Graduate Course syllabus

NR385

Instructors:

Ferdinando Villa, Roelof Boumans, Marta Ceroni, Serguei Krivov.

Knowledge about the connected socio-economic and ecological system covers: (1) the state of the system (data); (2) the workings of the system (models); (3) the way we relate to the system (policies, values, priorities, and development scenarios). Traditionally, only data are stored in databases and have intuitive search interfaces, while models, policies, values etc. are much less accessible; as an example, specific software education and investments are normally required to explore dynamic models. The broad mission at UVM’s Ecoinformatics Collaboratory (http://ecoinformatics.uvm.edu) is to (1) eliminate the distinctions between the ways these different classes of knowledge are accessed, contributed, and used and make access to all kinds of knowledge, not only data, as natural, intuitive, and integrated as possible. One of the EC’s goal is to change the role of the Web from a loosely indexed, unstructured "encyclopedia" to that of an open collaboratory where knowledge can be not only located without ambiguity, but also explored, produced, and communicated.

The fundamental tool to enable transparent integration of knowledge is the knowledge map or ontology, a formalization of concepts and relationships that is used to describe the conceptual structure of data and models. Knowledge maps describe the nature of the knowledge represented in data and models, and guide the integration of knowledge represented in different ways and coming from widely changing disciplinary sources. Knowledge maps can be also used with interactive graphical browsers to explore concepts and search for related information. When used as learning tools, knowledge maps allow users to navigate clearly organized concepts and follow logical relationships instead of “poking in the dark” to find data. Such maps, enriched with explanations of each concept and relationship, are tremendous learning tools that suit both the "hunter" and the "gatherer" style of database user.

The development of ontologies requires organizing, synthesizing and formalizing knowledge from literature and expert panels. The development of collaborative tools and techniques for ontology development is likely to become the main IT challenge of the decade. The EC has incorporated knowledge maps in breakthrough products like the Ecosystem Services Database, an NSF-funded integrated, web-accessible knowledge base to inform the economic valuation of ecosystem services, whose content has been largely provided through the work of UVM graduate students. An initial ESD prototype is now available online (http://esd.uvm.edu). A fast-increasing number of IT projects in ecology and biodiversity rely on the development of ontologies to enable seamless integration between knowledge from different sources. The EC is a partner in the largest IT projects in ecology and biodiversity to date.

The goal of this course is to provide students with technologies to manage, discover and re-think ecological, biodiversity and environmental information. With so much information on the Web, and of heterogeneous nature, managing it with conventional tools is becoming almost impossible. New tools and techniques are necessary to enable efficient access to data, sharing data, extracting information from data, and making use of the information. This course will deal with the management of information along the entire data stream, from data representation through data acquisition and knowledge discovery to maintenance of persistent repositories.

As secondary users, most of us are end-consumers of information without knowing the processes of representing and storing knowledge, and consequently we accept definitions and assumptions that others have labeled this information with. For example, the term biodiversity includes ecological diversity from individuals to communities but most of the existing knowledge we base policies on is archived in databases that only deal with species, so that conservation strategies neglect other important levels of biological diversity.

The main objective of the course is to develop some key computational and informatics skills to form the next generation of environmental researchers and managers. In particular, students will:

• Learn and creatively use a structured vocabulary that captures the concepts and relationships that constitute ecological knowledge for a specific problem or model;
• Use such knowledge to fully characterize the logical structure of an ecological dynamic model;
• Participate in creating an interactive, web-accessible simulation modeling environment to make the models developed accessible to students and researchers worldwide.

The skills learned during the course will be applicable to other realms, from academic studies to every-day life. Another important objective of this course is to start developing a protocol for collaborative knowledge modeling, a crucial requirement that has not been satisfactorily addressed to date. We plan to use the students' feedback to design protocols and software tools to make this key activity easy and practical. We expect that this effort will produce a peer-reviewed research paper on a major journal; all students will be given the opportunity to contribute to the paper after the course has ended.

We will use this course to start introducing the novel discipline of ecoinformatics in the UVM curriculum. We expect it to evolve into a full 3-credit ecoinformatics course to be taught at UVM on a regular basis, part of an integrated teaching offer that the Ecoinformatics Collaboratory is preparing for 2005.

This course will introduce students to the fledgling discipline of Ecoinformatics, with an emphasis on the collaborative development of knowledge maps (ontologies) in a classroom and workshop setting. The course will be taught 1h/week on days to be determined for 12 weeks. There will be a total of 12 lectures, 1h each, with most class time devoted to project activities. During the course, we expect students to become conversant with the following topics, by guided discussion, some formal lecturing, and lots of hands-on experimentation:

• Ecoinformatics and Bioinformatics: 2 emerging IT disciplines
• Data, metadata to ontologies: role of knowledge modeling in next-generation environmental research and policymaking.
• Knowledge representation: concepts, languages and tools.
• Enabling the incentive structure to data sharing: the open-source revolution.