Lack of interpretability limits the use of common unsupervised machine learning techniques in contexts where they are meant to augment managerial decision-making. DK Lee will present a generative deep learning model based on a Variational AutoEncoder (“InnoVAE”) that converts unstructured text about patents into an interpretable spatial representation of innovation space to enable high-resolution exploration at the patent, innovation, and firm level. The same methodology can be applied to various multi-modal business objects such as companies (strategy space), jobs (skill space), products (feature space), and more.

Download the Presentation

Watch the Recording

Read the Summary