learning architectures
Recently Published Documents


TOTAL DOCUMENTS

485
(FIVE YEARS 367)

H-INDEX

18
(FIVE YEARS 10)

2022 ◽  
Author(s):  
Stefan Bachhofner ◽  
Peb Ruswono Aryan ◽  
Bernhard Krabina ◽  
Robert David

This paper presents an on-going research where we studythe problem of embedding meta-data enriched graphs, with a focus onknowledge graphs in a vector space with transformer based deep neuralnetworks. Experimentally, we compare ceteris paribus the performance ofa transformer-based model with other non-transformer approaches. Dueto their recent success in natural language processing we hypothesizethat the former is superior in performance. We test this hypothesizesby comparing the performance of transformer embeddings with non-transformer embeddings on different downstream tasks. Our researchmight contribute to a better understanding of how random walks in-fluence the learning of features, which might be useful in the design ofdeep learning architectures for graphs when the input is generated withrandom walks.


Technologies ◽  
2021 ◽  
Vol 10 (1) ◽  
pp. 2
Author(s):  
Lazaros Alexios Iliadis ◽  
Spyridon Nikolaidis ◽  
Panagiotis Sarigiannidis ◽  
Shaohua Wan ◽  
Sotirios K. Goudos

Through the extensive study of transformers, attention mechanisms have emerged as potentially more powerful than sequential recurrent processing and convolution. In this realm, Vision Transformers have gained much research interest, since their architecture changes the dominant paradigm in Computer Vision. An interesting and difficult task in this field is the classification of artwork styles, since the artistic style of a painting is a descriptor that captures rich information about the painting. In this paper, two different Deep Learning architectures—Vision Transformer and MLP Mixer (Multi-layer Perceptron Mixer)—are trained from scratch in the task of artwork style recognition, achieving over 39% prediction accuracy for 21 style classes on the WikiArt paintings dataset. In addition, a comparative study between the most common optimizers was conducted obtaining useful information for future studies.


2021 ◽  
Author(s):  
Farhad Zamani ◽  
Retno Wulansari

Recently, emotion recognition began to be implemented in the industry and human resource field. In the time we can perceive the emotional state of the employee, the employer could gain benefits from it as they could improve the quality of decision makings regarding their employee. Hence, this subject would become an embryo for emotion recognition tasks in the human resource field. In a fact, emotion recognition has become an important topic of research, especially one based on physiological signals, such as EEG. One of the reasons is due to the availability of EEG datasets that can be widely used by researchers. Moreover, the development of many machine learning methods has been significantly contributed to this research topic over time. Here, we investigated the classification method for emotion and propose two models to address this task, which are a hybrid of two deep learning architectures: One-Dimensional Convolutional Neural Network (CNN-1D) and Recurrent Neural Network (RNN). We implement Gated Recurrent Unit (GRU) and Long Short-Term Memory (LSTM) in the RNN architecture, that specifically designed to address the vanishing gradient problem which usually becomes an issue in the time-series dataset. We use this model to classify four emotional regions from the valence-arousal plane: High Valence High Arousal (HVHA), High Valence Low Arousal (HVLA), Low Valence High Arousal (LVHA), and Low Valence Low Arousal (LVLA). This experiment was implemented on the well-known DEAP dataset. Experimental results show that proposed methods achieve a training accuracy of 96.3% and 97.8% in the 1DCNN-GRU model and 1DCNN-LSTM model, respectively. Therefore, both models are quite robust to perform this emotion classification task.


2021 ◽  
Author(s):  
Sajeel Aziz

The contributions of this paper are two-fold. We define unsupervised techniques for the panoptic segmentation of an image. We also define clusters which encapsulate the set of features that define objects of interest inside a scene. The motivation is to provide an approach that mimics natural formation of ideas inside the brain. Fundamentally, the eyes and visual cortex constitute the visual system, which is essential for humans to detect and recognize objects. This can be done even without specific knowledge of the objects. We strongly believe that a supervisory signal should not be required to identify objects in an image. We present an algorithm that replaces the eye and visual cortex with deep learning architectures and unsupervised clustering methods. The proposed methodology may also be used as a one-click panoptic segmentation approach which promises to significantly increase annotation efficiency. We have made the code available privately for review<sup>1</sup>.<div><br></div>


2021 ◽  
Author(s):  
Sajeel Aziz

The contributions of this paper are two-fold. We define unsupervised techniques for the panoptic segmentation of an image. We also define clusters which encapsulate the set of features that define objects of interest inside a scene. The motivation is to provide an approach that mimics natural formation of ideas inside the brain. Fundamentally, the eyes and visual cortex constitute the visual system, which is essential for humans to detect and recognize objects. This can be done even without specific knowledge of the objects. We strongly believe that a supervisory signal should not be required to identify objects in an image. We present an algorithm that replaces the eye and visual cortex with deep learning architectures and unsupervised clustering methods. The proposed methodology may also be used as a one-click panoptic segmentation approach which promises to significantly increase annotation efficiency. We have made the code available privately for review<sup>1</sup>.<div><br></div>


Water ◽  
2021 ◽  
Vol 13 (24) ◽  
pp. 3633
Author(s):  
Reed M. Maxwell ◽  
Laura E. Condon ◽  
Peter Melchior

While machine learning approaches are rapidly being applied to hydrologic problems, physics-informed approaches are still relatively rare. Many successful deep-learning applications have focused on point estimates of streamflow trained on stream gauge observations over time. While these approaches show promise for some applications, there is a need for distributed approaches that can produce accurate two-dimensional results of model states, such as ponded water depth. Here, we demonstrate a 2D emulator of the Tilted V catchment benchmark problem with solutions provided by the integrated hydrology model ParFlow. This emulator model can use 2D Convolution Neural Network (CNN), 3D CNN, and U-Net machine learning architectures and produces time-dependent spatial maps of ponded water depth from which hydrographs and other hydrologic quantities of interest may be derived. A comparison of different deep learning architectures and hyperparameters is presented with particular focus on approaches such as 3D CNN (that have a time-dependent learning component) and 2D CNN and U-Net approaches (that use only the current model state to predict the next state in time). In addition to testing model performance, we also use a simplified simulation based inference approach to evaluate the ability to calibrate the emulator to randomly selected simulations and the match between ML calibrated input parameters and underlying physics-based simulation.


2021 ◽  
Author(s):  
Sandaru Seneviratne ◽  
Artem Lenskiy ◽  
Christopher Nolan ◽  
Eleni Daskalaki ◽  
Hanna Suominen

Complexity and domain-specificity make medical text hard to understand for patients and their next of kin. To simplify such text, this paper explored how word and character level information can be leveraged to identify medical terms when training data is limited. We created a dataset of medical and general terms using the Human Disease Ontology from BioPortal and Wikipedia pages. Our results from 10-fold cross validation indicated that convolutional neural networks (CNNs) and transformers perform competitively. The best F score of 93.9% was achieved by a CNN trained on both word and character level embeddings. Statistical significance tests demonstrated that general word embeddings provide rich word representations for medical term identification. Consequently, focusing on words is favorable for medical term identification if using deep learning architectures.


Sensors ◽  
2021 ◽  
Vol 21 (24) ◽  
pp. 8218
Author(s):  
Issiaka Diaby ◽  
Mickaël Germain ◽  
Kalifa Goïta

The role of a service that is dedicated to road weather analysis is to issue forecasts and warnings to users regarding roadway conditions, thereby making it possible to anticipate dangerous traffic conditions, especially during the winter period. It is important to define pavement conditions at all times. In this paper, a new data acquisition approach is proposed that is based upon the analysis and combination of two sensors in real time by nanocomputer. The first sensor is a camera that records images and videos of the road network. The second sensor is a microphone that records the tire–pavement interaction, to characterize each surface’s condition. The two low-cost sensors were fed to different deep learning architectures that are specialized in surface state analysis; the results were combined using an evidential theory-based data fusion approach. This study is a proof of concept, to test an evidential approach for improving classification with deep learning, applied to only two sensors; however, one could very well add more sensors and make the nanocomputers communicate together, to analyze a larger urban environment.


2021 ◽  
Author(s):  
Daniel Franco-Barranco ◽  
Arrate Muñoz-Barrutia ◽  
Ignacio Arganda-Carreras

AbstractElectron microscopy (EM) allows the identification of intracellular organelles such as mitochondria, providing insights for clinical and scientific studies. In recent years, a number of novel deep learning architectures have been published reporting superior performance, or even human-level accuracy, compared to previous approaches on public mitochondria segmentation datasets. Unfortunately, many of these publications make neither the code nor the full training details public, leading to reproducibility issues and dubious model comparisons. Thus, following a recent code of best practices in the field, we present an extensive study of the state-of-the-art architectures and compare them to different variations of U-Net-like models for this task. To unveil the impact of architectural novelties, a common set of pre- and post-processing operations has been implemented and tested with each approach. Moreover, an exhaustive sweep of hyperparameters has been performed, running each configuration multiple times to measure their stability. Using this methodology, we found very stable architectures and training configurations that consistently obtain state-of-the-art results in the well-known EPFL Hippocampus mitochondria segmentation dataset and outperform all previous works on two other available datasets: Lucchi++ and Kasthuri++. The code and its documentation are publicly available at https://github.com/danifranco/EM_Image_Segmentation.


Sign in / Sign up

Export Citation Format

Share Document