Projektet vil udvikle en ny generation af hydrologiske modeller, der kraftigt forbedrer DMI’s nationale varsling af oversvømmelser. Metoderne baseres på scientific machine learning, hvor klassiske fysisk-baserede modeller integreres med maskinlæringskomponenter. Gevinsten er både langt hurtigere simuleringer og større nøjagtighed. Ud over et styrket dansk beredskab vil de nye metoder være til stor gavn for digitaliseringen af vand- og miljøsektoren gennem bedre brug af store mængder data.

Projektet udvikler en ny metode til at kortlægge mekaniske egenskaber i voksbelægninger ved hjælp af mikromekanisk karakterisering såsom viskoeplastiske og frakturmekaniske egenskaber under temperatur- og geometri-afhængige belastninger. Metoden skal bruges til at forstå og løse revnedannelse i PCDs nye biobaserede ostevoks, der opstår ved temperaturpåvirkning. Resultaterne vil forbedre voksens anvendelighed og understøtte PCDs udvikling af bæredygtig fødevareemballage.

Imagine reading a text that knows you are looking at it – one that adapts to how you read it, supporting your unique reading habits as you read. Our primary aim is to develop a reading app that can be used by beginning readers, particularly children who struggle with reading. A secondary potential outcome is for this application to become a radical innovation for supporting reading at large. The ScreenReads project will develop software that utilize biosensor data of eye-movements and speech recognition to identify how a child is reading in real-time. It will provide real-time support by adjusting the level of difficulty when it detects that a child is struggling. This will facilitate an experience of a fluent reading session. By understanding the nature of the struggle, the software will suggest relevant engaging exercises and provide digital support for children to supplement one-on-one adult assistance. As biosensor will inform the app about the skill level of the reader, it can reward and validate the specific desired behavior rather than rely on the clicking of buttons, as is otherwise the norm for existing learning tools. By improving reading skills through audio and visual support and biosensor-based reading exercises, we will create a learning tool that can evolve and expand existing teaching methods. The project is also likely to succeed as it is built on an existing reading application, EyeJustRead 1.0. The outcome will be a new reading application: EyeJustRead