The Machine Learning and Data Science (MLDS) Unit focuses on developing fundamental machine learning algorithms and using machine learning to solve important scientific problems. Currently, they are interested in statistical modeling of high-dimensional data, including kernels and deep learning models, as well as geometric machine learning algorithms such as Graph Neural Networks (GNN) and Optimal Transport Problems. In addition to developing ML models, they are also dedicated to creating new machine learning methods to automatically discover new scientific findings from data.
-
Curious about the concept of geometric machine learning algorithms
- Graph Neural Networks are one of them
-
New machine learning methods for automatically discovering scientific findings
- At what layer are they considered “new”?
-
https://oist.mlds.jp/2024/09/26/three-papers-accepted-by-neurips-2024/
-
From the machine learning and data science unit, we have three NeurIPS papers accepted! Congratulations!
-
Polyak Meets Parameter-free Clipped Gradient Descent. In NeurIPS, 2024.
-
Scanning Trojaned Models Using Out-of-Distribution Samples. In NeurIPS, 2024.
-
Learning Structured Representations with Hyperbolic Embeddings. In NeurIPS, 2024.
- Strong (blu3mo)(blu3mo)
-
-
It seems that Mr. Yamada also runs a company.
- Flat Minima.