Iterative Neural Scoring of Validated Insight Candidates

Allmin Susaiyah, Aki Harma, Ehud Reiter, Milan Petkovic

Research output: Contribution to conferenceUnpublished paperpeer-review


Automatic generation of personalised behavioural insight messages is useful in many applications, for example, health selfmanagement services based on a wearable and an app. Insights should be statistically valid, but also interesting and actionable for the user of the service. In this paper, we propose a novel neural network approach for joint modeling of these elements of the relevancy, that is, statistical validity and user preference, using synthetic and real test data sets. We
also demonstrate in an online learning scenario that the system can automatically adapt to the changing preferences of the user while preserving the statistical validity of the mined insights.
Original languageEnglish
Number of pages10
Publication statusAccepted/In press - 17 Aug 2020
EventIntelLanG : Intelligent Information Processing and Natural Language Generation - Santiago de Compostela, Spain
Duration: 7 Sept 20207 Sept 2020


CitySantiago de Compostela
Internet address

Bibliographical note

This work was supported by the Horizon H2020 Marie Skłodowska-Curie Actions Initial Training Network European Industrial Doctorates project under grant agreement No. 812882 (PhilHumans).


Dive into the research topics of 'Iterative Neural Scoring of Validated Insight Candidates'. Together they form a unique fingerprint.

Cite this