In this collaboration with Adobe Corporation, we are working to improve the evaluation of LLM outputs and the creation of personalized multimedia scientific documents with a human author in the loop.
Automating the creation of scientific diagrams from academic papers can streamline the development of tutorials, presentations, and posters.  Current text-to-image models struggle with generating accurate and visually appealing diagrams from long-context inputs. We develop tools to render these diagrams and images automatically from user input and build more efficient mechanisms to detect the quality of the LLM generations.