Projects

Motivated by the need for reasoning approaches that can scale well even on the most expressive, and large ontologies, neuro-symbolic approaches have received major attention in recent times. The idea is to combine the robustness and learning capabilities of the artificial neural networks along with the precise reasoning abilities of logic-based approaches. We are exploring a reinforcement learning solution that can take advantage of the exploration-exploitation trade-off to reduce the overall time complexity of the reasoning algorithms.

Although data is available in different forms and formats, they are related to each other. For example, air pollution and weather are connected. So are patient records and their test results. Unless a common structure is provided, along with integrating and connecting (linking) them, it would be hard to query and analyze the data. We make use of Semantic Web technologies such as OWL 2 and RDF to create such a platform.

Preterm babies in NICUs are monitored carefully to make sure that they are getting the right balance of fluids and nutrition. We are developing an ontology based nutrition guideline system that captures the information of neonates such as day of life, signs and symptoms. This system will recommend the feed amount and the nutrition for the preterm babies.

There are several existing ontology reasoners that span awide spectrum in terms of their performance and the expressivity that they support. In order to benchmark these reasoners to find and improve the performance bottlenecks, we ideally need several real-world ontologies that span the wide spectrum in terms of their size and expressivity. This is often not the case. One of the potential reasons for the ontology developers to not build ontologies that vary in terms of size and expressivity, is the performance bottleneck of the reasoners. To solve this chicken and egg problem, we need high quality ontology benchmarks that have good coverage of the OWL 2 language constructs, and can test the scalability of the reasoners by generating arbitrarily large ontologies. OWL2Bench is a work in that direction. It is an extension of the well-known University Ontology Benchmark (UOBM). OWL2Bench consists of the following - TBox axioms for each of the four OWL 2 profiles (EL, QL, RL, and DL), a synthetic ABox axiom generator that can generate axioms of arbitrary size, and a set of SPARQL queries that involve reasoning over the OWL 2 language constructs.

Most of the knowledge graph embedding models focus on capturing the structural properties of the graph and the interaction between the entities. They do not take into account the constraints and characteristics of the underlying ontology. Consequently, the embeddings produced by such methods are not suitable for reasoning tasks such as classification, satisfiability and consistency checking. So we are working on mapping the classes and the relations in an ontology to an n-dimensional vector space such that the relations between them are preserved.

Ontologies evolve over time due to changes in the domain and the requirements of the application. Maintaining an ontology over time and keeping it up-to-date with respect to the changes in the domain and the requirements of application is hard. But a high quality ontology can significantly reduce the effort and the cost of ontology maintenance. Ontology Design Patterns (ODPs) can be used to improve the quality of an ontology and make it more modular and reusable. But with increasing number of ODPs spread across different categories, it is not easy to determine the right set of ODPs to choose for a particular use case even for experts. This becomes even more difficult in the case of refactoring existing ontologies using the right set of ODPs. We built a tool named ODPReco that can recommend the possible ODPs to use in a given ontology by analyzing its lexical, structural, and behavioural aspects.

RDF graphs and Property Graphs (PG) are the two popular Knowledge Graph (KG) representation formats. SPARQL is the standard and the W3C recommended query language for RDF graphs. There are several query languages for PGs, but one of the most used query languages is Cypher. Although it is easier to model the data in the form of PGs when compared to RDF graphs, they are not standardized. So having a bridge between these two formats and query languages is very useful. In this project, we are working on building a SPARQL to Cypher query converter.

Building an ontology is not only a time-consuming process, but it is also confusing, especially for beginners and the inexperienced. Although ontology developers can take the help of domain experts in building an ontology, they are not readily available in several cases for a variety of reasons. Ontology developers have to grapple with several questions related to the choice of classes, properties, and the axioms that should be included. Apart from this, there are aspects such as modularity and reusability that should be taken care of. From among the thousands of publicly available ontologies and vocabularies in repositories such as Linked Open Vocabularies (LOV) and BioPortal, it is hard to know the terms (classes and properties) that can be reused in the development of an ontology. A similar problem exists in implementing the right set of ontology design patterns (ODPs) from among the several available. Generally, ontology developers make use of their experience in handling these issues, and the inexperienced ones have a hard time. In order to bridge this gap, we propose a tool named OntoSeer, that monitors the ontology development process and provides suggestions in real-time to improve the quality of the ontology under development. It can provide suggestions on the naming conventions to follow, vocabulary to reuse, ODPs to implement, and axioms to be added to the ontology.

Ontology learning is the process of building ontologies automatically from unstructured data sources such as text. Several ontology learning systems have been developed, but they have been trained and tested on different datasets. In order to evaluate and compare these systems as well as make progress in ontology learning, it is critical to have good benchmarks. In this project, we work on generating text with annotations that can be used to build ontologies with different types of axioms.