401 research outputs found
A new finite-horizon dynamic programming analysis of nonanticipative rate-distortion function for Markov sources
No full textRollout-based approximate dynamic programming for MDPs with information-theoretic constraints
No full textExploiting multi-granularity visual features for retinal layer segmentation in human eyes
No full textAccurate segmentation of retinal layer boundaries can facilitate the detection of patients with early ophthalmic disease. Typical segmentation algorithms operate at low resolutions without fully exploiting multi-granularity visual features. Moreover, several related studies do not release their datasets that are key for the research on deep learning-based solutions. We propose a novel end-to-end retinal layer segmentation network based on ConvNeXt, which can retain more feature map details by using a new depth-efficient attention module and multi-scale structures. In addition, we provide a semantic segmentation dataset containing 206 retinal images of healthy human eyes (named NR206 dataset), which is easy to use as it does not require any additional transcoding processing. We experimentally show that our segmentation approach outperforms state-of-the-art approaches on this new dataset, achieving, on average, a Dice score of 91.3% and mIoU of 84.4%. Moreover, our approach achieves state-of-the-art performance on a glaucoma dataset and a diabetic macular edema (DME) dataset, showing that our model is also suitable for other applications. We will make our source code and the NR206 dataset publicly available at (https://github.com/Medical-Image-Analysis/Retinal-layer-segmentation)
Supplemental Material - Characteristics and Quality of Diagnostic and Risk Prediction Models for Frailty in Older Adults: A Systematic Review
No full textSupplemental material for Characteristics and Quality of Diagnostic and Risk Prediction Models for Frailty in Older Adults: A Systematic Review by Yinyan Gao, MD, Yancong Chen, MD, Mingyue Hu, PhD, Ting Gan, PhD, Xuemei Sun, PhD, Zixuan Zhang, MD, Wenbo He, PhD, and Irene X. Y. Wu in Journal of Applied Gerontology</p
Single-cell transcriptomic profiling of goat milk somatic cells highlights immune heterogeneity and epithelial cell-related networks
Full text linkThe global goat milk market has expanded rapidly, driven by its reputed hypoallergenic properties and associated health benefits. Here we present the first single-cell RNA sequencing (scRNA-seq) atlas of somatic cells in mid-lactation Saanen goat milk, revealing cellular heterogeneity and immune-regulatory mechanisms. Analysis of 7276 high-quality cells from five biological replicates revealed seven populations: myofibroblasts, dendritic cells (DCs), epithelial cells (EPCs), monocytes, bone marrow-derived progenitor cells, neutrophils, and T cells. The proportion of EPCs varied markedly between individuals (23.18 %–94.09 %, p = 0.0295) and was positively correlated with somatic cell count (R = 0.6087, based on five biological replicates), suggesting a moderate association. Pseudotime analysis revealed two differentiation trajectories: high-epithelial samples (HPG; >80 % EPCs) were dominated by immune-cell- dominated, whereas low-epithelial samples (LPG; <60 % EPCs) primarily exhibited epithelial differentiation. Cell-cell communication analyses showed distinct signaling: SELL/CXCL-mediated immune pathways were activated in HPG, while TGF-β/SPP1 signaling—linked to cell migration and immune suppression—was upregulated in LPG. A total of 214 differentially expressed genes (DEGs) were identified. Pro-inflammatory factors, such as SAA and PAEP, were enriched in HPG, whereas anti-inflammatory markers, including SERPIN B3 and C3, were elevated in LPG. Notably, casein genes (CSN1S2, CSN2, CSN3) were markedly upregulated in immune cells of HPG (T cells, monocytes and DCs). In conclusion, this work unveils a key cellular biomarker for milk quality, which is expected to guide the dairy industry towards producing safer and hypoallergenic goat milk products
SANTOS Benchmark for Table Union Search
No full textThis record contains the datasets released with SIGMOD 2023 paper entitled "SANTOS: Relationship-based Semantic Table Union Search". We release two new tabular benchmarks to evaluate the table union search problem over the data lakes. Furthermore, we also release relabeled ground truth for an existing TUS benchmark by taking the binary relationship between the columns into account. Please visit our paper for further details.
If you use our dataset for your work, please cite our paper as:
Aamod Khatiwada, Grace Fan, Roee Shraga, Zixuan Chen, Wolfgang Gatterbauer, Renée J. Miller, and Mirek
Riedewald. 2023. SANTOS: Relationship-based Semantic Table Union Search. SIGMOD Conference 2023, ACM
@inproceedings{2023khatiwadasantos,
title = {SANTOS: Relationship-based Semantic Table Union Search},
author={Khatiwada, Aamod and Fan, Grace and Shraga, Roee and Chen, Zixuan and Gatterbauer, Wolfgang and Miller, Ren{\'e}e J and Riedewald, Mirek},
year = {2023},
publisher = {ACM},
booktitle = {SIGMOD Conference 2023},
}
You can find SANTOS implementation at: https://github.com/northeastern-datalab/santos
You can find the original TUS benchmark at: https://github.com/RJMillerLab/table-union-search-benchmark
Abstract: Existing techniques for unionable table search define unionability using metadata (tables must have the same or similar schemas) or column-based metrics (for example, the values in a table should be drawn from the same domain). In this work, we introduce the use of semantic relationships between pairs of columns in a table to improve the accuracy of union search. Consequently, we introduce a new notion of unionability that considers relationships between columns, together with the semantics of columns, in a principled way. To do so, we present two new methods to discover semantic relationship between pairs of columns: The first uses an existing knowledge base (KB), the second (which we call a “synthesized KB”) uses knowledge from the data lake itself. We adopt an existing Table Union Search benchmark and present new (open) benchmarks that represent small and large real data lakes. We show that our new unionability search algorithm called SANTOS outperforms a state-of-the-art union search that uses a wide variety of column-based semantics, including word embeddings and regular expressions. We show empirically in all benchmarks that our synthesized KB improves the accuracy of union search by representing relationship semantics that may not be contained in an available KB. This result hints at a promising future of creating a synthesized KBs from data lakes with limited KB coverage and using them for union search
Analysis the character of J.S.Bach’s Sinfonia in A major (BWV 798) – based on the dynamic and research on related acoustic parameters
No full textVarious annotations of the character of J.S.Bach’s clavier music in piano perform seem to be under discussion by the pianists and musicologist. However, the extent of dynamic and the way of dynamic changes that had been considered appear scarce. In the essay below, the extent of the dynamic and the way of dynamic changes of J.S.Bach’s Sinfonia in A major (BWV 798) in piano perform is under analysis and discuss
Recommended from our members
EFFORTS TOWARDS EFFECTIVE ROBOTIC STRAWBERRY HARVESTING
Full text linkStrawberry is one of the most important agricultural products in the United States and around the world. Traditionally the harvesting of these delicate fruits has been heavily reliant on a seasonal workforce, which is currently impacted by uncertainties in labor availability and increasing labor costs. The investigation of robotic strawberry harvesting has emerged as a promising alternative to address these labor-related challenges. This dissertation research focused on advancing the robotic strawberry harvesting technology by exploring the challenges posed by occlusion of ripe strawberries by other canopy parts such as leaves, vine, and immature fruit. The study introduced innovative solutions in three different areas. Firstly, the study advanced the machine vision system by leveraging the YOLO (You Only Look Once) family of deep learning models. Specifically, YOLOv5 and YOLO v8 models were modified (e.g., YOLOv5s-Straw(+C2f+SPPFCSP) and YOLOv8s(+C3x+head+αIoU)) to improve their performance in strawberry detection, which achieved the highest mean average precision (mAP) of 80.3% and the highest peak mAP of 83.2%, while maintaining real-time inference speeds on a laptop computer with GPU (RTX3070) and CPU (11800H). Secondly, this research addressed the critical task of determining the pickability of partially occluded strawberries using the YOLOv5s-cls model, which achieved an accuracy of 95.0% with a very low inference time of 2.8 ms. These machine vision models were then integrated with a robotic system designed for open-field strawberry harvesting. The robot included a commercial 6 DOF (Degree of Freedom) manipulator and a novel gripper and fan system, which were installed on a mobile platform. In the challenging open-field environments, this integrated harvester, particularly with the fan system, achieved a significantly higher picking success rate of 73.9% compared to the same achieved without the fan system (58.1%). This research findings demonstrated the potential of these technological innovations in addressing occlusion-related complexities, thereby enhancing efficiency and precision in robotic strawberry harvesting in open-field environments
Bayesian Deep Learning for Distilling Physical Laws from Videos
No full textAn end-to-end framework is developed to discover physical laws directly from videos, which can help facilitate the study on robust prediction, system stability analysis and gain the physical insight of a dynamic process. In this work, a video information extraction module is proposed to detect and collect the pixel position of moving objects, which would be further transformed into physical states we care about. A physical law discovery module is developed to learn closed-form expressions based on the extracted physical information. The video information extraction module takes advantage of contour detection and Hough transformation to extract position information. The physical law discovery module includes a deep neural network-like hierarchical structure Mathematical Operation Network (MathONet) which is consisted of basic mathematical operations. We develop a sparse Bayesian learning algorithm to learn both the topology and parameters of dynamic systems. Several simulated videos were generated to illustrate Newton’s law of motion, the state space of a Duffing oscillator, and the pendulum motion equation. By demonstrating on these examples, our method can discover the corresponding governing function without requiring much prior information.Mechanical Engineering | Vehicle Engineering | Cognitive Robotic
- …
