Papers by Jesús García Ramírez
IEEE Access
Deep Learning has been used for several applications including the analysis of medical images. So... more Deep Learning has been used for several applications including the analysis of medical images. Some transfer learning works show that an improvement in performance is obtained if a pre-trained model on ImageNet is transferred to a new task. Taking into account this, we propose a method that uses a pre-trained model on ImageNet to fine-tune it for Covid-19 detection. After the fine-tuning process, the units that produce a variance equal to zero are removed from the model. Finally, we test the features of the penultimate layer in different classifiers removing those that are less important according to the f-test. The results produce models with fewer units than the transferred model. Also, we study the attention of the neural network for classification. Noise and metadata printed in medical images can bias the performance of the neural network and it obtains poor performance when the model is tested on new data. We study the bias of medical images when raw and masked images are used for training deep models using a transfer learning strategy. Additionally, we test the performance on novel data in both models: raw and masked data. INDEX TERMS Covid-19 detection, model compression, transfer learning, deep learning explainability.
Lecture Notes in Computer Science, 2021
Deep reinforcement learning has obtained impressive performance in challenging tasks in recent ye... more Deep reinforcement learning has obtained impressive performance in challenging tasks in recent years. Nevertheless, it has important limitations such as long training times and the number instances that are needed to achieve acceptable performance. Transfer learning offers an alternative to alleviate these limitations. In this paper, we propose a novel method for transferring knowledge from more than one source tasks. First, we select the best source tasks using a regressor that predicts the performance of a pre-trained model in the target task. Then, we apply a selection of relevant convolutional kernels for the target task in order to find a target model with similar number of parameters compared to the source ones. According to the results, our approach outperforms the accumulated reward obtained when learning from scratch in 20.62% using lower parameters (about 56% of the total, depending on the specific game).
Nowadays, Facial Expression Recognition is an active research subarea of computer vision because ... more Nowadays, Facial Expression Recognition is an active research subarea of computer vision because of its applications in different human activities. In this paper a method for detecting edge information in human face images is introduced, where Robinson Edge technique is applied. In addition, a thresholding threshold values are proposed to reduce the noise in the images with the aim to improve results. Based on experiments of the first stage, it is possible to detect facial expression in eyes, eyebrows, mouth, and forehead with good accuracy.
In computer vision, facial images have several applications such as Facial Expression Recognition... more In computer vision, facial images have several applications such as Facial Expression Recognition and Face Recognition. The segmentation of Regions Of Interest (ROIs) in face images are relevant, because those provide information about facial expressions. In this paper a method to segment mouth and eyebrows in face images based on edge detection and pixel density is proposed. According to the experimental results, our approach extracts the ROIs in face images taken from different public datasets.
The reduction of energy consumption in buildings is one of the goals to improve energy efficiency... more The reduction of energy consumption in buildings is one of the goals to improve energy efficiency. One way to achieve energy savings in buildings is to develop intelligent control strategies for heating systems that are able to reduce power consumption without affecting the thermal comfort. An intelligent control system must be able to predict the temperature of the building in order to manage the heating system. In this paper, we present a rule-based model that is able to predict the indoor temperature for different values of k (hours ahead in time). The model has been learned with FRULER, a genetic fuzzy system that generates accurate and simple knowledge bases. Our approach has been validated with real data from a residential college.
Tecnología Educativa Revista CONAIC, 2021
Moodle es un sistema de administración del aprendizaje que provee servicios a los profesores, est... more Moodle es un sistema de administración del aprendizaje que provee servicios a los profesores, estudiantes y sector administrativo escolar. Moodle tiene una larga lista de desarrolladores alrededor del mundo y ofrece complementos para detectar el plagio. Este artículo propone un complemento para capturar, verificar y actualizar las referencias de un curso cuando el profesor lo crea, reconociendo así que el conocimiento científico se desarrolla con el aporte de mujeres y hombres honestos que hacen referencia por medio de citas a resultados publicados por otros autores. Moodle is a Learning Management System that provides several services to educators, learners and school administrators. Moodle has a large list of developers around the world and it provides also plugins to detect plagiarism. This paper proposes a plugin in order to capture, verify and update the references of a course when it is created by the educator; recognizing that the scientific knowledge is developed by the cont...
Neural Computing and Applications, 2021
Deep reinforcement learning (DRL) combines the benefits of deep learning and reinforcement learni... more Deep reinforcement learning (DRL) combines the benefits of deep learning and reinforcement learning. However, it still requires long training times and a large number of instances to reach an acceptable performance. Transfer learning (TL) offers an alternative to reduce the training time of DRL agents, using less instances and in some cases improving performance. In this work, we propose a transfer learning formulation for DRL across tasks. A relevant problem of TL that we address herein is how to select a proper pre-trained model that will be useful for the target task. We consider the entropy of feature maps in the hidden layers of the convolutional neural network and their actions spaces as relevant features to select a pre-trained model that is then fine-tuned for the target task. We report experimental results of the proposed source task selection methodology when using Deep Q-Networks for learning to play Atari games. Nevertheless, the proposed method could be used in other DRL algorithms (e.g., DDQN, C51, etc.) and also other domains. Results reveal that most of the time our proposed method is capable of selecting source tasks that improve the performance of a model trained from scratch. Additionally, we introduce a method for selecting the most relevant kernels for the target task, the results show that transferring a subset of the convolutional kernels results in similar performance to training the model from scratch while using less parameters.
Research in Computing Science, 2017
In this work, a methodology for classifying emotions (such as happiness, anger and surprise) base... more In this work, a methodology for classifying emotions (such as happiness, anger and surprise) based on face images is proposed. This methodology consist of three stages: in the pre-processing stage, edge detectors and threshold algorithms are used in order to find edge information about ROIs; in the second stage (feature extraction) numeric information of pre-processing images is extracted via interpolation methods; finally, in the classification stage supervised learners such as Neural Networks, Support Vector Machines and Naïve Bayes are tested. According to the results, our approach has acceptable accuracy in order to recognize emotions.
Research in Computing Science, 2016
es un problema que requiere procesamiento masivo, por esta razón el enfoque paralelo es una buena... more es un problema que requiere procesamiento masivo, por esta razón el enfoque paralelo es una buena opción para reducir los tiempos de ejecución. En este artículo se propone un algoritmo paralelo y se muestran los resultados de su implementación mediante la interfaz de paso de mensajes, para encontrar si es que una sub-cadena esta contenida dentro otra cadena de ADN o ARNi, así mismo se propone una forma de guardar los datos para que estos sean enviados dentro de la red de computadoras.
Journal of Intelligent & Fuzzy Systems, 2018
Facial Expression Recognition (FER) is a research area that has been interesting for computer sci... more Facial Expression Recognition (FER) is a research area that has been interesting for computer science community in recent years. In this paper, we propose a methodology for the three stages of a FER system. In the pre-processing stage a method based on edge detectors and thresholding operators for eyebrow and mouth segmentation is proposed; the next stage is feature extraction, we propose using polynomials as features for describing eyebrows and mouth regions. Finally, in classification stage different supervised learners such as: Neural Networks, K-Nearest Neighbors and C4.5 decision trees are tested in order to obtain a model for classifying three out of six basic emotions (anger, happiness and surprise). According to our results, the proposed approach has acceptable accuracy for predicting new examples.
Nature Chemical Biology, 2009
hormone receptors are diverse and different from animal hormone receptors. Second, regulated prot... more hormone receptors are diverse and different from animal hormone receptors. Second, regulated protein degradation plays a central role in hormone signaling. Several receptors are themselves enzymes in the ubiquitin-protein conjugation pathway. In addition, the levels of key downstream signaling proteins are regulated by ubiquitin-dependent degradation. Third, hormone signaling generally leads to major changes in transcription. Although there are likely to be nongenomic hormonal responses, in general these are not as well characterized as the transcriptional responses. Fourth, the hormones interact at multiple levels to regulate various growth and defense processes. In this article we review some of the exciting recent advances in hormone biology.
Journal of Integrative Plant Biology, 2008
Plant mature cells have the capability to reverse their state of differentiation and produce new ... more Plant mature cells have the capability to reverse their state of differentiation and produce new organs under cultured conditions. Two phases, dedifferentiation and redifferentiation, are commonly characterized during in vitro organogenesis. In these processes, cells undergo fate switch for several times regulated by both extrinic and intrinic factors, which is associated with reentry of cell cycle, the balance between euchromatin and heterochromatin, reprogramming of gene expression, etc. This short article reviews the advances in the mechanism on organ regeneration from plant somatic cells in molecular, genomic, and epigenetic aspects, aiming to provide important information on the mechanism underlying cell fate switch during in vitro plant organogenesis.
Annual Review of Plant Biology, 2012
Multicellular organisms possess pluripotent stem cells to form new organs, replenish the daily lo... more Multicellular organisms possess pluripotent stem cells to form new organs, replenish the daily loss of cells, or regenerate organs after injury. Stem cells are maintained in specific environments, the stem cell niches, that provide signals to block differentiation. In plants, stem cell niches are situated in the shoot, root, and vascular meristems—self-perpetuating units of organ formation. Plants' lifelong activity—which, as in the case of trees, can extend over more than a thousand years—requires that a robust regulatory network keep the balance between pluripotent stem cells and differentiating descendants. In this review, we focus on current models in plant stem cell research elaborated during the past two decades, mainly in the model plant Arabidopsis thaliana. We address the roles of mobile signals on transcriptional modules involved in balancing cell fates. In addition, we discuss shared features of and differences between the distinct stem cell niches of Arabidopsis.
Se propone la representación mediante el modelo de<br> las Composiciones Paralelas de Alto ... more Se propone la representación mediante el modelo de<br> las Composiciones Paralelas de Alto Nivel o CPANs del patrón<br> de comunicación/interacción denominado Farm y su adecuación<br> a un nuevo modelo de CPAN llamado Cpan ARNi para<br> demostrar su importancia de uso en la solución de problemas<br> reales en áreas como la bioinformática y realizar la búsqueda<br> exhaustiva paralela (BEP) de cadenas ARN que ayuden a la<br> secuencia y ensamble de ADNs en la construcción de gnomas;<br> todo esto a través de un enfoque de Programación Paralela<br> Estructurada basado en el concepto de Objetos Paralelos. Se<br> muestra el modelo del Farm en sus dos implementaciones, la<br> genérica como CpanFarm y la particular como CpanARNi bajo<br> el paradigma de la Orientación a Objetos para la solución del<br> problema de la búsqueda de cadenas ARNi en particular. Los<br> CPANs que aquí se muestran, contienen un c...
(España), por primera vez. Después de reuniones varias cargadas de ideas y muchas ilusiones, un g... more (España), por primera vez. Después de reuniones varias cargadas de ideas y muchas ilusiones, un grupo de profesores, mayoritariamente del Departamento de Ciencias Humanas y Sociales-pero con la participación de otros pertenecientes a diferentes departamentos también implicados en los estudios de Magisterio-, se plantea editar una revista de educación con el objetivo de investigar sobre los procesos de enseñanzaaprendizaje en su integralidad. Por ello, desde el primer momento los comités de redacción y científico han estado compuestos por profesionales de la educación de todas las áreas; aunque más específicamente, por aquellos que impartían docencia e investigaban en los campos más cercanos a los estudios de Magisterio. En aquel tiempo se impartían cuatro especialidades: 1. Educación Infantil. 2. Educación Musical. 3. Educación Física. Y 4. Lengua Extranjera. Debido a estos planes de
Advances in Computational Intelligence, 2021
Research in Computing Science, 2016
Facial expression recognition and its applications are currently research topics in computer scie... more Facial expression recognition and its applications are currently research topics in computer science. When analyzing face images, it is necessary to apply an algorithm for pre-processing them in order to extract features from each facial expression. In this paper related to our initial research phase, we propose variations and results from algorithms for binarizing and thresholding face expression images. In addition, we present earlier results about the parallel implementation of our proposed algorithms.
Explican el rol que cumple el reajuste por valuación en la presentación de los estados financiero... more Explican el rol que cumple el reajuste por valuación en la presentación de los estados financieros del BCRP
Annual review of plant biology, 2014
Plant stature and development are governed by cell proliferation and directed cell growth. These ... more Plant stature and development are governed by cell proliferation and directed cell growth. These parameters are determined largely by cell wall characteristics. Cellulose microfibrils, composed of hydrogen-bonded β-1,4 glucans, are key components for anisotropic growth in plants. Cellulose is synthesized by plasma membrane-localized cellulose synthase complexes. In higher plants, these complexes are assembled into hexameric rosettes in intracellular compartments and secreted to the plasma membrane. Here, the complexes typically track along cortical microtubules, which may guide cellulose synthesis, until the complexes are inactivated and/or internalized. Determining the regulatory aspects that control the behavior of cellulose synthase complexes is vital to understanding directed cell and plant growth and to tailoring cell wall content for industrial products, including paper, textiles, and fuel. In this review, we summarize and discuss cellulose synthesis and regulatory aspects of ...
Nature Reviews Molecular Cell Biology, 2007
The cellular microenvironment that provides the signals and physical support to maintain stem cel... more The cellular microenvironment that provides the signals and physical support to maintain stem cells. Asymmetrical divisions Mitotic divisions that generate distinct fates in the two daughter cells; in the case of stem cells, one daughter preserves the stem-cell state. Transit-amplifying cells Dividing stem-cell daughters that are fated for differentiation and that can retain the ability to divide. Totipotent Cells that can produce progeny that can form an entire organism.
Uploads
Papers by Jesús García Ramírez