วีดีโอ

for creating a binary mask which tool or method is used?

Hi, Do you have the dataset published somewhere we could access?

流利度还是差很多，感觉是一个单词一个单词在念啊~

Good for you, we can't test it, it is not open..

The model is not openly available. My institute does not subscribe to CITI training and thus I am not able to access the model on PhysioNet. Why does this model need to be behind such lockdown?

Is it available & accessible for everyone?

Why not make the PowerPoint slides and programs ，dataset from the video publicly available? let us study

This is why you need strong moderators. Let her finish her presentation, then ask your questions. She handled it so well and professionally, but the interruptions are beyond unprofessional.

🎉🎉🎉

Have questions at the end. The questioners had no idea what they were talking about and derailed the presentation. Great work Qianqian Xie, you did fantastic

Good video, but the lady who kept interrupting was too loud and spoke very fast. I couldn't understand what she was trying to achieve while constantly interrupting. Please change the format to allow questions to be asked at set intervals or at the end of the presentation.

Thank you for this resource presentation. I would like to ask if you observed any performance gaps depending on the datasets used, particularly with the MTsample dataset, which is synthetic.

Prompt engineering differs from prompt tuning

Excellent presentation, well understandable

thankyou for the session, please provide the code.

Thanks stanf and Mihir! Amazing!

084 Veum Drive

Please can you share the paper,GitHub, or something? This is very interesting

Can we get a certificate on completing the course or no?

Very ingenious to use the image to make more data about the sample

3 months of cuda development! wow, that's rough. hope we get it working well in compilers

This is really good

Good job stanford MedAI and Dr. Zifeng Wang. Keep contributing

impressive

希望有中文字幕，英文听力不好。

Great video! IIRC Paul co-created the current CMU Multimodal course, which is available on TH-cam. This is 80% overlap with lecture 2.

Tri Dao yyds!

Interesting idea

优秀的一仔

at 5:18, Bayesian inference point, it should be \delta_x log p(y|x) on the right hand side of equation. In other words, \delta_x log p(x|y) = \delta_x log p(x) + \delta_x log p(y|x)

so good

Very interesting!

Wow, impressed by Max...I hope I can call him that!

Insightful contribution. Keep it up Dr. Anwar.

impressive presentations. thank you

This topic is really interesting! I can understand easily although I am not good at data set!

great presentation!

Thanks.

Excellent presentation and impressive research, i only wonder why SSMs are recurrently efficient? (video timestamp : 32:27) Suppose k is the token len of input history. The general sequence model takes k square (s.t. transformer) time complexity. On the other hand, SSMs still need to encode all stateful history "recurrently". The S4 paper also aims to deal with this issue (multiply A, k-1 times to create a K bar matrix, it also ends in nearly k square) by diagonalizing the matrix. So, it seems SSMs recurrent aren't "naturally" efficient, but require some linear algebra technique. Any suggestion will be appreciated!!

he is one of the heads of this new Mamba architecture

quality, well structured presentation. Thanks Jason.

Awesome.. What's GAN to convert CT sinogram to image, thank you

Super interesting! Thanks for the presentation. I work in game development for now, but cool to see how things are going in the ML world 😊

excellent presentation. Thank you

Excellent presentation

couldn't imagine how interesting and informative this video is.

Where did you get the data set sir

Thanks for such a informative session

If I am not mistaken, this covers Foundation Models for images and language. Any attempt to build foundation models for time-series health data? Thanks in advance

It is a grear work ! Thank you!