5 papers
Taming Visual Neglect: A Variational Information Bottleneck Framework for Adaptive Attention in Multimodal In-Context Learning
Kaito Tanaka, Yuji Nishimura, Keisuke Matsuda +1
Large vision-language models exhibit strong in-context learning (ICL) capabilities, yet when and why visual context helps multimodal ICL remains poorly understood. Empirical studie…
AE-LLM: Adaptive Efficiency Optimization for Large Language Models
Kaito Tanaka, Masato Ito, Yuji Nishimura +2
Large Language Models (LLMs) have achieved remarkable success across diverse applications, yet their deployment remains challenging due to substantial computational costs, memory r…
Contextual Discrepancy-Aware Contrastive Learning for Robust Medical Time Series Diagnosis in Small-Sample Scenarios
Kaito Tanaka, Aya Nakayama, Masato Ito +2
Medical time series data, such as EEG and ECG, are vital for diagnosing neurological and cardiovascular diseases. However, their precise interpretation faces significant challenges…
Semantic-Preserving Cross-Style Visual Reasoning for Robust Multi-Modal Understanding in Large Vision-Language Models
Aya Nakayama, Brian Wong, Yuji Nishimura +1
The "style trap" poses a significant challenge for Large Vision-Language Models (LVLMs), hindering robust semantic understanding across diverse visual styles, especially in in-cont…
XDR-LVLM: An Explainable Vision-Language Large Model for Diabetic Retinopathy Diagnosis
Masato Ito, Kaito Tanaka, Keisuke Matsuda +1
Diabetic Retinopathy (DR) is a major cause of global blindness, necessitating early and accurate diagnosis. While deep learning models have shown promise in DR detection, their bla…