Computational Perception Models of Scientific Presentations
Project Status: completed
This project intents to identify key shortcomings of current computational models for embedding the context of concepts as presented in media documents. The way how humans parse the content of documents is fundamentally different to algorithmic approaches. We intend to a) identify prototypical models of how human perceive concepts. To identify these models, a reception study with about 25 test persons will be conducted. In this project we will focus on scientific presentations, like powerpoint, that contain a mixture of text and images. b) Formalize those models into computable algorithms. c) Identify the conceptual differences between current state-of-the-art algorithms for concept embeddings to the models derived from human perception.
from: 1 Juni 2018
until: 31 Dezember 2018
Funding: YIN Grant