Narrative Understanding with Knowledge Graphs
Computer scientist Inès Blin s hows that giving AI a structured memory of facts and relationships ca n make its answers more accurate and easier to verify.
Blin’s research focused on how AI systems can better support human sense-making. People naturally understand the world by linking new e vents or information to what they already know, and turning this into a coherent narrative (for example: what happened, why it happened, a nd why it matters). Today’s AI can generate fluent text, but it may also invent details or struggle to explain where its answers come fr om. I investigated how combining structured knowledge (such as knowle dge graphs) with other AI methods can help systems build narratives t hat are more reliable, transparent, and useful across different domai ns. The motivation was to design AI that can use structured memories to support explanations, generate hypotheses, and analyse debates in ways that people can trust.
Useful alternatives by AI
Her research showed that giving AI a structured memory of
facts and relationships can make its answers more accurate and easier
to verify. Instead of generating text directly, the systems Blin bui
lt first collect relevant information and organise it into a structur
ed map of key entities and their connections. She tested this approac
h in three domains: history, social media discussions, and social sci
ence research. In the historical domain, structuring information impr
oved the relevance of what the system retrieved and reduced factual e
rrors. In the social media domain, it helped make complex debates eas
ier to explore. In the social science domain, AI-generated hypotheses
did not always outperform human ones, but they often added useful al
ternatives, showing strong potential for human–AI collaboration.
These findings matter for anyone who uses AI to understand comple x topics, especially when trust and clarity are important. For everyd ay users, structured narrative representations can help AI explain hi storical events in a clearer and more accurate way, instead of produc ing confident but incorrect answers. For expert users, the same appro ach can support tasks such as summarising large public debates, like social media discussions about inequality, or helping researchers gen erate new ideas in social science. In practice, this could lead to to ols that help users quickly navigate large amounts of information, un derstand the main viewpoints, and see how claims connect to evidence. These applications are realistic in the near future, because they bu ild on existing AI systems and improve them with structured knowledge .
Collaboration with domain experts
Blin co
nducted her research using a mix of literature review, computer-based
experiments, and user studies. First, she reviewed existing research
on narratives and how they can be represented computationally. Then
she developed methods to retrieve relevant information and convert it
into structured knowledge representations, and tested them across se
veral real-world use cases. She evaluated the results using quantitat
ive measures as well as qualitative analysis. For the qualitative ana
lyses, she ran user studies to assess how helpful the system outputs
were, both for AI-generated hypotheses in social science and for the
quality of answers in the historical domain. Lastly, Blin collaborate
d with domain experts when needed to ensure the results were meaningf
ul and realistic in practice.
More information on the thesi s
DESCRIPTION: Computer scientist Inès Blin shows that giving A I a structured memory of facts and relationships can make its answers more accurate and easier to verify. Blin’s research focus ed on how AI systems can better support human sense-making. People na turally understand the world by linking new events or information to what they already know, and turning this into a coherent narrative (f or example: what happened, why it happened, and why it matters). Toda y’s AI can generate fluent text, but it may also invent details or struggle to explain where its answers come from. I investigated how c ombining structured knowledge (such as knowledge graphs) with other A I methods can help systems build narratives that are more reliable, t ransparent, and useful across different domains. The motivation was t o design AI that can use structured memories to support explanations, generate hypotheses, and analyse debates in ways that people can tru st. Useful alternatives by AI