第8回学生ゼミを開催しました The 8th student seminar was held






In this seminar, Qiu-san gave a presentation on the variational autoencoder in Chapter 12, and Sakaban-san gave a presentation on the adversarial generative network and normaliing flow.

The variational autoencoder model was more difficult to derive than the models we have dealt with so far, but I think it became easier to understand by following it carefully as we all discussed the purpose of the model, assumptions, derivation of losses, roles, and so on. We also touched on several generative models such as VAE, GAN, and normalizing flow, and were able to learn about the assumptions and characteristics of each, deepening our understanding of generative models.

This will be the last seminar of the summer. I would like to thank everyone for their hard work over the past three months.

The next seminar will be a laboratory seminar, where two fourth-year students will give presentations for the mid-term review of their graduation theses.


七夕の飾りつけを行いました✨ Tanabata Festival was held.






On July 8, we held a Tanabata decoration event at the Regional Planning and Information Laboratory.

Despite the intense heat, we went shopping to the 100-yen shop and wrote our wishes in strips of paper.

By the way, I prayed for the recovery of my fingernail that I hurt in May. I hope it will be back to normal soon. I hope that the wishes of our readers will also reach the stars☆!

Despite the heat, I will keep working hard on my research so that my wish will come true!



第7回学生ゼミが開かれました The 7th student seminar was held









We learned how to use the tools such as Git, Github, Docker and PyTorch(※), which are necessary for research from Katayama-san.

It is very painstaking work for those with no experience of using such tools to make SSH connections to servers and build virtual environments with Docker, so I truly appreciate that the senior staffs taught me those tasks carefully from the very beginning!

The next seminar will deal with GANs and normalization flows in Chapter 12 of ‘Deep Learning’.

※PyTorch…Open source machine learning library for Python. It has the largest market share in recent years. In our laboratory, we mainly use it to build and train models.