement is treated in isolation from others, and the only quantitative measure (or ranking) is a confidence score that the statement is valid. This paper aims to overcome these limitations by introducing a multi-faceted model of CSK statements and methods for joint reasoning over sets of inter-related statements. Our model captures four different dimensions of CSK statements: plausibility, typicality, remarkability and salience, with scoring and ranking along each dimension. For example, hyenas drinking water is typical but not salient, whereas hyenas eating carcasses is salient. For reasoning and ranking, we develop a method with soft constraints, to couple the inference over concepts that are related in a taxonomic hierarchy. The reasoning is cast into an integer linear programming (ILP), and we leverage the theory of reduction costs of a relaxed LP to compute informative rankings. Our evaluation shows that we can consolidate existing CSK collections into much cleaner and more expressive knowledge.

AKBC 2020 Conference Submissions

Exploiting Semantic Relations for Fine-grained Entity Typing

Revisiting Evaluation of Knowledge Base Completion Models

Enriching Knowledge Bases with Interesting Negative Statements

Joint Reasoning for Multi-Faceted Commonsense Knowledge