Evotec Investor Day Presentation Deck

evotec Example MOLECULE DISCOVERY J.DISCOVERY TM • Discriminator neural network is lightly trained on human faces PAGE 90 Generative Adversarial Networks (GANS) to create faces in silico Go to www.thispersondoesnotexist.com to find out more • Generator creates images that sometimes fools the Discriminator, and learns from this experience • Discriminator is trained with more real human faces, forcing the Generator to improve . Eventually Generator can fool a human Training faces Noise MOLECULE DESIGN J.MD Generator PROCESS & PRODUCT DESIGN JP3 Discriminator Generated faces Real/ Synthetic MANUFACTURING DESIGN J.PODⓇ AB CD

View entire presentation