DAIREL is the Data Science and Artificial Intelligence research group at the University of Wolverhampton in the UK. The group focuses on all aspects of data, information and the interaction with it. Our aim is to reduce information overload caused by the huge amount of heterogeneous data constantly generated in all disciplines, and support users in accessing relevant information effectively and efficiently. We provide expertise and conduct research in deep machine learning, generative AI, user-centric information retrieval and natural language processing.

Recent developments in large language models, artificial intelligence, machine learning, human-computer interaction and information retrieval are continuing to transform our information society on several levels. To exploit the opportunities and address the challenges these developments bring with it, DAIREL’s research bridges the gap between data-driven and user-driven approaches to support users’ interaction with data and information in Big Data collections, in a fair and unbiased manner. DAIREL therefore follows a holistic view that integrates user as well as system aspects, which should not be seen in isolation. In the nature of the thought experiment of Schrödinger’s Cat, the different n-grams of the DAIREL acronym can be interpreted differently at the same time:

  • DA stands for Data, the building block of our work and the fuel of our information society; data can be structured, semi-structured and unstructured, textual or multimedia, as well as heterogeneous - turning data into information is part of our mission;
  • AI stands for Artificial Intelligence, including deep machine learning, reinforcement learning, transformers, embeddings, generative AI, probabilistic logics and other ‘intelligent’ data processing and generative technologies;
  • I stands for Interaction, to emphasise our human-centric approach to investigate how users are interacting with data and information for their tasks, and to provide fair and unbiased solutions for effective access to, and interaction with, data and information;
  • IR stands for Information Retrieval, the task of satisfying users’ information needs;
  • REL stands for Relevance, a core concept in Information Retrieval with its aim to divide relevant from irrelevant information;
  • L stands for Language, including (generative) natural language processing, word embeddings and large language models.

We think all these different parts that make DAIREL should not be seen in isolation, but need to be integrated.

Research Topics

DAIREL’s mission is to provide solutions to support users in their different tasks and domains when dealing with data and information. Our research topics and interests include

  • User-centric Information Access and Retrieval
    • Retrieval-augmented generation
    • Interactive Models for multimodal information access inspired by Quantum Mechanics
    • Integration of cognitive theories (e.g, Polyrepresentation and Information Foraging)
    • Bibliometric-enhanced Information Retrieval
    • Cross-cultural Information Retrieval
    • Conversational agents
  • Artificial Intelligence and Machine Learning
    • Probabilistic Logics
    • Reinforcement Learning for Multimodal Retrieval
    • Deep Machine Learning for Information Retrieval and Recommender Systems
    • Neural models for Natural Language Processing
    • Responsible AI and Information Retrieval
  • Text Analysis and Natural Language Processing
    • Large Language Models and Embeddings
    • Information Extraction
    • Text Categorisation

Example Applications

  • Digital Humanities
  • Digital Libraries and Digital Museums
  • Authorship Attribution
  • Scholarly Search and Recommendation
  • Cyberstalking Detection
  • Sarcasm Detection
  • Expert Search
  • Machine Translation