Knowledge graph is ubiquitous and plays an important role in many real-world applications, including recommender systems, question answering, fact-checking, and so on. However, most of the knowledge graphs are incomplete which can hamper their practical usage. Fortunately, knowledge graph completion (KGC) can mitigate this problem by inferring missing edges in the knowledge graph according to the existing information. In this paper, we propose a novel KGC method named ABM (Attention-Based Message passing) which focuses on predicting the relation between any two entities in a knowledge graph. The proposed ABM consists of three integral parts, including (1) context embedding, (2) structure embedding, and (3) path embedding. In the context embedding, the proposed ABM generalizes the existing message passing neural network to update the node embedding and the edge embedding to assimilate the knowledge of nodes' neighbors, which captures the relative role information of the edge that we want to predict. In the structure embedding, the proposed method overcomes the shortcomings of the existing GNN method (i.e., most methods ignore the structural similarity between nodes.) by assigning different attention weights to different nodes while doing the aggregation. Path embedding generates paths between any two entities and treats these paths as sequences. Then, the sequence can be used as the input of the Transformer to update the embedding of the knowledge graph to gather the global role of the missing edges. By utilizing these three mutually complementary strategies, the proposed ABM is able to capture both the local and global information which in turn leads to a superb performance. Experiment results show that ABM outperforms baseline methods on a wide range of datasets.
more »
« less
Complex Factoid Question Answering with a Free-Text Knowledge Graph
We introduce delft, a factoid question answering system which combines the nuance and depth of knowledge graph question answering approaches with the broader coverage of free-text. delft builds a free-text knowledge graph from Wikipedia, with entities as nodes and sentences in which entities co-occur as edges. For each question, delft finds the subgraph linking question entity nodes to candidates using text sentences as edges, creating a dense and high coverage semantic graph. A novel graph neural network reasons over the free-text graph—combining evidence on the nodes via information along edge sentences—to select a final answer. Experiments on three question answering datasets show delft can answer entity-rich questions better than machine reading based models, bert-based answer ranking and memory networks. delft’s advantage comes from both the high coverage of its free-text knowledge graph—more than double that of dbpedia relations—and the novel graph neural network which reasons on the rich but noisy free-text evidence.
more »
« less
- NSF-PAR ID:
- 10212078
- Date Published:
- Journal Name:
- WWW '20: Proceedings of The Web Conference 2020
- Page Range / eLocation ID:
- 1205 to 1216
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
This work investigates the challenge of learning and reasoning for Commonsense Question Answering given an external source of knowledge in the form of a knowledge graph (KG). We propose a novel graph neural network architecture, called Dynamic Relevance Graph Network (DRGN). DRGN operates on a given KG subgraph based on the question and answers entities and uses the relevance scores between the nodes to establish new edges dynamically for learning node representations in the graph network. This explicit usage of relevance as graph edges has the following advantages, a) the model can exploit the existing relationships, re-scale the node weights, and influence the way the neighborhood nodes’ representations are aggregated in the KG subgraph, b) It potentially recovers the missing edges in KG that are needed for reasoning. Moreover, as a byproduct, our model improves handling the negative questions due to considering the relevance between the question node and the graph entities. Our proposed approach shows competitive performance on two QA benchmarks, CommonsenseQA and OpenbookQA, compared to the state-of-the-art published results.more » « less
-
We are now over four decades into digitally managing the names of Earth's species. As the number of federating (i.e., software that brings together previously disparate projects under a common infrastructure, for example TaxonWorks) and aggregating (e.g., International Plant Name Index, Catalog of Life (CoL)) efforts increase, there remains an unmet need for both the migration forward of old data, and for the production of new, precise and comprehensive nomenclatural catalogs. Given this context, we provide an overview of how TaxonWorks seeks to contribute to this effort, and where it might evolve in the future. In TaxonWorks, when we talk about governed names and relationships, we mean it in the sense of existing international codes of nomenclature (e.g., the International Code of Zoological Nomenclature (ICZN)). More technically, nomenclature is defined as a set of objective assertions that describe the relationships between the names given to biological taxa and the rules that determine how those names are governed. It is critical to note that this is not the same thing as the relationship between a name and a biological entity, but rather nomenclature in TaxonWorks represents the details of the (governed) relationships between names. Rather than thinking of nomenclature as changing (a verb commonly used to express frustration with biological nomenclature), it is useful to think of nomenclature as a set of data points, which grows over time. For example, when synonymy happens, we do not erase the past, but rather record a new context for the name(s) in question. The biological concept changes, but the nomenclature (names) simply keeps adding up. Behind the scenes, nomenclature in TaxonWorks is represented by a set of nodes and edges, i.e., a mathematical graph, or network (e.g., Fig. 1). Most names (i.e., nodes in the network) are what TaxonWorks calls "protonyms," monomial epithets that are used to construct, for example, bionomial names (not to be confused with "protonym" sensu the ICZN). Protonyms are linked to other protonyms via relationships defined in NOMEN, an ontology that encodes governed rules of nomenclature. Within the system, all data, nodes and edges, can be cited, i.e., linked to a source and therefore anchored in time and tied to authorship, and annotated with a variety of annotation types (e.g., notes, confidence levels, tags). The actual building of the graphs is greatly simplified by multiple user-interfaces that allow scientists to review (e.g. Fig. 2), create, filter, and add to (again, not "change") the nomenclatural history. As in any complex knowledge-representation model, there are outlying scenarios, or edge cases that emerge, making certain human tasks more complex than others. TaxonWorks is no exception, it has limitations in terms of what and how some things can be represented. While many complex representations are hidden by simplified user-interfaces, some, for example, the handling of the ICZN's Family-group name, batch-loading of invalid relationships, and comparative syncing against external resources need more work to simplify the processes presently required to meet catalogers' needs. The depth at which TaxonWorks can capture nomenclature is only really valuable if it can be used by others. This is facilitated by the application programming interface (API) serving its data (https://api.taxonworks.org), serving text files, and by exports to standards like the emerging Catalog of Life Data Package. With reference to real-world problems, we illustrate different ways in which the API can be used, for example, as integrated into spreadsheets, through the use of command line scripts, and serve in the generation of public-facing websites. Behind all this effort are an increasing number of people recording help videos, developing documentation, and troubleshooting software and technical issues. Major contributions have come from developers at many skill levels, from high school to senior software engineers, illustrating that TaxonWorks leads in enabling both technical and domain-based contributions. The health and growth of this community is a key factor in TaxonWork's potential long-term impact in the effort to unify the names of Earth's species.more » « less
-
null (Ed.)This work deals with the challenge of learning and reasoning over multi-hop question answering (QA). We propose a graph reasoning network based on the semantic structure of the sentences to learn cross paragraph reasoning paths and find the supporting facts and the answer jointly. The proposed graph is a heterogeneous document-level graph that contains nodes of type sentence (question, title, and other sentences), and semantic role labeling sub-graphs per sentence that contain arguments as nodes and predicates as edges. Incorporating the argument types, the argument phrases, and the semantics of the edges originated from SRL predicates into the graph encoder helps in finding and also the explainability of the reasoning paths. Our proposed approach shows competitive performance on the HotpotQA distractor setting benchmark compared to the recent state-of-the-art models.more » « less
-
Background: Relationships between bio-entities (genes, proteins, diseases, etc.) constitute a significant part of our knowledge. Most of this information is documented as unstructured text in different forms, such as books, articles and on-line pages. Automatic extraction of such information and storing it in structured form could help researchers more easily access such information and also make it possible to incorporate it in advanced integrative analysis. In this study, we developed a novel approach to extract bio-entity relationships information using Nature Language Processing (NLP) and a graph-theoretic algorithm. Methods: Our method, called GRGT (Grammatical Relationship Graph for Triplets), not only extracts the pairs of terms that have certain relationships, but also extracts the type of relationship (the word describing the relationships). In addition, the directionality of the relationship can also be extracted. Our method is based on the assumption that a triplet exists for a pair of interactions. A triplet is defined as two terms (entities) and an interaction word describing the relationship of the two terms in a sentence. We first use a sentence parsing tool to obtain the sentence structure represented as a dependency graph where words are nodes and edges are typed dependencies. The shortest paths among the pairs of words in the triplet are then extracted, which form the basis for our information extraction method. Flexible pattern matching scheme was then used to match a triplet graph with unknown relationship to those triplet graphs with labels (True or False) in the database. Results: We applied the method on three benchmark datasets to extract the protein-protein-interactions (PPIs), and obtained better precision than the top performing methods in literature. Conclusions: We have developed a method to extract the protein-protein interactions from biomedical literature. PPIs extracted by our method have higher precision among other methods, suggesting that our method can be used to effectively extract PPIs and deposit them into databases. Beyond extracting PPIs, our method could be easily extended to extracting relationship information between other bio-entities.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1145/3366423.3380197
