Deep Probabilistic Generative Models

Dear students,

Tomorrow we will have the first panel session of the seminar. We have uploaded a PDF file compiling the pending points from the last session along with your submitted questions, so you can prepare for the discussion.

Students who presented papers in this block will act as “experts” and lead the discussion, but all students are expected to participate actively. Everyone will be evaluated based on their assigned role.

The session will begin with the presentation of the paper “VAE with a VampPrior”, followed by the panel discussion.

Looking forward to seeing you all tomorrow!

Your seminar team

Dear Students,

Due to the illness of the three presenters scheduled for today’s session, we have decided to cancel and postpone the first block sessions by one week. This means that the paper presentations of the first block will take place on November 18, and the panel discussion on November 25, at the usual time slot.

Remember that you can reach out to get feedback on your presentations, but please do so at least one week before your presentation to ensure we can schedule a session and you have time to incorporate the feedback.

We’ve also noticed some confusion regarding submissions, so here’s a clarification:

• For presentations:

  • Only the presenters need to submit their slides in PDF format.

  • The submission on CMS portal for slides will open a few days before the presentation date, with a deadline on the day of the presentation.

  • Students who are not presenting do not need to submit anything prior to the presentations.

• For panel sessions:

  • Only students who did not present in that block are required to submit a question for the discussion. Presenters may submit questions as well, but it is optional, as they will act as panelists.

  • Submission for questions will open the day after the presentations, with a deadline two days before the panel session (Sunday).

  • All questions will be compiled into a single file and uploaded to CMS the day before the panel, so everyone has time to prepare.

  • Remember, the goal of the panel discussions is to engage in a broader conversation about the framework covered during the block. While discussing individual papers is fine, we encourage questions that go further: compare approaches, analyze advantages and limitations, explore practical applications, and consider future research directions. Please do not limit your questions to the specific papers presented.

Also, remember that attendance is mandatory, and you may miss no more than two sessions without a valid justification.

Thank you all for your understanding today. See you all next week!

Best,

Your seminar team

Dear Students,

After reviewing your paper preferences, we have finalized the list of paper assignments. You can find your matriculation number listed before each reference. Presentations will take place during the presentation session corresponding to each block. Remember that your presentation should be 15 minutes long, followed by a Q&A with questions from both the audience and the instructors.

If you notice that you’ve been assigned a paper you marked as “I would not present” or find any other error, please let us know as soon as possible.

Additionally, we’ve noticed that only three of you have registered on LSF so far. Please remember that registration is required by November 4, one week before the first presentation session.

• BLOCK 1:
  • [7058394] Tomczak, J. & Welling, M.. (2018). VAE with a VampPrior. Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 84:1214-1223. Link: https://proceedings.mlr.press/v84/tomczak18a.html.
  • [7070291] NVAE: A Deep Hierarchical Variational Autoencoder. Advances in Neural Information Processing Systems, 33, 19667-19679. Link: https://proceedings.neurips.cc/paper_files/paper/2020/file/e3b21256183cf7c2c7a66be163579d37-Paper.pdf
  • [7075715] Kingma, D. P., & Dhariwal, P. (2018). Glow: Generative Flow with Invertible 1x1 Convolutions. Advances in Neural Information Processing Systems, 31. Link: https://proceedings.neurips.cc/paper_files/paper/2018/file/d139db6a236200b21cc7f752979132d0-Paper.pdf
  • [7076153] Zhai, S., ZHANG, R., Nakkiran, P., Berthelot, D., Gu, J., Zheng, H., ... & Susskind, J. M. Normalizing Flows are Capable Generative Models. In Forty-second International Conference on Machine Learning. Link: https://openreview.net/pdf?id=2uheUFcFsM
• BLOCK 2:
  • [7062300] Rombach, R., Blattmann, A., Lorenz, D., Esser, P., & Ommer, B. (2022). High-Resolution Image Synthesis with Latent Diffusion Models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 10684-10695). Link: https://openaccess.thecvf.com/content/CVPR2022/papers/Rombach_High-Resolution_Image_Synthesis_With_Latent_Diffusion_Models_CVPR_2022_paper.pdf
  • [7076165] Ho, J., & Salimans, T. Classifier-Free Diffusion Guidance. In NeurIPS 2021 Workshop on Deep Generative Models and Downstream Applications. Link: https://openreview.net/pdf?id=qw8AKxfYbI
  • [7026925] Chen, X., Duan, Y., Houthooft, R., Schulman, J., Sutskever, I., & Abbeel, P. (2016). InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets. Advances in Neural Information Processing Systems, 29. Link: https://proceedings.neurips.cc/paper_files/paper/2016/file/7c9d0b1f96aebd7b5eca8c3edaa19ebb-Paper.pdf
  • [7074899] Casanova, A., Careil, M., Verbeek, J., Drozdzal, M., & Romero Soriano, A. (2021). Instance-Conditioned GAN. Advances in Neural Information Processing Systems, 34, 27517-27529. Link: https://proceedings.neurips.cc/paper_files/paper/2021/file/e7ac288b0f2d41445904d071ba37aaff-Paper.pdf
• BLOCK 3:
  • [7086510] Palumbo, E., Daunhawer, I., & Vogt, J. E. (2023). MMVAE+: Enhancing the Generative Quality of Multimodal VAEs without Compromises. In The Eleventh International Conference on Learning Representations. Link: https://openreview.net/pdf?id=sdQGxouELX
  • [7073034] Campbell, A., Yim, J., Barzilay, R., Rainforth, T., & Jaakkola, T. (2024, July). Generative Flows on Discrete State-Spaces: Enabling Multimodal Flows with Applications to Protein Co-Design. In International Conference on Machine Learning (pp. 5453-5512). PMLR. Link: https://proceedings.mlr.press/v235/campbell24a.html
  • [7057757] Kang, M., Zhu, J. Y., Zhang, R., Park, J., Shechtman, E., Paris, S., & Park, T. (2023). Scaling up GANs for Text-to-Image Synthesis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 10124-10134). Link: https://openaccess.thecvf.com/content/CVPR2023/papers/Kang_Scaling_Up_GANs_for_Text-to-Image_Synthesis_CVPR_2023_paper.pdf
  • [7081324] Chun, S., Kim, W., Park, S., & Yun, S. (2025) Probabilistic Language-Image Pre-Training. In The Thirteenth International Conference on Learning Representations. Link: https://openreview.net/pdf?id=D5X6nPGFUY

If you have any questions, don’t hesitate to reach out. 

Best regards,
Your seminar team