Narrative Understand ing with Knowledge Graphs
Computer scientist Inès Blin shows that giving AI a structured memory of facts and relations hips can make its answers more accurate and easier to verify.
Blin’s research focused on how AI systems can better suppor t human sense-making. People naturally understand the world by linkin g new events or information to what they already know, and turning th is into a coherent narrative (for example: what happened, why it happ ened, and why it matters). Today’s AI can generate fluent text, but it may also invent details or struggle to explain where its answers come from. I investigated how combining structured knowledge (such as knowledge graphs) with other AI methods can help systems build narra tives that are more reliable, transparent, and useful across differen t domains. The motivation was to design AI that can use structured me mories to support explanations, generate hypotheses, and analyse deba tes in ways that people can trust.
Useful alternatives
by AI
Her research showed that giving AI a structured mem
ory of facts and relationships can make its answers more accurate and
easier to verify. Instead of generating text directly, the systems B
lin built first collect relevant information and organise it into a s
tructured map of key entities and their connections. She tested this
approach in three domains: history, social media discussions, and soc
ial science research. In the historical domain, structuring informati
on improved the relevance of what the system retrieved and reduced fa
ctual errors. In the social media domain, it helped make complex deba
tes easier to explore. In the social science domain, AI-generated hyp
otheses did not always outperform human ones, but they often added us
eful alternatives, showing strong potential for human–AI collaborat
ion.
These findings matter for anyone who uses AI to understand complex topics, especially when trust and clarity are important. For everyday users, structured narrative representations can help AI exp lain historical events in a clearer and more accurate way, instead of producing confident but incorrect answers. For expert users, the sam e approach can support tasks such as summarising large public debates , like social media discussions about inequality, or helping research ers generate new ideas in social science. In practice, this could lea d to tools that help users quickly navigate large amounts of informat ion, understand the main viewpoints, and see how claims connect to ev idence. These applications are realistic in the near future, because they build on existing AI systems and improve them with structured kn owledge.
Collaboration with domain experts
Blin conducted her research using a mix of literature review, compute
r-based experiments, and user studies. First, she reviewed existing r
esearch on narratives and how they can be represented computationally
. Then she developed methods to retrieve relevant information and con
vert it into structured knowledge representations, and tested them ac
ross several real-world use cases. She evaluated the results using qu
antitative measures as well as qualitative analysis. For the qualitat
ive analyses, she ran user studies to assess how helpful the system o
utputs were, both for AI-generated hypotheses in social science and f
or the quality of answers in the historical domain. Lastly, Blin coll
aborated with domain experts when needed to ensure the results were m
eaningful and realistic in practice.
More information on the thesis
DESCRIPTION: Narrative Understanding with Knowledge Graphs Com puter scientist Inès Blin shows that giving AI a structured memory o f facts and relationships can make its answers more accurate and easi er to verify.Blin’s research focused on how AI systems can better support human sense-making. People naturally understand the w orld by linking new events or information to what they already know, and turning this into a coherent narrative (for example: what happene d, why it happened, and why it matters). Today’s AI can generate fl uent text, but it may also invent details or struggle to explain wher e its answers come from. I investigated how combining structured know ledge (such as knowledge graphs) with other AI methods can help syste ms build narratives that are more reliable, transparent, and useful a cross different domains. The motivation was to design AI that can use structured memories to support explanations, generate hypotheses, an d analyse debates in ways that people can trust.Useful altern atives by AI