Multimodal Video Search by Examples
Lead Research Organisation:
University of Cambridge
Department Name: Engineering
Abstract
Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
People |
ORCID iD |
| Mark Gales (Principal Investigator) |
Publications
BannĂ² S
(2024)
Towards End-to-End Spoken Grammatical Error Correction
BannĂ² S
(2023)
Towards End-to-End Spoken Grammatical Error Correction
Liusie A
(2022)
University of Cambridge at TREC CAsT 2022
Ma R
(2023)
Adapting an Unadaptable ASR System
Ma R
(2023)
Adapting an Unadaptable ASR System
Ma R
(2025)
ASR Error Correction Using Large Language Models
| Description | The search functionality supported by the MVSE demonstrator, image, video, speech and audio, are useful tools for BBC journalists to retrieve videos for archives. An evaluation was undertaken at the BBC to check the functionality of the tool, with the results analysed by the University of Ulster. |
| Exploitation Route | The demonstrator is being further refined for use by journalists within the BBC, but the underlying technology is broadly applicable. |
| Sectors | Digital/Communication/Information Technologies (including Software) |
| URL | https://mvse.ecit.qub.ac.uk |
| Description | BBC R&D Demonstrate Partner |
| Organisation | British Broadcasting Corporation (BBC) |
| Country | United Kingdom |
| Sector | Public |
| PI Contribution | THE CUED MVSE team has developed underlying speech and text analysis tools for the BBC demonstrator. |
| Collaborator Contribution | BBC are the commercial partner for the MVSE project. They are involved in the development and evaluation of the MVSE demonstrator. Member of the BBC have been directly involved in supplying data, and specifying the scope of the project. |
| Impact | The initial release of the demonstrator to the BBC was made in April 2022. The initial release was not designed for formal evaluation, but to demonstrate a proof-of-concept. A version of the demonstrator was delivered to the BBC at the end of 2023 for formal evaluation. |
| Start Year | 2021 |
| Description | Invited talk at Second Language Acquisition with Speech and Language Technologies (SLASaLT) |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | Invited talk presented research undertaken at Cambridge University. The majority of the talk was related to language learning, research funded by Automated Language Teaching and Assessment (ALTA), with contributions from the MVSE project research. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://sites.google.com/sheffield.ac.uk/slasalt-workshop-2024 |
| Description | MMVC Workshop |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | This was a workshop held at BMVC 2022. he purpose was to bring together researchers on multimodal video search. The work shop was hybrid i nature with approximately 20 people attending in-person and 20 people on-line. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://mvse.ecit.qub.ac.uk |
| Description | MVSE annual workshop |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Professional Practitioners |
| Results and Impact | This was then annual MVSE workshop with representatives from Belfast, Cambridge, Surrey and Ulster Universities and the BBC, approximately 30 people in attendance. The workshop was aimed at determining the form of the demonstrator to be evaluated by the BBC journalists. |
| Year(s) Of Engagement Activity | 2023 |
| Description | Multimodal AI Community UK |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Professional Practitioners |
| Results and Impact | Bringing together researchers and practitioners from AI, data science, and various scientific/application domains to build a vibrant community on Multimodal AI |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://twitter.com/MultimodalAI_UK/status/1708877483834048719 |