Xiangyu Gao

Dequan Wang, Xiaosong Wang, Lilong Wang et al. · berkeley

Foundation models, often pre-trained with large-scale data, have achieved paramount success in jump-starting various vision and language applications. Recent advances further enable adapting foundation models in downstream tasks efficiently using only a few training samples, e.g., in-context learning. Yet, the application of such learning paradigms in medical image analysis remains scarce due to the shortage of publicly accessible data and benchmarks. In this paper, we aim at approaches adapting the foundation models for medical image classification and present a novel dataset and benchmark for the evaluation, i.e., examining the overall performance of accommodating the large-scale foundation models downstream on a set of diverse real-world clinical tasks. We collect five sets of medical imaging data from multiple institutes targeting a variety of real-world clinical tasks (22,349 images in total), i.e., thoracic diseases screening in X-rays, pathological lesion tissue screening, lesion detection in endoscopy images, neonatal jaundice evaluation, and diabetic retinopathy grading. Results of multiple baseline methods are demonstrated using the proposed dataset from both accuracy and cost-effective perspectives.

Xiangyu Gao, Yaping Sun, Dongyu Wei et al.

With the proliferation of edge computing, efficient AI inference on edge devices has become essential for intelligent applications such as autonomous vehicles and VR/AR. In this context, we address the problem of efficient remote object recognition by optimizing feature transmission between mobile devices and edge servers. We propose an online optimization framework to address the challenge of dynamic channel conditions and device mobility in an end-to-end communication system. Our approach builds upon existing methods by leveraging a semantic knowledge base to drive multi-level feature transmission, accounting for temporal factors and dynamic elements throughout the transmission process. To solve the online optimization problem, we design a novel soft actor-critic-based deep reinforcement learning system with a carefully designed reward function for real-time decision-making, overcoming the optimization difficulty of the NP-hard problem and achieving the minimization of semantic loss while respecting latency constraints. Numerical results showcase the superiority of our approach compared to traditional greedy methods under various system setups.

Jinhe Shi, Xiangyu Gao, Chenyu Ha et al.

Adverse drug events (ADEs) are a serious health problem that can be life-threatening. While a lot of studies have been performed on detect correlation between a drug and an AE, limited studies have been conducted on personalized ADE risk prediction. Among treatment alternatives, avoiding the drug that has high likelihood of causing severe AE can help physicians to provide safer treatment to patients. Existing work on personalized ADE risk prediction uses the information obtained in the current medical visit. However, on the other hand, medical history reveals each patient's unique characteristics and comprehensive medical information. The goal of this study is to assess personalized ADE risks that a target drug may induce on a target patient, based on patient medical history recorded in claims codes, which provide information about diagnosis, drugs taken, related medical supplies besides billing information. We developed a HTNNR model (Hierarchical Time-aware Neural Network for ADE Risk) that capture characteristics of claim codes and their relationship. The empirical evaluation show that the proposed HTNNR model substantially outperforms the comparison methods, especially for rare drugs.