- Date:
- 2008-2011
- Themes:
- Machine Learning, Security & Trust, Agent Based Computing
- Funding:
- The Virtual Centre of Excellence in Mobile and Personal Communications, Technology Strategy Board
The 'Instant Knowledge' research project is jointly funded by MobileVCE, the Technology Strategy Board (TSB) and the EPSRC, carrying out research into the field of machine learning and intelligent agents. The project itself centres around the notion of a personal distributed environment involving mobile devices (such as phones, PDA's, laptops etc.) and using context-sensitive information to extract and share knowledge in a useful way; whilst simultaneously addressing any security and privacy issues.
This is a multi-disciplinary project lead by the MobileVCE organisation, whose membership consists of several industrial members, including many mobile phone operators. The University of Strathclyde and Royal Holloway, University of London are the two other academic partners involved in the project.
The role of Southampton in the project is to provide machine learning and agent expertise in order to collate context features in an intelligent way so that information between agents can be shared either via ad-hoc meetings (e.g. over bluetooth or short-range wireless) or via a central repository (for example a central store on a corporate network.
Primary investigator
- cjs
Partners
- Mobile VCE
- Royal Holloway, University of London
- University of Strathclyde
Associated research group
- Information: Signals, Images, Systems Research Group
- Date:
- 2008-2009
- Themes:
- Accessible Technologies, Learning Technologies
- Funding:
- JISC
Multimedia has become technically easier to create (e.g. recording lectures) but while users can bookmark , search, link to, or tag a COMPLETE podcast or video available on the web, they cannot bookmark, search, link to, or tag A PARTICULAR SECTION WITHIN that podcast or video. Users cannot, for example, easily find or return to a particular point in a recording or associate their notes or resources with a specific section of a recording. As an analogy, users would clearly find a text book difficult to use if it had no contents page, index or page numbers. Therefore the growing amount of knowledge available in multimedia format has yet to achieve the level of interconnection and manipulation achieved for text documents via the World Wide Web and so realize the exciting opportunities for learning that can occur in ââ¬ËWeb 2.0ââ¬â¢ and ââ¬Ësocial softwareââ¬â¢ environments.
The MACFOB project aims to develop a web-based multimedia annotation tool that will meet the important and pervasive user need of making multimedia web resources (e.g. podcasts) easier to access, search, manage, and exploit for students, teachers and other users through developing and deploying technologies that support the creation of synchronised notes, bookmarks, tags, images and text captions.
User needs requirements for MACFOB have been established through user studies which have shown how teachers and students benefit from making annotations to help search and manipulate recordings of lectures or classes to find and make use of information. Consultations with user groups have confirmed the user needs requirements for both teaching and administration and interest in involvement in the project. We will work with a number of interested users throughout the project, in order to build a generalised Multimedia Annotation tool that is simple enough for teachers and learners to appropriate for their own use.
Primary investigator
Secondary investigators
- sk07v
- Gary Wills
- Lester Gilbert
- David Millard
- jk2
Associated research groups
- Learning Societies Lab
- Electronic and Software Systems
- Web and Internet Science
- Date:
- 2006-2011
- Themes:
- e-Defence, Semantic Web, Knowledge Technologies
- Funding:
- Ministry of Defence, U.S. Army Research Laboratory
The International Technology Alliance (ITA) is a consortium of academic of industrial partners that seeks to investigate issues relating to future coalition capabilities, particularly those affecting US and UK forces. The ITA comprises 4 technical areas: network theory (TA1), security within a system of systems (TA2), sensor information processing and delivery (TA3), and distributed coalition planning and decision-making (TA4). Within these technical areas, the research undertaken by the School of Electronics and Computer Science falls within a number of focus areas. These include: semantically-mediated data fusion, semantic integration and interoperability, user interfaces for the Semantic Web, and collaborative planning and plan enrichment.
Semantically-Mediated Data Fusion
The future battlespace will feature a range of sophisticated sensor systems that promise to increase the level of battlespace resolution afforded to commanders at all levels of the command chain. In order to make best use of available sensor resources, coalition forces will benefit from data fusion processes that are able to exploit semantically-enriched representations of domain-relevant information. Semantic representations could benefit data fusion in two ways. Firstly, they may facilitate the exploitation of external, contextual information in ways that biases the analysis of low-level feature vectors. Secondly, semantic representations may facilitate the exploitation of fusion-related outcomes by providing a set of stable, symbolic atoms that align themselves with the conceptual infrastructure of the domain of discourse. The ITA Semantically-Mediated Data Fusion initiative aims to evaluate the contribution of semantic technologies to data fusion using a combination of real-world datasets, pattern recognition techniques and domain ontologies.
Semantic Integration and Interoperability
Semantic integration and interoperability are important capabilities for coalition forces. Coalition operations will typically involve the exploitation and exchange of semantically heterogeneous information, especially when the operational context demands close cooperation with non-military agencies (e.g. diplomatic, humanitarian and civil authorities). The concern is that even in situations where the physical transfer of information is supported by network infrastructures; the meaning of information content may be lost or distorted as it traverses organizational, national and cultural boundaries. The focus of our research in respect of the ITA Semantic Integration and Interoperability initiative is to investigate approaches to information exchange that maximally exploit the potential of the Semantic Web to provide a framework for the consensual interpretations of entities, events and actions across force elements and between coalition partners. The ultimate aim is to provide a foundation for coalition inter-operability, enabling semantic integration with respect to both digital datalink networks and unstructured, non-military information sources.
Collaborative Planning and Plan Representation
The aim of the Collaborative Planning and Plan Enrichment initiative is to investigate a number of representational issues that emerge in coalition planning contexts. In particular, the project aims to understand the representational requirements of plans, especially with respect to the representation of critical information (e.g. command intent) that may influence the ability (or propensity) of agents to understand, accept and/or implement a plan. Key research issues include (but are not necessarily limited to) an analysis of the factors that may undermine shared understanding in coalition planning contexts, an analysis of the factors that influence the acceptability or usability of a coalition plan, and an analysis of optimal modes of communicating plan-relevant information in a heterogeneous agent environment.
Primary investigator
Secondary investigators
- ps02v
- tdh
- Peter Houghton
- db06v
- dm07v
- Jim Hendler
- Cheryl Giammanco
- Ali Bahrami
- Michael Dorneich
- Jie Bao
- Katia Sycara
Partners
- IBM United Kingdom Limited
- Defence Science and Technology Laboratory [dstl]
- Rensselaer Polytechnic Institute
- U.S. Army Research Laboratory
- Honeywell
- Boeing
- Carnegie Mellon University
- Applied Research Associates
Associated research group
- Intelligence, Agents, Multimedia Group
- Date:
- 2006-2009
- Themes:
- e-Defence, Web Science, Knowledge Technologies, Semantic Web
- Funding:
- Ministry of Defence
OntoMediate is a 2.5 year research project, funded by the UK Ministry of Defence as part of the Data and Information Fusion Defence Technology Centre (DIF DTC) initiative. The project focuses on the creation of information brokerage services that incorporate ontology mediation techniques as the basis for meaning-preserving modes of information exchange between diverse user communities. Part of the research effort associated with the project aims to explore the extent to which collaborative environments can be used to support the identification of inter-relationships between ontologies. Such an approach seeks to surmount traditional barriers to information exchange by providing a collaborative framework for the identification of ontology alignments and the active translation of information content. The ultimate goal of the project is to facilitate a common understanding of information content as a precursor to better modes of agent collaboration, communication and shared situation awareness.
Primary investigator
Secondary investigators
- ha
- gc3
Associated research group
- Intelligence, Agents, Multimedia Group
- Date:
- 2004-2006
- Themes:
- Knowledge Technologies, e-Defence, Semantic Web
- Funding:
- Ministry of Defence
Situation awareness is a critical success factor in military operations. Even when the operational context is not directly adversarial, as is the case in most humanitarian and peace-keeping missions, awareness of the temporal unfolding of events, the strategic displacement of military assets and the ability to anticipate the actions of other (sometimes competing) agencies, all serve to underpin the successful realization of operational objectives. The need for improved situation awareness is particularly important when one considers the increasingly sophisticated technological backdrop against which military operations are typically undertaken. The advent of network-enabled capabilities (NEC) and the growth of the internet as a medium for information dissemination, affords great opportunities for situation awareness, but it also presents some relatively new and distinct challenges. One challenge relates to the need to distinguish relevant information from background noise (the concern here is that highly relevant information may be swamped by less relevant information). Another concern relates to the rate of information dissemination in todayââ¬â¢s media-intensive environment. The concern here is that the dynamics of the situation picture may warrant rapid switching between different problem-solving goals. When goal switching is mandated by changing operational commitments then different subsets of information will need to be dynamically integrated or aggregated to support changing situation awareness concerns.
AKTiveSA is a 3 year project that forms part of the Data and Information Fusion Defence Technology Centre (DIF DTC) initiative. Its aim is to explore knowledge-based approaches to information fusion and enhanced situation awareness in military operational contexts other than war (MOOTW), specifically humanitarian assistance and disaster relief. As part of the effort to improve situation awareness, AKTiveSA exploits domain ontologies that enhance information exchange and inter-operability between diverse agencies and user communities (e.g. NGOs, UN agencies, coalition forces, government departments, etc). It also introduces the idea of semantic filters that provide filtered views of the situation picture to support task-relevant information processing. Technological components, called knowledge monitors, support situation awareness by actively monitoring the situation picture for events that impact on current plans or problem-solving goals. These components are defined in terms of semantic queries that execute as background processes, constantly monitoring the totality of the information space and alerting users to relevant changes via a variety of output formats, e.g. emails, RSS, SMS, etc. AKTiveSA also devotes considerable attention to user interface design issues and this is reflected in the development of the AKTiveSA Technical Demonstrator System (TDS), a sophisticated user interface that combines 3-D visualization components with novel ontology browsing and navigation techniques. The scientific and technological outcomes of the AKTiveSA project support other e-Defence research initiatives within the School of Electronics and Computer Science at the University of Southampton. These include the SEMIOTIKS initiative, which focuses on information extraction, knowledge processing and information visualization; and the MIMEX initiative, which concentrates on situation awareness enhancement and information exploitation in hostile information environments.
Primary investigators
Secondary investigators
- ar5
- ps02v
Associated research group
- Intelligence, Agents, Multimedia Group
- Date:
- 2006-2009
- Themes:
- e-Defence, Web Science, Semantic Web, Knowledge Technologies
- Funding:
- Ministry of Defence
SEMIOTIKS is a 3 year research programme sponsored by General Dynamics and QinetiQ as part of the MoD-funded Data and Information Fusion Defence Technology Centre (DIF DTC) initiative. SEMIOTIKS aims to address the challenges faced by military and civilian agencies in leveraging the potential of large scale information repositories to support enhanced situation awareness and information superiority. The project combines state-of-the-art approaches to text analysis and resource classification with semantic technologies that support information retrieval, knowledge extraction, knowledge discovery, text summarization, knowledge dissemination and decision-making. A key aim of the project is to facilitate the identification, classification and processing of unstructured textual resources by capitalizing on the availability of natural language processing (NLP) and machine learning (ML) techniques. NLP technologies are used to extract relational information from information resources in a form that is suitable for knowledge processing, while ML techniques support the classification of resources with respect to the elements of a domain ontology. Both capabilities are supported by a ML-based approach to semantic annotation in which a system progressively learns to annotate text fragments based on user-supplied exemplars.
In addition to the attempt to improve the acquisition of domain-relevant information, SEMIOTIKS also aims to support the user with respect to the manipulation, processing and dissemination of knowledge. One of the research objectives of the project is to investigate techniques for efficient and scalable rule-based processing on the Semantic Web, and the results of this work are informing the development of a generic decision support system that provides both a situation assessment and planning capability. Other research focus areas include the development of knowledge monitors that extend the situation monitoring capabilities of the AKTiveSA Technical Demonstrator System (TDS). These components alert users about any changes or events in the situation picture that might impact on a userââ¬â¢s problem-solving goals. SEMIOTIKS also aims to extend the user interaction capabilities of the AKTiveSA TDS. In this case, a rich suite of user interface components are being developed which support the human end user with respect to the manipulation of task-relevant information. Specific focus areas for visualization and interaction research include the development of visual query languages, graphical query designers, multi-touch displays, and novel approaches to ontology navigation and visualization.
Primary investigator
Secondary investigators
- ps02v
- ar5
- sfl
Partner
- QinetiQ
Associated research group
- Intelligence, Agents, Multimedia Group
- Date:
- 2001-2007
- Themes:
- Semantic Web, Content Based Retrieval, Knowledge Technologies
- Funding:
- EPSRC, Department
The overall aim of the ArtEquAKT project is to use knowledge acquisition and analysis techniques to extract information from web pages on a given subject domain and construct an knowledge base overlaid with an ontology. The ontology can then be used to construct stories on the fly, by using story templates to walk the ontology and extract appropriate knowledge from the knowledge base underneath. Eventually it is hoped that the ontologies might be built automatically, although initially they will be manually constructed.
Primary investigators
Secondary investigators
- pass
- Nigel Richard Shadbolt
- David De Roure
- Sanghee Kim
Associated research group
- Intelligence, Agents, Multimedia Group
- Date:
- 2007-
- Themes:
- ELearning, Platforms and Tools, Web Science
The Learning Societies Toolkit is an initiative within the Learning Societies Lab to create a new generation of e-learning tools for the Web Literate generation.
The new generation of students entering our institutions are more than computer literate, they are used to taking ownership of the information systems that they use, to create an individual mix of applications and systems that both forms their digital identity and shapes their virtual experience. To these web literate students the old-style monolithic Virtual Learning Environments (VLEs) are to restrictive and rather than help the student/teacher relationship they can get in the way, creating a false distance that hinders learning and undermines the educational experience.
Learners and teachers are already rebelling, by moving away from institutional provision and using online tools such as public discussion forums, social sites, wikis and blogs. Personal Learning Environments (PLEs) are one potential solution, existing PLEs accept an open set of data sources (such as public repositories and commercial news sources), but they still try to gather the students experience into one institutionally owned place.
The Learning Societies Toolkit is an attempt to maintain a sense of institutional provision, but to allow the student to take ownership of the overall experience. The toolkit is a collection of e-learning tools that:
- Are loosely coupled
- Are functionally focused
- Have open applicability
- Can be appropriated as needed
- Are available as a public service
- Are Open Source (so can be a private service)
Currently there are number of projects what will contribute to the toolkit, including PeerPigeon, ASDEL and Faroes.
Primary investigator
Associated research group
- Learning Societies Lab
- Date:
- 2006-2007
- Themes:
- Semantic Web, Pervasive Computing and Networks, Educational Enhancement, Human Computer Interaction, Web Science
- Funding:
- NCeSS
This project will demonstrate how Semantic Grid technologies can be used to provide enhanced techniques for data collection and use within a grid-enabled environment, thus enhancing the capacity to address substantive social science research. It will achieve this by conducting a case study in which semantic annotation (i.e. machine-processable annotation using Semantic Web technologies) is used both in capturing and working with the digital record, in support of subsequent qualitative and quantitative analysis. As a demonstrator this will extend and inform existing NCeSS activities in video annotation and new forms of digital record being developed at the MixedMediaGrid and Digtal Record nodes, as well as producing a training output in the application of grid technologies for the wider social science community. This case study will set the agenda for future work in this area.
The case study focuses on the research and practice of skills-based learning in the context of health care education. The project utilises a significant facility in the School of Nursing and Midwifery in Southampton that provides a powerful testbed for the study, with very extensive data collection and replay capabilities. This enables new forms of data capture and analysis involving multiple simultaneous video and audio sources and video focus groups. Of particular relevance to the Digital Record and MixedMediaGrid nodes is the rich form of record including audiovisual data but also data streams from pervasive devices (e.g. the instrumented simulated patients), control event streams, scenario storyboards and multiple layers of annotation. The case study provides a social science research context in which to investigate the new techniques.
Primary investigators
- David De Roure
- Mary Gobbi
- J. W. McDonald
Secondary investigators
Associated research groups
- Intelligence, Agents, Multimedia Group
- Learning Societies Lab
- Date:
- 2007-2008
- Themes:
- ELearning, Platforms and Tools, Digital Libraries, Semantic Web
- Funding:
- JISC
The main theme of EdScene is to develop scenarios using semantic web to support all roles (teachers, students, quality assurance people, management, perspective employers, etc) in an academic system.
The University of Southampton has taken the strategic decision to develop a repository for educational materials using its well established EPrints research repository software as the framework. EdScene aims to describe EdSpace scenarios in ECS, University of Southampton, in particular functional activities of various roles such as students, teachers, employers and external professional bodies. This is to illustrate the expected added-value of a community repository holding educational materials together with metadata and semantics.
Semantic Web technologies are being exploited to extend and improve the existing metadata tagging mechanism. An educational ontology explicitly defined contextual conceptualization of the educational domain, which can be then used to annotate educational artefacts such as lecture resources, programme specifications, modules and assessments. This, together with semantics tools, gives the users a choice to make their resources more machine-processable (aggregation, reasoning, etc) by adding an enriched layer of the semantic web linking these educational artefacts with metadata and formal semantics to support key functional requirements.
Overall the main objective of this project is to support all stakeholders of an academic system for solving their intelligent queries. The following are some of the examples that the project aims to perform:
Students choosing options wish to access ratings and comments from previous students
Although VLE allows students to access the e-portfolio of the programme and modules in order to select, rate and comments throughout their participation of the educational learning activities, the results often do not have a permanent URL for access outside the VLE. The EdSpace vision of using a repository in stead is to overcome this shortcomings, aiming to allow permanent data cruration and convenient access of resources from other system.
Teachers needs to understand what they need to teach
Teachers, in particular those new to the school, can access to existing programme and modules annotations while designing their modules to identify popular subjects, fulfill required learning outcomes (at module and programme level) and avoid repetition of learning materials.
Employer wishes to understand students transcript
Employers at recruitment often find it difficult to match-make their job skill demands with graduates transcripts. The repository (and the semantic mark-up) should be able to help them scrutinize job candidates transcripts in terms of the programme, learning outcome and module specification.
Professional bodies such as QAA and BCS wish to understand CS in ECS
With the aim to safeguard and help to improve the academic standards and quality, Quality Assurance Agency for Higher Education needs to work closely with the education institute to define academic standard and quality. The repository framework, in particular the educational ontology as the conceptualization backbone, will reflect QAA's standard. They will wish to assure themselves that learning outcomes are actually being met. This will require them to access assessment materials such as assignments (the instructions supplied to students, the student work submitted, and the marks/feedback obtained) and exams (both the papers supplied to students and the marked student scripts). In addition, they also need to understand the learning infrastructure (timetables, library provision, lab provision -- hardware and software, CVs of academic and technical staff, staff and student numbers, minutes of academic committees, etc.)
Primary investigators
- hcd
- Andrew Gravell
- sk07v
- ft
Associated research group
- Learning Societies Lab