Biomedical and E-Health Abstractions and Modelling: Engineering and Research (BEAMER)
Date: 18th June 2010,
Time: 9:30 am
Location: Royal Society of Edinburgh, 22-26 George Street, Edinburgh EH2 2PQ
Booking Status: Expired
This workshop is sponsored by the Scottish Informatics & Computing Science Alliance (SICSA)
The workshop will cover
• Exploration of issues arising from different levels of abstraction in the collection of raw biomedical data and the analysis of, and reasoning over, these data sets.
• Fostering collaborations between the research groups involved in the collection, processing and storage of raw biomedical data and groups focusing on the representation and reasoning of knowledge that can be derived from this data.
• Encouragement of interaction between life science and computational research groups.
Background:
In our communities we do a wide spectrum of work from gathering and interpreting data through to developing sophisticated reasoning algorithms designed to manipulate knowledge. The gap between data and knowledge can be very wide and the different areas of expertise, data gathering and reasoning, often do not interact. The consequence is that many reasoning algorithms are based on abstractions that are disconnected from reality, and those who are working with large data sets are building different abstractions of the data that cannot be exploited by these algorithms. This makes it hard for the two communites to benefit from each others' work.
The abstract symbols used in model-checking, planning, constraint reasoning, and other combinatorial problem solving systems, must be grounded in reality if these approaches are to be applied to data rich domains such as bio-informatics. A predicate logic, or process algebra, representation being used by a planner or a model-checker, has to map down to patterns in data that imply the statements in the abstract representation. The data itself, after it has been cleaned and analysed, must imply truths that correspond to the abstract statements. Bridging this gap poses many research challenges and draws on expertise in data analysis, signal processing, machine learning, management of uncertainty and knowledge representation.
Participants:
The SICSA community includes experts in combinatorial problem solving based on abstract representations. The SULSA community includes experts in data collection and analysis, signal processing and pattern recognition. We propose a workshop to bring together the skills and concerns of both these communities, with the objective of learning how to acquire meaningful abstract representations by automating the cleaning, processing, partitioning and classification of raw data.
This workshop will be of value to all those in the life sciences who want to use machine learning and reasoning algorithms to identify and reason about abstract patterns in their data. It will also be of value to computer scientists who develop abstract reasoning algorithms without knowing how to ground the abstract symbols in their representations.
** For program please see link below **
