School of Electronic Engineering and Computer Science

Dr Thomas Roelleke

Thomas

Senior Lecturer

Email: t.roelleke@qmul.ac.uk
Telephone: +44 20 7882 7988
Room Number: Peter Landin, CS 423
Website: http://www.eecs.qmul.ac.uk/~thor
Office Hours: Wednesday 11:00-13:00

Teaching

Database Systems (Postgraduate)

Introduction to databases and their language systems in theory and practice. The main topics covered by the module are: The principles and components of database management systems. The main modelling techniques used in the construction of database systems. Implementation of databases using an object-relational database management system. SQL, the main relational database language. Object-Oriented database systems. Future trends, in particular information retrieval and data warehouses. There are 2 timetabled lectures a week, and 1 hour tutorial per week (though not every week). There will be timetabled laboratory sessions (2 hours a week) for approximately 4 weeks.

Database Systems (Undergraduate)

This module is an introduction to databases and their language systems in theory and practice. The main topics covered by the module are: the principles and components of database management systems; the main modelling techniques used in the construction of database systems; implementation of databases using an object-relational database management system; the main relational database language; Object-Oriented database systems; future trends, in particular information retrieval, data warehouses and data mining.There are two timetabled lectures a week, and one-hour tutorial per week (though not every week). There will be timetabled laboratory sessions (two hours a week) for approximately five weeks.

Research

Research Interests:

My research focuses on two related areas: 1. information retrieval (IR) models and 2. the integration of database (DB) and IR technologies.

IR models are related to probability theory and the sound derivation of IR models leads to new and general approaches to rank any object, to reason about complex knowledge sources, and to make decisions. Many results of my research over the past 10 years are summarised in the book "IR Models: Foundations and Relationships", Morgan Claypool Publishers, 2013. Currently, my main research interest is in generalisations of probability theory in order to obtain a "new" theory that joins probabilistic and information-theoretic reasoning.

The integration of DB and IR is an ongoing research challenge, though, in principle, DB and IR do the same: manage and retrieve data. I have developed probabilistic object-relational, logic-based knowledge representations that are useful for solving tasks in the domain of "semantic" (knowledge-rich) information management tasks. This led to the "Relational Bayes", a patented technology (VLDB Journal 2008).

Based on the insights into probabilistic reasoning and IR models, and based on the benefits from a seamless DB+IR, we have developed an information management system that we provide to selected collaborators and customers.

Publications

  • Frommholz I, Roelleke T (2016). Scalable DB+IR Technology: Processing Probabilistic Datalog with HySpirit.. nameOfConference
  • Roelleke T, Kaltenbrunner A, Baeza-Yates R (2015). Harmony Assumptions in Information Retrieval and Social Networks. nameOfConference
  • Milajevs D, Sadrzadeh M, Roelleke T (2015). IR meets NLP: On the semantic similarity between subject-verb-object phrases. nameOfConference
  • Roelleke T, Bonzanini M, Alvarez MM (2013). On the modelling of ranking algorithms in probabilistic datalog. nameOfConference
  • Martinez-Alvarez M, Bonzanini M, Roelleke T (2013). Mathematical specification and logic modelling in the context of IR. nameOfConference
  • Martinez-Alvarez M, Bellogin A, Roelleke T (2013). Document difficulty framework for semi-automatic text classification. nameOfConference
  • Bonzanini M, Martinez-Alvarez M, Roelleke T (2013). Extractive summarisation via sentence removal: Condensing relevant sentences into a short summary. nameOfConference
  • Bonzanini M, Martinez-Alvarez M, Roelleke T (2012). Investigating the use of extractive summarisation in sentiment classification. nameOfConference
  • Bonzanini M, Martinez-Alvarez M, Roelleke T (2012). Opinion summarisation through sentence extraction: An investigation with movie reviews. nameOfConference
  • Martinez-Alvarez M, Yahyaei S, Roelleke T (2012). Semi-automatic document classification: Exploiting document difficulty. nameOfConference
  • Azzam H, Yayhaei, Roelleke et al. (2012). A Schema-driven Approach for Knowledge-oriented Retrieval and Query Formulation. KEYS 2012, The 3rd International Workshop on Keyword Search and Structured Data
  • Martinez-Alvarez M, Roelleke T (2011). A descriptive approach to classification. nameOfConference
  • Azzam H, Roelleke T (2011). A Generic Data Model for Schema-Driven Design in Information Retrieval Applications. nameOfConference
  • Yahyaei S, Bonzanini M, Roelleke T (2011). Cross-Lingual Text Fragment Alignment Using Divergence from Randomness. nameOfConference
  • Smeraldi F, Martinez-Alvarez M, Frommholz I et al. (2011). On the probabilistic logical modelling of quantum and geometrically–inspired IR. nameOfConference
  • Martinez-Alvarez M, Roelleke T (2010). Modelling probabilistic inference networks and classification in probabilistic datalog. nameOfConference
  • Klampanos IA, Wu HZ, Roelleke T et al. (2010). Logic-Based Retrieval: Technology for Content-Oriented and Analytical Querying of Patent Data. nameOfConference
  • Gurrin C, He YL, Kazai G et al. (2010). Recent Developments in Information Retrieval. nameOfConference
  • Forst JF, Tombros A, Roelleke T (2009). Less Is More: Maximal Marginal Relevance as a Summarisation Feature. nameOfConference
  • Wu HZ, Roelleke T (2009). Semi-subsumed Events: A Probabilistic Semantics of the BM25 Term Frequency Quantification. nameOfConference
  • Amer-Yahia S, Hiemstra D, Roelleke T et al. (2008). DB&IR Integration: Report on the Dagstuhl Seminar "Ranked XML Querying". nameOfConference
  • Roelleke T, Wu H, Wang J et al. (2008). Modelling retrieval models in a probabilistic relational algebra with a new operator: the relational Bayes. nameOfConference
  • ROELLEKE T, Wang J (2008). TF-IDF Uncovered: A Study of Theories and Probabilities. 31st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval
  • ROELLEKE T, Wang J (2006). A Parallel Derivation of Probabilistic Retrieval Models. 29th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, Seattle, US
  • Rolleke T, Tsikrika T, Kazai G (2006). A general matrix framework for modelling Information Retrieval. nameOfConference
  • Wang J, Roelleke T (2006). Context-specific frequencies and discriminativeness for the retrieval of structured documents. nameOfConference
  • Amer-Yahia S, Case P, Rolleke T et al. (2005). Report on the DB/IR panel at SIGMOD 2005. nameOfConference
  • ROELLEKE T, de Vries A (2005). Relevance Information: A Loss of Entropy but a Gain for IDF?. 28th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, Salvador, Brazil
  • Szlavik Z, Rolleke T (2005). Building and experimenting with a heterogeneous collection. nameOfConference
  • Lalmas M, Rolleke T (2004). Modelling vague content and structure querying in XML retrieval with a probabilistic object-relational framework. nameOfConference
  • ROELLEKE T (2003). A Frequency-based and a Poisson-based Definition of the Probability of Being Informative. 26th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, Toronto, Canada
  • Lalmas M, Rolleke T (2003). Four-valued knowledge augmentation for structured document retrieval. nameOfConference
  • LALMAS M, Roelleke T, Ruthven I (2003). Abductive retrieval for multimedia information seeking. 10th International Conference on Human - Computer Interaction, HCI International, Crete, Greece, vol. 4
  • Pearmain A, Lalmas M, Moutogianni E et al. (2002). Using MPEG-7 at the consumer terminal in broadcasting. nameOfConference
  • Lalmas M, Roelleke T (2002). Four-valued knowledge augmentation for representing structured documents. nameOfConference
  • Lalmas L, ROELLEKE T, Fuhr N (2002). Intelligent Hypermedia Retrieval. nameOfConference
  • Roelleke T, Lalmas M, Kazai G et al. (2002). The accessibility dimension for structured document retrieval. nameOfConference
  • Kazai G, Lalmas M, Rolleke T (2001). A model for the representation and focussed retrieval of structured documents based on fuzzy aggregation. nameOfConference
  • Lalmas M, Rolleke T, Turra F et al. (2001). Concepts for a graphical user interface for hypermedia retrieval. nameOfConference