_hdsi-seminar-series-124-spurious-or-causal-studying-failure-modes-of-deep-learning-and-ways-to-fix-them-with-soheil-feiz.jpg "HDSI Seminar Series | Spurious or Causal? Studying Failure Modes of Deep Learning and Ways to Fix Them with Soheil Feizi from mathematics for deep learning Video Screenshot Preview")
GalleryPlay Video:
(Note: The default playback of the video is HD VERSION. If your browser is buffering the video slowly, please play the REGULAR MP4 VERSION or Open The Video below for better experience. Thank you!)
⏲ Duration: 36 min 49 sec
✓ Published: 16-Nov-2021
✓ Published: 16-Nov-2021
Description:
Over the last decade, deep models have enjoyed wide empirical success. However, innpractice, these models are not reliable due to their sensitivity against adversarial ornnatural input distributional shifts as well as a lack of meaningful reasoning behind theirnpredictions. In this talk, I will show that a root cause of these issues is the heavynreliance of deep models on non-causal and spurious features in their inferences. I willnthen explain our progress in understanding failure modes of deep
Over the last decade, deep models have enjoyed wide empirical success. However, innpractice, these models are not reliable due to their sensitivity against adversarial ornnatural input distributional shifts as well as a lack of meaningful reasoning behind theirnpredictions. In this talk, I will show that a root cause of these issues is the heavynreliance of deep models on non-causal and spurious features in their inferences. I willnthen explain our progress in understanding failure modes of deep
Share with your friends:
Whatsapp | Viber | Telegram | Line | SMS
Email | Twitter | Reddit | Tumblr | Pinterest
Related Videos
_ml-was-hard-until-i-learned-these-5-secrets.webp "Boris Meinardus")
_wei-ligang-at-michael-goedhuis.webp "Also available to view on GalleryLOG at gallerylog.com/michael-goedhuis--wei-ligang.htmlnnOn View: Wei Ligang at Michael GoedhuisnnBorn in Datong, Shanxi, in 1964, Wei Ligang has been at the forefront of contemporary ink painting’s development from its beginning on the mainland, and he was one of the organizers of the June 1999 “Bashu Parade: ’99 Chengdu Retrospective of Chinese Modern Calligraphy at the End of the 20th Century” exhibition. In 1981, at the age of 17, Wei was admitted to")
_calculus-1-full-college-course.webp "freeCodeCamp.org")
_math-videos-how-to-learn-basic-arithmetic-fast-online-tutorial-lessons.webp "The Organic Chemistry Tutor")
⏲ 1 hour 12 minutes 9 seconds 👁 2.2M
_the-map-of-mathematics.webp "Domain of Science")
_pytorch-for-deep-learning-amp-machine-learning-full-course.webp "freeCodeCamp.org")
_building-thinking-classrooms.webp "We know that problem solving is an effective way for students to learn to think mathematically and to acquire deep knowledge and understanding of the mathematics they are learning. Simply problematizing the mathematics curriculum, however, does not help constitute the practice that teachers want or students need. Equally, infusion of problem-based learning into the mathematics curriculum does not help with the transformations we want to see in our classrooms. What we need are a set of practices")
_11-introduction-to-machine-learning.webp "MIT OpenCourseWare")
_ml-was-hard-until-i-learned-these-5-secrets.jpg)
_wei-ligang-at-michael-goedhuis.jpg)
_calculus-1-full-college-course.jpg)
_math-videos-how-to-learn-basic-arithmetic-fast-online-tutorial-lessons.jpg)
_the-map-of-mathematics.jpg)
_pytorch-for-deep-learning-amp-machine-learning-full-course.jpg)
_building-thinking-classrooms.jpg)
_11-introduction-to-machine-learning.jpg)
![[+]](/images/menu.gif){kind=small}