Doctor en Ciencias de la Ingeniería, mención Ciencia de la Computación, de la Pontificia Universidad Católica de Chile. Su tesis titulada «Un Enfoque de Reconocimiento Visual basado en Modelos Jerarquicos Composicionales Profundos, Eficientes y Compactos y Técnicas de Máximo Margen Estructuralmente Regularizadas», centrada en tópicos avanzados de Aprendizaje de Máquina y Reconocimiento Visual, recibió el premio a la mejor tesis doctoral del año 2016 de la Pontificia Universidad Católica de Chile en el área de Tecnología, Ingeniería y Procesos Productivos. Recibió además su grado de Magíster en Ciencias de la Ingeniería, mención Ciencia de la Computación y su título de Ingeniero Civil de Computación, ambos de la Pontificia Universidad Católica de Chile, en 2009 y 2007 respectivamente.
Desde el año 2016 es Profesor Asistente de los departamentos de Ingeniería de Transporte y Logística y de Ciencia de la Computación. Sus áreas de interés se centran en el desarrollo y aplicación de nuevos métodos de aprendizaje de máquina a problemas relacionados con Smart Cities y Sistemas Inteligentes de Transporte, donde se aproveche el uso de volúmenes masivos de datos de fuentes heterogéneas.
This is the dataset used in the paper: Automatic Document Screening of Medical Literature Using W... more This is the dataset used in the paper: Automatic Document Screening of Medical Literature Using Word and Text Embeddings in an Active Learning Setting. It is composed of: - Pre-trained models using active learning for document screening on HealthCLEF and Epistemonikos datasets. - Epistemonikos and HealthCLEF datasets containing medical questions and relevant/non relevant articles. - Embeddings and Document Representations used for experiments on both datasets. Scripts to run experiments can be found at: https://github.com/afcarvallo/active_learning_document_screening <strong>Paper abstract:</strong> Document screening is a fundamental task within Evidence-based Medicine (EBM), a practice that provides scientific evidence to support medical decisions. Several approaches have tried to reduce physicians' workload of screening and labeling vast amounts of documents to answer clinical questions. Previous works tried to semi-automate document screening, reporting promising...
Santiago's public transit system uses a Passenger Service Quality Index (ICA) to measure the ... more Santiago's public transit system uses a Passenger Service Quality Index (ICA) to measure the quality of service offered by buses companies. Parts of this index are related to bus driver's behavior, and are obtained in a superficial and very subjective manner. The main objective of this research is to formulate a new methodology that uses data provided by inertial measurement units to classify drivers' behavior. This is achieved by means of a classification method: decision trees. Data are collected to evaluate the method and results show that the use of decision trees delivers good performance and an interpretable output that allows further analysis. The proposal uses elements from the ICA index and produces a methodology that is simple, objective and capable of being implemented on a large scale with good performance at a low cost.
Abstract People’s perceptions of the built environment influence the way they use and navigate it... more Abstract People’s perceptions of the built environment influence the way they use and navigate it. Understanding these perceptions may be useful to inform the design, management and planning process of public spaces. Recently, several studies have used data collected at a massive scale and machine learning methods to quantify these perceptions, showing promising results in terms of predictive performance. Nevertheless, most of these models can be of little help in understanding users’ perceptions due to the difficulty associated with identifying the importance of each attribute of landscapes. In this work, we propose a novel approach to quantify perceptions of landscapes through discrete choice models, using semantic segmentations of images of public spaces, generated through machine learning algorithms, as explanatory variables. The proposed models are estimated using the Place Pulse dataset, with over 1.2 million perceptual indicators, and are able to provide useful insights into how users perceive the built environment as a function of its features. The models obtained are used to infer perceptual variables in the city of Santiago, Chile, and show they have a significant correlation with socioeconomic indicators.
Abstract. The Bag-of-Visual-Words (BoVW) model is a popular approach for visual recognition. Used... more Abstract. The Bag-of-Visual-Words (BoVW) model is a popular approach for visual recognition. Used successfully in many different tasks, simplicity and good performance are the main reasons for its popularity. The central aspect of this model, the visual dictionary, is used to build mid-level representations based on low level image descriptors. Classifiers are then trained using these mid-level representations to perform categorization. While most works based on BoVW models have been focused on learning a suitable dictionary or on proposing a suitable pooling strategy, little effort has been devoted to explore and improve the coupling between the dictionary and the top-level classifiers, in order to gen-erate more discriminative models. This problem can be highly complex due to the large dictionary size usually needed by these methods. Also, most BoVW based systems usually perform multiclass categorization using a one-vs-all strat-egy, ignoring relevant correlations among classes. T...
This article describes the participation and results of the PUC Chile team in the Turberculosis t... more This article describes the participation and results of the PUC Chile team in the Turberculosis task in the context of ImageCLEFmedical challenge 2021. We were ranked 7th based on the kappa metric and 4th in terms of accuracy. We describe three approaches we tried in order to address the task. Our best approach used 2D images visually encoded with a DenseNet neural network, which representations were concatenated to finally output the classification with a softmax layer. We describe in detail this and other two approaches, and we conclude by discussing some ideas for future work.
This article describes PUC Chile team’s participation in the Concept Detection task of ImageCLEFm... more This article describes PUC Chile team’s participation in the Concept Detection task of ImageCLEFmedical challenge 2021, which resulted in the team earning the fourth place. We made two submissions, the first one based on a naive approach which resulted in a F-1 score of 0.141, and an improved version which leveraged the Perceptual Similarity among images and obtained a final F-1 score of 0.360. We describe in detail our data analysis, our different approaches, and conclude by discussing some ideas for future work.
This article describes PUC Chile team’s participation in the Caption Prediction task of ImageCLEF... more This article describes PUC Chile team’s participation in the Caption Prediction task of ImageCLEFmedical challenge 2021, which resulted in the team winning this task. We first show how a very simple approach based on statistical analysis of captions, without relying on images, results in a competitive baseline score. Then, we describe how to improve the performance of this preliminary submission by encoding the medical images with a ResNet CNN, pre-trained on ImageNet and later fine-tuned with the challenge dataset. Afterwards, we use this visual encoding as the input for a multi-label classification approach for caption prediction. We describe in detail our final approach, and we conclude by discussing some ideas for future work.
Federated Learning aims to train distributed deep models without sharing the raw data with the ce... more Federated Learning aims to train distributed deep models without sharing the raw data with the centralized server. Similarly, in Split Learning, by partitioning a neural network and distributing it across several physical nodes, activations and gradients are exchanged between physical nodes, rather than raw data. Nevertheless, when a neural network is partitioned and distributed among physical nodes, failure of physical nodes causes the failure of the neural units that are placed on those nodes, which results in a significant performance drop. Current approaches focus on resiliency of training in distributed neural networks. However, resiliency of inference in distributed neural networks is less explored. We introduce ResiliNet, a scheme for making inference in distributed neural networks resilient to physical node failures. ResiliNet combines two concepts to provide resiliency: skip hyperconnection, a concept for skipping nodes in distributed neural networks similar to skip connect...
Continuous learning occurs naturally in human beings. However, Deep Learning methods suffer from ... more Continuous learning occurs naturally in human beings. However, Deep Learning methods suffer from a problem known as Catastrophic Forgetting (CF) that consists of a model drastically decreasing its performance on previously learned tasks when it is sequentially trained on new tasks. This situation, known as task interference, occurs when a network modifies relevant weight values as it learns a new task. In this work, we propose two main strategies to face the problem of task interference in convolutional neural networks. First, we use a sparse coding technique to adaptively allocate model capacity to different tasks avoiding interference between them. Specifically, we use a strategy based on group sparse regularization to specialize groups of parameters to learn each task. Afterward, by adding binary masks, we can freeze these groups of parameters, using the rest of the network to learn new tasks. Second, we use a meta learning technique to foster knowledge transfer among tasks, enco...
When a neural network is partitioned and distributed across physical nodes, failure of physical n... more When a neural network is partitioned and distributed across physical nodes, failure of physical nodes causes the failure of the neural units that are placed on those nodes, which results in a significant performance drop. Current approaches focus on resiliency of training in distributed neural networks. However, resiliency of inference in distributed neural networks is less explored. We introduce ResiliNet, a scheme for making inference in distributed neural networks resilient to physical node failures. ResiliNet combines two concepts to provide resiliency: skip connection in residual neural networks, and a novel technique called failout, which is introduced in this paper. Failout simulates physical node failure conditions during training using dropout, and is specifically designed to improve the resiliency of distributed neural networks. The results of the experiments and ablation studies using three datasets confirm the ability of ResiliNet to provide inference resiliency for dist...
This is the dataset used in the paper: Automatic Document Screening of Medical Literature Using W... more This is the dataset used in the paper: Automatic Document Screening of Medical Literature Using Word and Text Embeddings in an Active Learning Setting. It is composed of: - Pre-trained models using active learning for document screening on HealthCLEF and Epistemonikos datasets. - Epistemonikos and HealthCLEF datasets containing medical questions and relevant/non relevant articles. - Embeddings and Document Representations used for experiments on both datasets. Scripts to run experiments can be found at: https://github.com/afcarvallo/active_learning_document_screening <strong>Paper abstract:</strong> Document screening is a fundamental task within Evidence-based Medicine (EBM), a practice that provides scientific evidence to support medical decisions. Several approaches have tried to reduce physicians' workload of screening and labeling vast amounts of documents to answer clinical questions. Previous works tried to semi-automate document screening, reporting promising...
Santiago's public transit system uses a Passenger Service Quality Index (ICA) to measure the ... more Santiago's public transit system uses a Passenger Service Quality Index (ICA) to measure the quality of service offered by buses companies. Parts of this index are related to bus driver's behavior, and are obtained in a superficial and very subjective manner. The main objective of this research is to formulate a new methodology that uses data provided by inertial measurement units to classify drivers' behavior. This is achieved by means of a classification method: decision trees. Data are collected to evaluate the method and results show that the use of decision trees delivers good performance and an interpretable output that allows further analysis. The proposal uses elements from the ICA index and produces a methodology that is simple, objective and capable of being implemented on a large scale with good performance at a low cost.
Abstract People’s perceptions of the built environment influence the way they use and navigate it... more Abstract People’s perceptions of the built environment influence the way they use and navigate it. Understanding these perceptions may be useful to inform the design, management and planning process of public spaces. Recently, several studies have used data collected at a massive scale and machine learning methods to quantify these perceptions, showing promising results in terms of predictive performance. Nevertheless, most of these models can be of little help in understanding users’ perceptions due to the difficulty associated with identifying the importance of each attribute of landscapes. In this work, we propose a novel approach to quantify perceptions of landscapes through discrete choice models, using semantic segmentations of images of public spaces, generated through machine learning algorithms, as explanatory variables. The proposed models are estimated using the Place Pulse dataset, with over 1.2 million perceptual indicators, and are able to provide useful insights into how users perceive the built environment as a function of its features. The models obtained are used to infer perceptual variables in the city of Santiago, Chile, and show they have a significant correlation with socioeconomic indicators.
Abstract. The Bag-of-Visual-Words (BoVW) model is a popular approach for visual recognition. Used... more Abstract. The Bag-of-Visual-Words (BoVW) model is a popular approach for visual recognition. Used successfully in many different tasks, simplicity and good performance are the main reasons for its popularity. The central aspect of this model, the visual dictionary, is used to build mid-level representations based on low level image descriptors. Classifiers are then trained using these mid-level representations to perform categorization. While most works based on BoVW models have been focused on learning a suitable dictionary or on proposing a suitable pooling strategy, little effort has been devoted to explore and improve the coupling between the dictionary and the top-level classifiers, in order to gen-erate more discriminative models. This problem can be highly complex due to the large dictionary size usually needed by these methods. Also, most BoVW based systems usually perform multiclass categorization using a one-vs-all strat-egy, ignoring relevant correlations among classes. T...
This article describes the participation and results of the PUC Chile team in the Turberculosis t... more This article describes the participation and results of the PUC Chile team in the Turberculosis task in the context of ImageCLEFmedical challenge 2021. We were ranked 7th based on the kappa metric and 4th in terms of accuracy. We describe three approaches we tried in order to address the task. Our best approach used 2D images visually encoded with a DenseNet neural network, which representations were concatenated to finally output the classification with a softmax layer. We describe in detail this and other two approaches, and we conclude by discussing some ideas for future work.
This article describes PUC Chile team’s participation in the Concept Detection task of ImageCLEFm... more This article describes PUC Chile team’s participation in the Concept Detection task of ImageCLEFmedical challenge 2021, which resulted in the team earning the fourth place. We made two submissions, the first one based on a naive approach which resulted in a F-1 score of 0.141, and an improved version which leveraged the Perceptual Similarity among images and obtained a final F-1 score of 0.360. We describe in detail our data analysis, our different approaches, and conclude by discussing some ideas for future work.
This article describes PUC Chile team’s participation in the Caption Prediction task of ImageCLEF... more This article describes PUC Chile team’s participation in the Caption Prediction task of ImageCLEFmedical challenge 2021, which resulted in the team winning this task. We first show how a very simple approach based on statistical analysis of captions, without relying on images, results in a competitive baseline score. Then, we describe how to improve the performance of this preliminary submission by encoding the medical images with a ResNet CNN, pre-trained on ImageNet and later fine-tuned with the challenge dataset. Afterwards, we use this visual encoding as the input for a multi-label classification approach for caption prediction. We describe in detail our final approach, and we conclude by discussing some ideas for future work.
Federated Learning aims to train distributed deep models without sharing the raw data with the ce... more Federated Learning aims to train distributed deep models without sharing the raw data with the centralized server. Similarly, in Split Learning, by partitioning a neural network and distributing it across several physical nodes, activations and gradients are exchanged between physical nodes, rather than raw data. Nevertheless, when a neural network is partitioned and distributed among physical nodes, failure of physical nodes causes the failure of the neural units that are placed on those nodes, which results in a significant performance drop. Current approaches focus on resiliency of training in distributed neural networks. However, resiliency of inference in distributed neural networks is less explored. We introduce ResiliNet, a scheme for making inference in distributed neural networks resilient to physical node failures. ResiliNet combines two concepts to provide resiliency: skip hyperconnection, a concept for skipping nodes in distributed neural networks similar to skip connect...
Continuous learning occurs naturally in human beings. However, Deep Learning methods suffer from ... more Continuous learning occurs naturally in human beings. However, Deep Learning methods suffer from a problem known as Catastrophic Forgetting (CF) that consists of a model drastically decreasing its performance on previously learned tasks when it is sequentially trained on new tasks. This situation, known as task interference, occurs when a network modifies relevant weight values as it learns a new task. In this work, we propose two main strategies to face the problem of task interference in convolutional neural networks. First, we use a sparse coding technique to adaptively allocate model capacity to different tasks avoiding interference between them. Specifically, we use a strategy based on group sparse regularization to specialize groups of parameters to learn each task. Afterward, by adding binary masks, we can freeze these groups of parameters, using the rest of the network to learn new tasks. Second, we use a meta learning technique to foster knowledge transfer among tasks, enco...
When a neural network is partitioned and distributed across physical nodes, failure of physical n... more When a neural network is partitioned and distributed across physical nodes, failure of physical nodes causes the failure of the neural units that are placed on those nodes, which results in a significant performance drop. Current approaches focus on resiliency of training in distributed neural networks. However, resiliency of inference in distributed neural networks is less explored. We introduce ResiliNet, a scheme for making inference in distributed neural networks resilient to physical node failures. ResiliNet combines two concepts to provide resiliency: skip connection in residual neural networks, and a novel technique called failout, which is introduced in this paper. Failout simulates physical node failure conditions during training using dropout, and is specifically designed to improve the resiliency of distributed neural networks. The results of the experiments and ablation studies using three datasets confirm the ability of ResiliNet to provide inference resiliency for dist...
The School of Engineering at the Pontificia Universidad Católica de Chile, one of the leading... more The School of Engineering at the Pontificia Universidad Católica de Chile, one of the leading engineering academic institutions in Latin America and ranked among the top four emerging leaders for engineering education worldwide, invites outstanding candidates for a full-time faculty joint position in the area of Digital Transformation between the Department of Industrial and Systems Engineering and the Department of Computer Science. A successful Digital Transformation researcher must have a twofold profile: a strong background in Management (e.g., Organizational Behavior, Strategy, Entrepreneurship, Innovation, and/or Sociology), as well as in Computing and Information Systems.
Duties
High-quality teaching at undergraduate and graduate levels. We expect the applicant to be able to teach courses in both the Department of Industrial and Systems Engineering (e.g., Technology Strategy, Entrepreneurship, and Innovation, Competitive Intelligence) and the Department of Computer Science (e.g., Information Systems, Digital Transformation and Information Technology, Information Technology Management, IT Strategy).
Independente research: the applicant should conduct research in any of the topics included in Digital Transformation. Additional duties include knowledge transfer, outreach, and university administrative tasks.
For more information about the application process, you can visit dtpos.ing.puc.cl, “Faculty Position in Digital Transformation”.
Uploads
Papers by Hans Lobel
Duties
High-quality teaching at undergraduate and graduate levels. We expect the applicant to be able to teach courses in both the Department of Industrial and Systems Engineering (e.g., Technology Strategy, Entrepreneurship, and Innovation, Competitive Intelligence) and the Department of Computer Science (e.g., Information Systems, Digital Transformation and Information Technology, Information Technology Management, IT Strategy).
Independente research: the applicant should conduct research in any of the topics included in Digital Transformation. Additional duties include knowledge transfer, outreach, and university administrative tasks.
For more information about the application process, you can visit dtpos.ing.puc.cl, “Faculty Position in Digital Transformation”.