PhD Cross-modal learning for information retrieval

Cross-Modal Retrieval is the task of learning and retrieving data across different modalities, such as from images, video, audio, and text. It is a challenging task to learn representations from different modalities in the shared subspace. The project may focus on:

(1) Developing methods for learning relations of entities in cross-modal spaces comprising images and text. Identifying the limitations of existing cross-modal retrieval algorithms.
(2) Investigating bias in cross-modal transfer learning and developing strategies for detecting and mitigating bias
(3) Proposing and implementing strategies for cross-modal lifelong learning and for forgetting information when required.
(4) Development of a tool for the search, and retrieval of multi-modal data. The tool will have query expansion and relevance feedback capabilities.

Applicants must have, or expect to achieve, a 2:1 Honours degree (or equivalent) in computer science, information science, or mathematics and have strong programming skills. Students with a first-class honours degree who don't have an MSc degree are welcome to apply

IELTS - 6.5 overall with 6.0 in each individual element (reading, writing, listening and speaking).
TOEFL - iBT - Overall score of 92, with not less than 22 in each test.