TheFraunhoferInstituteapproachedustopresenttheirexpertiseinartificialintelligenceandmachinelearninginanewandinnovativewayontheweb.
Client
Fraunhofer Institut - ML2R
Industry
Public Sector and Non-Profit
Services
Technical ConsultingUX DesignVisual DesignSoftware DevelopmentDevOpsProducing
Research that sparks curiosity.
Research often remains invisible.
ML2R aimed to make complex AI research accessible to a broad audience from society, authorities, and business. Long papers do not reach all target groups today. The goal was to ignite curiosity, promote understanding, and increase measurable interaction through web touchpoints.
We developed an interactive web experience that presents the study 'How to Teach AI Humor?' as a narrative experiment. Users experience a 3D cartoon as an emotional introduction and can then explore the research interactively.
An interactive web experience explains AI through humor.
The 3D animated cartoon video serves as the emotional entry point: a short, structured punchline explains how humor is modeled semantically and topologically. The video and accompanying texts are available in German and English.
A package of story, visuals, and interaction.
The result includes a 3D animated cartoon video, an interactive timeline, scientific infographics, and a WebGL visualization (rotating Venn sphere). It features a basic and an expert mode, mobile optimization, and bilingual content.
The rotating sphere makes abstract data tangible: users tap or click to open spheres and discover variables such as vector relationships, embeddings, and dynamic behavior. Scientific graphics remain explanatory and accessible.
After the launch, the average dwell time increased by about 30% and the interaction rate within the experience rose by around 25%. Both scientists and users reported better understanding, and the experience generated additional media attention.
Measurable impact on reach and understanding.
This is how the process went.
In discovery, we identified relevant use cases and target groups; in the prototype, we validated the story and interaction ideas; in the build, we integrated WebGL visualizations and optimized assets; and in the launch, we delivered a bilingual, high-performance experience.
The timeline synchronizes joke sequences with ML concepts: vectors, embeddings, topological relationships, and dynamic behavior become visible along the narrative, allowing both laypeople and professionals to follow and dive deeper.
Technically, the experience is based on WebGL rendering with optimized assets from Blender, responsive design, and a clear structure for basic and expert modes. Integration and performance were the focus to ensure that the visualizations remain accessible on mobile devices.
Technology and insights lead to real learning success.
Interested in your own research experience?
Talk to us about your research showcase.
We create interactive experiences that make research accessible and present expertise in an engaging way. Contact us for concepts, prototypes, and implementations to showcase your research tailored to your audience.