Computer Science Department, School of Informatics,
Community Grids Lab (CGL), Pervasive Technology Labs (PTL),
Indiana University, Bloomington (IUB).
Dr. Aktas's Research Interest span into two areas: Systems Science and Data Science. In former area, his emphasis has been on Distributed Systems and Grid Computing. Within this emphasis, he mainly focuses on Distributed Resource Management, Grid/Web Information Systems, Information Integration and Federation in Grid/Web Information Services, Geographic Information Systems, Service Oriented Architecture based Distributed Systems for Science and Business Applications, Semantic Grid/Web based Systems, and Handling Distributed Data/Metadata. In latter area, his emphasis has been on Data/Web Mining and Information Retrieval. Within this emphasis, he mainly focuses on the Machine Learning Techniques to Mine Massive Datasets, Web Search, Personalized Web Search, Conversational Recommendation Systems for Information Retrieval. His research interests also span into Semantic Grid/Web, Web Search and Information Retrieval, and Data and Web Mining.

Geographic Information Systems (GIS) integrate hardware, software, and data for capturing, managing, analyzing, and displaying all forms of geographically referenced information. The Open Geospatial Consortium (OGC) provides several important, open specifications for building GIS services that will allow us to leverage other groups’ efforts. Relevant OGC data model specifications include the Geographic Markup Language (GML) and the SensorML family of data and metadata information models. OGC service specifications include the Web Feature Service (WFS), which provides access to archival data and abstract map features, and the Web Map Service (WMS), which generates human-comprehensible overlay maps. A WMS may generate all images locally from abstract information stored in remote WFS instances, or it may combine its own image capabilities with maps from other WMS instances. Dr. Aktas and his colleagues completed comprehensive projects to build GIS services using current Grid and Web Service standards. By following Web Service standards, they incorporated general Web Service functionality, including security, reliability, and information management [link].

Service Oriented Architecture (SOA) principles have gained great importance in e-business and e-science applications in recent years. Following this trend, Dr. Aktas's past and ongoing research has mainly evolved around designing and developing SOA-based systems to address resource management requirements of large scale data-intensive application domains such as Geographical Information System Grids [pdf], Sensor and Collaboration Grids [pdf]. As the scale and complexity of such SOA-based systems increased, an emerging need has appeared for advanced methodologies to locate desired services. As the rich-interacting SOA-based systems emerged and services within these systems started collaborating with each other within a work-flow session to produce a common functionality, another need has appeared for storing, querying, and sharing metadata generated as result of service interactions. To address these needs, independent projects have developed their own solutions to Information Services over the years. However, these solutions are not interoperable with each other, target vastly different systems and address diverse sets of requirements.

Web Information Systems has been the main focus of Dr. Aktas's research to address aforementioned challenges. He designed, developed and evaluated a Hybrid Web Information System for SOA-based systems to manage all kinds of information that may be associated with services. The Hybrid Web Information System provides unication, federation and interoperability of major grid/web information system implementations. It runs one layer above the existing implementations of information systems and integrates one to many local information systems under one hybrid architecture. To achieve high performance, the Hybrid System employs different techniques such as in-memory based storage solutions [pdf].

XML Metadata Service is a standardized approach for building grid/web information systems. To facilitate testing of the Hybrid System, Dr. Aktas designed and implemented two major grid/web information systems: a) Extended UDDI XML Metadata Service, b) WS-Context XML Metadata Service. The Extended UDDI XML Metadata Service improves out-of-box UDDI Specication by providing various improvements: metadata-oriented query capabilities, leasing capabilities, geo-spacial domain compatible registry capabilities, domain-independent query capabilities. The WS-Context XML Metadata Service implements an extended version of WS-Context Specication. This implementation supports unique requirements of dynamic metadata management of rich-interacting systems such as notification, high-performance, up-to-date registry. Integrated with two major grid/web information systems, the Hybrid System supports both the scalability of large amounts of relatively slowly varying data and high performance rapidly updated data in dynamically assembled service collections [pdf].

Handling Distributed Data/Metadata has been a core research issue in distributed computing. To address the challenges in this area and to decentralize the Hybrid Web Information System, described in previous paragraph, Dr. Aktas designed and developed a distributed replica hosting environment that provides its clients the best available performance while providing a robust system which consumes minimum computing and network resources. This system provides solutions to core elements of designing a replica hosting system: replica content placement, optimization techniques, request distribution and consistency enforcement [pdf].

Semantic Web attempts to define a metadata information model for the Web to aid in information retrieval and aggregation, and provides advanced capabilities intended to enable knowledge representation and limited machine reasoning. In this area, Dr. Aktas designed and developed a recommendation system that benefits from expressiveness and reasoning capabilities of Semantic Web languages. To achieve this he designed an ontology for earthquake modeling applications. This system suggests resources of interest when the resources may be too difficult to locate with traditional retrieval systems. This recommender approach uses Conversational Case-Based Reasoning (CCBR), in which recommendations are made on an as-needed basis by doing reasoning from the current set of cases. To achieve a standard representation, He adopted Semantic Web languages such as Resource Description Framework (RDF) as the representation syntax of metadata, enabling RDF representation of CCBR cases to provide a standard means of representation [pdf].

Personalized Web Search has gained great popularity to improve search effectiveness in recent years. The objective of personalized search is to provide users with information tailored to their individual contexts. The PageRank algorithm attempts to provide an objective global estimate of Web page importance. However, the importance of Web pages is subjective for different users and thus can be better determined if the PageRank algorithm takes into consideration user preferences. In order to provide personalized web search and address limitations of global PageRank, Dr. Aktas and his colleagues introduced a novel methodology to personalize PageRank scores based on hyperlink features readily available from Web page URLs [pdf].

Links to Projects:

  • FTHPIS: fault tolerant, high performance metadata management in Grids
  • ExtendedUDDI: managing quasi-static metadata associated to grid/web services
  • WSContext: managing dynamic metadata of rich interacting systems
  • CrisisGrid: building GIS based service oriented architectures
  • GlobalMMCS: service-oriented architecture based audio-video conferencing system
  • Vlab: Virtual Laboratory for Earth and Planetary Materials - Vlab project
  • SERVODiscovery: metadata discovery for distributed earthquake simulation codes
  • Personalized Search: improving PageRanking algorithm to personalize web search
  • OKC: online knowledge center project
Home Page
Click Me

Mehmet S. Aktas