Join CDS Women in Data Science for Another Lunch and Learn Talk:
Friday, July 24th 12pm-1pm EST
Speaker: Dr. Elena Sizikova, CDS Faculty
Neural network models recently became ubiquitous for object recognition, detection, and other automatic computer vision tasks and they have also been successful in healthcare and related technology applications. However, understanding and comparing networks to humans remains a challenging and unsolved problem and an active research area today. We study and show how networks differ from humans on the task of reading, in hopes of building an accurate computational model that accounts for different speeds of reading, such as slower reading resulting from dyslexia. We summarize related work in this domain and present two particular differences between the two models, based on perception of clutter (efficiency) and robustness to noise.
All genders are welcome to attend. Please sign up at this link to receive the Zoom invitation.