Thanks to the large pre-trained vision-language models (VLMs) like CLIP, we can craft a
zero-shot classifier by discrete prompt design, eg, the confidence score of an image being"[…

In this paper, we propose a multiscale hierarchical attention approach for supervised video
summarization. Different from most existing supervised methods which employ bidirectional …

In this paper, we propose a dynamic graph modeling approach to learn spatial-temporal
representations for video summarization. Most existing video summarization methods extract …

Recent advancements in large language models (LLMs) have led to the creation of
intelligent agents capable of performing complex tasks. This paper introduces a novel LLM-based …

Introduction Although the interaction between tumor immune microenvironment and
angiogenesis has been well established, evidence supporting the chemo-free combination of …

Sulfurimonas species are commonly isolated from sulfidic habitats and numerous 16S rRNA
sequences related to Sulfurimonas species have been identified in chemically distinct …

Multi-modal large language models have demonstrated impressive performances on most
vision-language tasks. However, the model generally lacks the understanding capabilities for …

One of the hallmark mechanisms activated by type I interferons (IFNs) in human tissues
involves cleavage of intracellular RNA by the kinase homology endoribonuclease RNase L. We …

Background and Aims Oxaliplatin (OXA) is one of the most common chemotherapeutics in
advanced hepatocellular carcinoma (HCC), the resistance of which poses a big challenge. …

Aims Histone methyltransferase G9a has been primarily understood as a co‐repressor of
gene expression, but it has been shown that G9a also positively regulates nuclear receptor‐…