Jordan Robinson

I'm an Artificial Intelligence (AI) researcher that specialises in bespoke multi-agent Large Language Model framework development, Computational Models of Argument, Argumentation Theory, and Natural Language Processing. My research regards the design, development and testing of decision-support systems for anyone, such as intelligence analysts, investigators and businesses, who need to conduct defeasible reasoning under uncertainty. I recently finished writing my thesis and passed my viva voce with minor corrections. I'm now working at the University of Liverpool as a postdoc, researching the persuasiveness of Large Language Models in a number of settings.

Education

• Computer Science, PhD - University of Liverpool (2020 - 2024)
• My PhD project concerned the design, development and testing of AI tools, techniques and algorithms to provide decision- and reasoning-support to intelligence analysts, with a specific focus on aiding the all-source analyst when conducting hypothesis generation and evaluation. Summaries of each chapter are listed below:
  • A survey of the literature related to intelligence analysis, argumentation theory, computational models of argument, and AI tools for intelligence analyses.
  • A proposal for an argumentation scheme from multiple causes to effect to represent hypotheses that can be supported by a varying number of items of information, evidence, and assumptions.
  • A study which assessed the feasibility of leveraging large language models for hypothesis generation.
  • An algorithm for the refinement of intelligence analysts' prior probabilistic beliefs using the acceptability of arguments discovered through the semantic evaluation of probabilistic argumentation frameworks.
  • An algorithm for conducting sensitivity analysis on Dung's original abstract argumentation framework.
• The project was funded by the EPSRC Centre for Doctoral Training in Distributed Algorithms (DA CDT) and Dstl. Being a part of the DA CDT meant that I had the pleasure of undertaking a three-month placement during my time as a student, some of the highlights were:
• Setting out a requirement for £200-300k of research on a topic closely aligned to my PhD research.
• Using my expertise to provide technical support to a number of ongoing projects.
• Going to a hackathon, sponsored by an unnamed Big Tech company, where I was able to speak with intelligence analysts and use some cool tools.
• (Living in a house with a thatched roof!)
• I was supervised by Katie Atkinson (University of Liverpool), Simon Maskell (University of Liverpool), Chris Reed (University of Dundee), and Todd Robinson (Dstl)
• A copy of my thesis (.pdf)
• Mechanical Engineering, MEng - University of Liverpool (2016 - 2020)

Employment

• Postdoctoral Research Associate, Department of Computer Science, University of Liverpool (January 2025 - Present)

Conference and Workshop Papers

• 2024
• J Robinson, K Atkinson, S Maskell, C Reed. On Diagnostic Arguments in Abstract Argumentation. In Proc. of COMMA. 2024. (.pdf)
• J Robinson, K Atkinson, S Maskell, C Reed. Identifying Diagnostic Arguments in Abstract Argumentation. In Proc. of COMMA. 2024. (.pdf)
• 2022
• L Bengel, E Bezou-Vrakatseli, L Blümel, F Castagna, G D'Agostino, D Odekerken, MS Patil, J Robinson, H Wu, A Xydis. Online Handbook of Argumentation for AI: Volume 3. 2022. (.pdf)
• 2021
• J Robinson. Distributed Hypothesis Generation and Evaluation. JURIX Doctoral Consortium 2021. (Won best student paper award, .pdf)

Public Engagements

• An article I wrote about starting a PhD during a pandemic (link)

Contact Information

• Feel free to contact me via email: Jordan {dot} Robinson {at} liverpool {dot} ac {dot} uk