Zu-hua Gao

Gexin Huang, Yanting Yang, Myeongkyun Kang et al.

Vision-language models (VLMs) excel on visual question answering and multimodal reasoning benchmarks. Yet their capability on ultra-resolution images - where critical evidence is tiny, subtle, spatially distant, or distributed - remains unclear. Existing evaluations largely report final-answer accuracy, offering limited insight into whether models acquire and integrate the necessary visual evidence. We introduce UltraVR, a diagnostic benchmark for evidence-grounded visual reasoning over ultra-resolution images. UltraVR spans four high-value scenarios: CCTV surveillance, remote sensing (RS), whole-slide image (WSI) pathology, and industrial anomaly detection (AD). These domains pose complementary challenges: fine-grained object grounding in crowded CCTV scenes, long-range spatial comparison in RS, multi-scale evidence navigation in WSI, and subtle irregularity detection in repetitive industrial layouts. Beyond standard QA triples, each instance includes a structured ground-truth chain of thought with step-level questions, intermediate answers, and reasoning labels. These labels decompose reasoning into evidence grounding, local perception, quantification, evidence integration, and decision inference, enabling process-level diagnosis over black-box scoring. Using UltraVR, we evaluate frontier VLMs and show that current models remain far from reliable on ultra-resolution reasoning. Importantly, the structured annotations allow us to localize failures across the visual-to-decision pipeline: errors concentrate in evidence grounding and local perception, while downstream inference often recovers when intermediate visual facts are supplied. These findings demonstrate UltraVR as a diagnostic testbed for measuring not only whether VLMs answer correctly, but where their ultra-resolution reasoning process breaks.

Beidi Zhao, SangMook Kim, Hao Chen et al.

Multiple Instance Learning (MIL) has advanced WSI analysis but struggles with the complexity and heterogeneity of WSIs. Existing MIL methods face challenges in aggregating diverse patch information into robust WSI representations. While ViTs and clustering-based approaches show promise, they are computationally intensive and fail to capture task-specific and slide-specific variability. To address these limitations, we propose PTCMIL, a novel Prompt Token Clustering-based ViT for MIL aggregation. By introducing learnable prompt tokens into the ViT backbone, PTCMIL unifies clustering and prediction tasks in an end-to-end manner. It dynamically aligns clustering with downstream tasks, using projection-based clustering tailored to each WSI, reducing complexity while preserving patch heterogeneity. Through token merging and prototype-based pooling, PTCMIL efficiently captures task-relevant patterns. Extensive experiments on eight datasets demonstrate its superior performance in classification and survival analysis tasks, outperforming state-of-the-art methods. Systematic ablation studies confirm its robustness and strong interpretability. The code is released at https://github.com/ubc-tea/PTCMIL.

Abicumaran Uthamacumaran, Mohamed Abdouh, Kinshuk Sengupta et al.

Patient-derived extracellular vesicles (EVs) that contains a complex biological cargo is a valuable source of liquid biopsy diagnostics to aid in early detection, cancer screening, and precision nanotherapeutics. In this study, we predicted that coupling cancer patient blood-derived EVs to time-resolved spectroscopy and artificial intelligence (AI) could provide a robust cancer screening and follow-up tools. Methods: Fluorescence correlation spectroscopy (FCS) measurements were performed on 24 blood samples-derived EVs. Blood samples were obtained from 15 cancer patients (presenting 5 different types of cancers), and 9 healthy controls (including patients with benign lesions). The obtained FCS autocorrelation spectra were processed into power spectra using the Fast-Fourier Transform algorithm and subjected to various machine learning algorithms to distinguish cancer spectra from healthy control spectra. Results and Applications: The performance of AdaBoost Random Forest (RF) classifier, support vector machine, and multilayer perceptron, were tested on selected frequencies in the N=118 power spectra. The RF classifier exhibited a 90% classification accuracy and high sensitivity and specificity in distinguishing the FCS power spectra of cancer patients from those of healthy controls. Further, an image convolutional neural network (CNN), ResNet network, and a quantum CNN were assessed on the power spectral images as additional validation tools. All image-based CNNs exhibited a nearly equal classification performance with an accuracy of roughly 82% and reasonably high sensitivity and specificity scores. Our pilot study demonstrates that AI-algorithms coupled to time-resolved FCS power spectra can accurately and differentially classify the complex patient-derived EVs from different cancer samples of distinct tissue subtypes.

Abicumaran Uthamacumaran, Samir Elouatik, Mohamed Abdouh et al.

The early detection of cancer is a challenging problem in medicine. The blood sera of cancer patients are enriched with heterogeneous secretory lipid bound extracellular vesicles (EVs), which present a complex repertoire of information and biomarkers, representing their cell of origin, that are being currently studied in the field of liquid biopsy and cancer screening. Vibrational spectroscopies provide non-invasive approaches for the assessment of structural and biophysical properties in complex biological samples. In this pilot study, multiple Raman spectroscopy measurements were performed on the EVs extracted from the blood sera of 9 patients consisting of four different cancer subtypes (colorectal cancer, hepatocellular carcinoma, breast cancer and pancreatic cancer) and five healthy patients (controls). FTIR (Fourier Transform Infrared) spectroscopy measurements were performed as a complementary approach to Raman analysis, on two of the four cancer subtypes. The AdaBoost Random Forest Classifier, Decision Trees, and Support Vector Machines (SVM) distinguished the baseline corrected Raman spectra of cancer EVs from those of healthy controls (18 spectra) with a classification accuracy of above 90 percent when reduced to a spectral frequency range of 1800 to 1940 inverse cm and subjected to a 50:50 training: testing split. FTIR classification accuracy on 14 spectra showed an 80 percent classification accuracy. Our findings demonstrate that basic machine learning algorithms are powerful applied intelligence tools to distinguish the complex vibrational spectra of cancer patient EVs from those of healthy patients. These experimental methods hold promise as valid and efficient liquid biopsy for artificial intelligence-assisted early cancer screening.