Researches going on

Semantic Core

We have a strong research background on ontology alignment of semantic web to facilitate semantic knowledge integration as we have created the fastest and scalable algorithm, Anchor-Flood, in this domain. However, we are still addressing complex mapping , uncertainity management and dynamic block management in ontology alignment. Furthermore, our researches go on some optimization introducing efficient persistent memory model. Billion triple semantic knowledge integration heavily depends on optimization and scalability of algorithms and on generation of automatic weight to ontology relations.

Linked Data (Semantic Knowledge-base)

Machine understandable informtion is the motto of linked data. We have already started a bunch of researches on linked data. For example, we are proud of our heroitic liberation war and its freedom fighters. An honorable freedom fighter was being trained in a particular place and by a group of people and later on, they were fighting at places under a commender at a particular time and there were many historical events. The total scenerio is captured by introducing linked data which is understandable by machine.
Furthermore, we research on information retrieval (IR) in semantically tagged image, linked data for stock exchange and semantic music portal and personalization

Bioinformatics

When someone is affected by a disease, it affects the disease causative genes (genotypes) and the genes is responsible for physiological changes (called phenotypes). In turn, a patient goes to see a doctor, let the doctor know the phenotypes and the doctor diagnose the disease. In this regard, we are developing a disease predicting system which analyzes phenotype-genotype and genotype-disease bipartite graphs probabilistically. To make the system more realistic we adapt term-binding to the system and medical image annotating and classifying tools.

Geographic Information Processing

Geographic information has a wide effect on many information processing. In our country, there are a number of division, each of which contains districts. Districs again contain Upazillas and Upazilla contains a number of unions and a village is associated or linked to a union. Furthermore, a place may have political name and traditional names as well. We research to construct spatial information hierarch of Bangladesh, geo-spatial linked data on Bangladesh and geo-spatial data visualization.

Natural Language Processing

We still have a lot of primitive researches yet to start with Bangla language. Along with these primitive work, we start researching on news classification such as academician would only be interested academic information, some other might like sports, or politics or other topics. Our system would provide them the required information only from different online news papers. Furthermore, we are researching on construction of Bangla WordNet and construction of Bangla ConceptNet.