dlcp2025:restricted:programme
Differences
This shows you the differences between two versions of the page.
Both sides previous revisionPrevious revisionNext revision | Previous revision | ||
dlcp2025:restricted:programme [04/06/2025 18:06] – [Section 3. Machine Learning in Natural Sciences] admin | dlcp2025:restricted:programme [06/06/2025 18:19] (current) – [39. ML-Based Optimum Sub-system Size Heuristic for the GPU Implementation of the Tridiagonal Partition Method] admin | ||
---|---|---|---|
Line 1: | Line 1: | ||
- | ====== | + | ====== |
// | // | ||
- | **Внимание!** В данный список включены заявленные доклады, | + | **Attention!** This list includes the announced papers included in the conference program. |
- | //Если кто-то не нашел себя в списке, просьба сообщить по почте | + | //If someone did not find themselves in the list, please inform us by email [[dlcp@sinp.msu.ru]]// |
===== Section 1. Machine Learning in Fundamental Physics ===== | ===== Section 1. Machine Learning in Fundamental Physics ===== | ||
Line 30: | Line 30: | ||
===== Section 3. Machine Learning in Natural Sciences ===== | ===== Section 3. Machine Learning in Natural Sciences ===== | ||
- | ==== Восстановление приповерхностной влажности атмосферы над океаном с применением методов машинного обучения ==== | ||
- | |||
- | //// | ||
- | |||
- | Влажность воздуха в приповерхностном слое атмосферы над океаном является ключевым климатическим параметром, | ||
==== 37. Neutron spectrum unfolding with deep learning models for tabular data ==== | ==== 37. Neutron spectrum unfolding with deep learning models for tabular data ==== | ||
Line 44: | Line 39: | ||
The research was carried out within the framework of the state assignment of the Ministry of Science and Higher Education of the Russian Federation (topic No. 124112200072-2). | The research was carried out within the framework of the state assignment of the Ministry of Science and Higher Education of the Russian Federation (topic No. 124112200072-2). | ||
+ | ==== 39. ML-Based Optimum Sub-system Size Heuristic for the GPU Implementation of the Tridiagonal Partition Method ==== | ||
+ | |||
+ | //Milena Veneva (1) \\ (1) RIKEN Center for Computational Science, R-CCS, 7-1-26 Minatojima-minami-machi, | ||
+ | |||
+ | The parallel partition algorithm for solving systems of linear algebraic equations (SLAEs) suggested in [1] is an efficient numerical technique for solving SLAEs with tridiagonal coefficient matrices which consists of three stages. It works by splitting the original matrix into smaller sub-matrices and solving these smaller SLAEs in parallel. Originally designed for use with a large number of processors, this algorithm was implemented using MPI (Message Passing Interface) technology in [1]. The development of HPC applications typically consists of two key phases: writing code that functions correctly and then optimizing that code to enhance performance. The nature of the parallel partition method is such that the initial SLAE with N unknowns is partitioned into a number of sub-systems with m unknowns each. The size of the SLAE N that the user solves is usually determined by the size of the problem they need to solve, while the size of the sub-system within the parallel partition method m is a parameter that needs to be tuned. We present one of the optimizations made to our CUDA [2] implementation, | ||
+ | |||
+ | [1] Austin, T.~M., and Berndt, M., and Moulton, J.~D., A Memory Efficient Parallel Tridiagonal Solver, Preprint LA-VR-03-4149, | ||
+ | [2] NVIDIA, NVIDIA CUDA C++ Programming Guide. https:// | ||
+ | [3] Fix, H., and Joseph, L., Discriminatory Analysis. Nonparametric Discrimination: | ||
+ | |||
+ | ==== 44. Natural Image Classification via Quasi-Cyclic Graph Ensembles and Random-Bond Ising Models with Enhanced Nishimori Temperature Estimation ==== | ||
+ | |||
+ | // | ||
+ | |||
+ | Recent advances have demonstrated the effectiveness of spectral clustering on the beta-Hessian of Graham matrices constructed from quasi-cyclic graphs in the context of Random-Bond Ising Models (RBIMs). Notably, at paper [1] showed that combining LDPC-inspired graph with VGG16-extracted features from GAN-generated two-class images (e.g., dog vs. cat) significantly outperformed Erdős–Rényi baselines in clustering accuracy, improving overlap from 73.21% to 90.60%—and up to 93.23% when using cosine similarity [2]. In this paper, we extend these insights to natural multi-class datasets, specifically ImageNet-10 and ImageNet-100. We introduce a refined approach to estimating the Nishimori temperature and propose a mixture-of-graphs model built from an ensemble of optimized RBIMs. These models leverage diverse quasi-cyclic graph families—including Spherical graphs and Multi-Edge Type LDPC graphs—to create sparse, expressive interaction structures. Feature embeddings are extracted from a lightweight MobileNetV2-based CNN, compressing 1280-dimensional activations to 32–64 feature maps per image. Using ensembles of 3 to 9 graph models, our approach achieves classification accuracies of up to 98.7% on ImageNet-10 and 82.5% on ImageNet-100 under optimal conditions with 32-dimensional embeddings. We demonstrate that significant parameter reduction in the MLP classification head (from 1280 to 32) improves both computational efficiency and robustness to feature puncturing. Furthermore, | ||
+ | |||
+ | [1] Usatyuk, V.S., Sapozhnikov, | ||
+ | [2] Dall' | ||
dlcp2025/restricted/programme.1749060372.txt.gz · Last modified: by admin