ML Paper Reading & Discussion: 24/003

ML Paper Reading & Discussion: 24/003
DetailsAgendaSpeakersCollaborationsComments
Details

Join us for our weekly AI/ML paper reading session, where enthusiasts, students, and professionals gather to dive into the latest and most impactful research in artificial intelligence and machine learning. Each week, we’ll explore a selected paper, discussing its methodologies, findings, and real-world applications in a collaborative and friendly environment.

This session is perfect for:

  • Anyone looking to deepen their understanding of AI/ML through hands-on analysis of research.
  • Those who want to stay updated with cutting-edge developments in the field.
  • Beginners, who are eager to gain insights into how research shapes AI/ML advancements.

Format: Each session will include a brief overview of the selected paper, followed by group discussions to clarify concepts, share insights, and ask questions.

Let’s build a vibrant AI/ML community together in Lucknow!

Event Link: meet.google.com/jvf-ypxq-jzc . Please, Join 5 mintes early at 6:55pm.  

Agenda

RMSNorm: A Faster, Efficient Alternative to Layer Normalization. Speaker:- Harshita Singh

12:30

13:00

The session will begin with an Introduction to LayerNorm, providing a brief overview of its core principles, benefits, and limitations. This will be followed by an exploration of RMSNorm, detailing its fundamental concept and the differences it brings compared to LayerNorm. The discussion will highlight the Key Benefits of RMSNorm, emphasizing its computational efficiency and practical advantages. Additionally, the session will introduce Partial RMSNorm (pRMSNorm) as an innovative variation. Finally, the Conclusion will recap RMSNorm's advantages, leaving participants with a clear understanding of its applications and benefits.

13:00

13:30

Discussion

Comments

Be the first to post a message!

Sunday, 17th Nov, 2024

6:00 PM - 7:00 PM IST

online

Cookies

This website uses cookies to improve your online experience. By continuing to use this website, you agree to our use of cookies. If you would like to, you can change your cookie settings at any time. Our Privacy Notice provides more information about what cookies we use.