Grapevine red blotch virus (GRBV) is a single-stranded DNA virus that causes grapevine red blotch disease (GRBD). Symptoms of GRBD include red discoloring of leaf margins and red veins on the underside of the leaves. GRBD affects the plant's metabolism and photosynthetic pathways, and disrupts carbon translocation and chlorophyll concentrations....
The fermentation of honey-based musts has been a feature of civilization for millennia. Typically comprised of 75-90% (w/w) simple (i.e. fermentable) sugars, the glucose and fructose present are readily fermented by either pitched or adventitious micro-organisms. Nevertheless, despite its venerable heritage, honey-based fermentations are prone to inconsistency, and it is...
Labeling videos is costly, time-consuming and tedious. These costs can escalate in applications such as medical diagnosis or autonomous driving where we need domain expertise for annotation. Few-shot action recognition aims to solve this problem by annotation-efficient learning mechanisms.
This thesis presents MetaUVFS as the first Unsupervised Meta-learning algorithm for...
Pesticides benefit agriculture by increasing crop yield, quality, and security. However, pesticides may inadvertently harm bees, which are valuable as pollinators. Thus, candidate pesticides in development pipelines must be assessed for toxicity to bees.
Leveraging a data set of 382 molecules with toxicity labels from honey bee exposure experiments, we...
Simultaneous speech translation (SimulST) is widely useful in many cross-lingual communication scenarios, including multinational conferences and international traveling. Since text-based simultaneous machine translation (SimulMT) has achieved great success in recent years. The conventional cascaded approach for SimulST uses a pipeline of streaming ASR followed by simultaneous MT but suffers from...
The utility of high-throughput, computational screening has become an invaluable asset to the field of materials science. In the hierarchy of computational methods, the most accurate methods are often the most computationally expensive. However, as both the efficiency and fidelity of numerical techniques advance, high-quality screening of large materials datasets...
Learning to recognize objects is a fundamental and essential step in human perception and understanding of the world. Accordingly, research of object discovery across diverse modalities plays a pivotal role in the context of computer vision. This field not only contributes significantly to enhancing our understanding of visual information but...
The performance of deep learning frameworks could be significantly improved through considering the particular underlying structures for each dataset. In this thesis, I summarize our three work about boosting the performance of deep learning models through leveraging structures of the data. In the first work, we theoretically justify that, for...
The Internet of Things (IoT) paradigm brought an ever-increasing dependence on low-power devices to collect sensor data and transmit that information to the cloud, placing greater demand on connectivity and lifespan. In response, rapid worldwide innovation demonstrates the trade-offs in processing, communication, and energy consumption with diverse approaches to low-power...
This paper addresses the high model complexity and overconfident frame labeling of state-of-the-art (SOTA) action segmenters. Their complexity is typically justified by the need to sequentially refine action segmentation through multiple stages of a deep architecture. However, this multistage refinement does not take into account uncertainty of frame labeling predicted...