arXiv subjects

Masataka Sato

Publications and source records attributed to Masataka Sato.

4 recordsLinked to original sources

arXiv2026

Contrastive Learning for Multi Label ECG Classification with Jaccard Score Based Sigmoid Loss

Recent advances in large language models (LLMs) have enabled the development of multimodal medical AI. While models such as MedGemini achieve high accuracy on VQA tasks like USMLE MM, their performance on ECG based tasks remains limited, and some models, such as MedGemma, do not support ECG data at all. Interpreting ECGs is inherently challenging, and diagnostic accuracy can vary depending on the interpreter's experience. Although echocardiography provides rich diagnostic information, it requires specialized equipment and personnel, limiting its availability. In this study, we focus on constructing a robust ECG encoder for multimodal pretraining using real world hospital data. We employ SigLIP, a CLIP based model with a sigmoid based loss function enabling multi label prediction, and introduce a modified loss function tailored to the multi label nature of ECG data. Experiments demonstrate that incorporating medical knowledge in the language model and applying the modified loss significantly improve multi label ECG classification. To further enhance performance, we increase the embedding dimensionality and apply random cropping to mitigate data drift. Finally, per label analysis reveals which ECG findings are easier or harder to predict. Our study provides a foundational framework for developing medical models that utilize ECG data.

cs.LG↗

arXiv2025

Application of Contrastive Learning on ECG Data: Evaluating Performance in Japanese and Classification with Around 100 Labels

The electrocardiogram (ECG) is a fundamental tool in cardiovascular diagnostics due to its powerful and non-invasive nature. One of the most critical usages is to determine whether more detailed examinations are necessary, with users ranging across various levels of expertise. Given this diversity in expertise, it is essential to assist users to avoid critical errors. Recent studies in machine learning have addressed this challenge by extracting valuable information from ECG data. Utilizing language models, these studies have implemented multimodal models aimed at classifying ECGs according to labeled terms. However, the number of classes was reduced, and it remains uncertain whether the technique is effective for languages other than English. To move towards practical application, we utilized ECG data from regular patients visiting hospitals in Japan, maintaining a large number of Japanese labels obtained from actual ECG readings. Using a contrastive learning framework, we found that even with 98 labels for classification, our Japanese-based language model achieves accuracy comparable to previous research. This study extends the applicability of multimodal machine learning frameworks to broader clinical studies and non-English languages.

cs.AI↗

arXiv2012

Phase transition without global ordering in a hierarchical scale-free network

We study the site-bond percolation on a hierarchical scale-free network, namely, the decorated (2,2)-flower, by using the renormalization group technique. The phase diagram essentially depends on the fraction of occupied sites. Surprisingly, when each site is unoccupied even with a small probability, the system permits neither the percolating phase nor the nonpercolating phase, but rather only critical phases. Although the order parameter always remains zero, a transition still exists between the critical phases that is characterized by the value of the fractal exponent, which measures the degree of criticality; the system changes from one critical state to another with the jump of the fractal exponent at the transition point. The phase boundary depends on the fraction of occupied sites. When the fraction of unoccupied sites exceeds a certain value, the transition line between the critical phases disappears, and a unique critical phase remains.

cond-mat.dis-nn↗

arXiv2010

Generating-function approach for bond percolations in hierarchical networks

We study bond percolations on hierarchical scale-free networks with the open bond probability of the shortcuts $\tilde{p}$ and that of the ordinary bonds $p$. The system has a critical phase in which the percolating probability $P$ takes an intermediate value $0<P<1$. Using generating function approach, we calculate the fractal exponent $ψ$ of the root clusters to show that $ψ$ varies continuously with $\tilde{p}$ in the critical phase. We confirm numerically that the distribution $n_s$ of cluster size $s$ in the critical phase obeys a power law $n_s \propto s^{-τ}$, where $τ$ satisfies the scaling relation $τ=1+ψ^{-1}$. In addition the critical exponent $β(\tilde{p})$ of the order parameter varies as $\tilde{p}$, from $β\simeq 0.164694$ at $\tilde{p}=0$ to infinity at $\tilde{p}=\tilde{p}_c=5/32$.

cond-mat.dis-nn↗