About the project
This project will explore how multimodal large language models (LLMs) can understand arguments across text, audio, images, and video. You'll develop methods using LLMs, vision-language models (VLMs), multimodal reasoning, and argument mining to identify claims, evidence, and relationships.
Multimodal LLMs can process text, speech, images, and video, but still struggle to understand how claims, evidence, and counterarguments connect across these different forms of communication. Addressing this problem requires models that can identify argumentative structures, integrate information across modalities, reason with this information and explain what evidence supports their conclusions.
In this project, you'll develop and evaluate new methods for multimodal argument understanding. Depending on your interests and experience, you could investigate:
- large language models and multimodal reasoning for identifying claims, premises, evidence, and argumentative relationships
- connecting spoken or written claims with visual evidence
- analysing arguments across long complex discussions
Possible methods include multimodal reasoning, multimodal CoT distillation, multimodal retrieval, structured prediction, retrieval-augmented generation, and combinations of neural and symbolic approaches.
You'll have flexibility to shape precise methods, modalities, and applications. Potential directions include:
- developing new datasets and evaluation methods
- improving cross-domain generalisation
- testing whether models use multimodal evidence faithfully
- designing more interpretable and reliable systems
Applications might include online discussion, political debate, public deliberation, or other forms of written, spoken, and visual communication.
The project offers the opportunity to contribute across natural language processing, multimodal machine learning, computer vision, and audio analysis. You'll develop a coherent technical contribution around your strengths while gaining experience of interdisciplinary research and the wider implications of AI for real-world communication.
You'll join the School of Electronics and Computer Science. You'll have access to professional-development opportunities that may include teaching experience, engagement with networks such as the Alan Turing Institute and Responsible AI UK programme, and support from Future Worlds to explore the commercial potential of your research.
The School of Electronics and Computer Science is committed to promoting equality, diversity inclusivity as demonstrated by our Athena SWAN award. We welcome all applicants regardless of their gender, ethnicity, disability, sexual orientation or age, and will give full consideration to applicants seeking flexible working patterns and those who have taken a career break. The University has a generous maternity policy, onsite childcare facilities, and offers a range of benefits to help ensure employees’ well-being and work-life balance. The University of Southampton is committed to sustainability and has been awarded the Platinum EcoAward.
