Knowledge Technologies (COMP90049)

AIMS

Much of the world's knowledge is stored in the form of unstructured data (e.g. text) or implicitly in structured data (e.g. databases). In this subject, students will learn algorithms and data structures for extracting, retrieving and analysing explicit knowledge from various data sources, with a focus on the web. Topics include: data encoding and mark-up, web crawling, regular expressions, document indexing, text retrieval, clustering, classification and prediction, pattern mining, and approaches to evaluation of knowledge technologies.

INDICATIVE CONTENT

Introduction to Knowledge Technologies; String search; Genomics; Text processing and search; Web search and retrieval; Introduction to Data Mining; Introduction to basic Probability; Classification; Association Rules; Clustering; Evaluation measures.

Examples of projects that students may complete are:

• A method for automatically predicting the geo-location of a Twitter user on the basis of their posts
• An automatic method for tagging multilingual Wikipedia documents with Wikipedia categories
• A search engine for Twitter data, which takes into account the time stamp of the query and documents
• A search engine for web user forum data
• A search engine servicing mixed monolingual queries (as in monolingual queries from a range of languages) over a large-scale document collection
• Classification and prediction of some real world problems using machine learning techniques
• Unification and resolution
• Nondeterminism, search, and backtracking