Luiz Schirmer - Academia.edu (original) (raw)

Papers by Luiz Schirmer

Proceedings of the 20th International Conference on Enterprise Information Systems, 2018

Working with average-sized event logs is still a major task in process mining, where the main goa... more Working with average-sized event logs is still a major task in process mining, where the main goal is to gain process-related insights based on event logs created by a wide variety of systems. An event log contains a sequence of events for every case that was handled by the system. Several discovery algorithms have been proposed and work well in specific cases but fail to be generic strategies. Moreover, there is no evidence that the existing strategies can handle events with a large number of variants. For this reason, a generic approach is needed to allow experts to explore event log data and decompose information into a series of smaller problems, to identify outliers and relations between the analyzed cases. In this paper we present a visual filtering approach for event logs that makes process analysis tasks more feasible and tractable. To evaluate our approach, we have developed a visual filtering tool and used it with the event log from BPI Challenge 2017.

Cornell University - arXiv, Aug 24, 2022

We present MR-Net, a general architecture for multiresolution neural networks, and a framework fo... more We present MR-Net, a general architecture for multiresolution neural networks, and a framework for imaging applications based on this architecture. Our coordinate-based networks are continuous both in space and in scale as they are composed of multiple stages that progressively add finer details. Besides that, they are a compact and efficient representation. We show examples of multiresolution image representation and applications to texture magnification, minification, and antialiasing. This document is the extended version of the paper [PNS + 22]. It includes additional material that would not fit the page limitations of the conference track for publication.

Computers & Graphics

We introduce a neural implicit framework that exploits the differentiable properties of neural ne... more We introduce a neural implicit framework that exploits the differentiable properties of neural networks and the discrete geometry of point-sampled surfaces to approximate them as the level sets of neural implicit functions. To train a neural implicit function, we propose a loss functional that approximates a signed distance function, and allows terms with high-order derivatives, such as the alignment between the principal directions of curvature, to learn more geometric details. During training, we consider a non-uniform sampling strategy based on the curvatures of the point-sampled surface to prioritize points with more geometric details. This sampling implies faster learning while preserving geometric accuracy when compared with previous approaches. We also use the analytical derivatives of a neural implicit function to estimate the differential measures of the underlying point-sampled surface.

Anais Estendidos da XXXIV Conference on Graphics, Patterns and Images (SIBRAPI Estendido 2021)

This thesis proposes new architectures for deep neural networks with attention enhancement and mu... more This thesis proposes new architectures for deep neural networks with attention enhancement and multilinear algebra methods to increase their performance. We also explore graph convolutions and their particularities. We focus here on the problems related to real-time human pose estimation. We explore different architectures to reduce computational complexity, and, as a result, we propose two novel neural network models for 2D and 3D pose estimation. We also introduce a new architecture for Graph attention networks called Semantic Graph Attention.

2021 34th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI)

This survey presents methods that use neural networks for implicit representations of 3D geometry... more This survey presents methods that use neural networks for implicit representations of 3D geometry — neural implicit functions. We explore the different aspects of neural implicit functions for shape modeling and synthesis. We aim to provide a theoretical analysis of 3D shape reconstruction using deep neural networks and introduce a discussion between researchers interested in this research field.

2021 34th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI)

We propose a novel gating mechanism applied to Semantic Graph Convolutions for 3D applications, n... more We propose a novel gating mechanism applied to Semantic Graph Convolutions for 3D applications, named Semantic Graph Attention. Semantic Graph Convolutions learn to capture semantic information such as local and global node relationships, not explicitly represented in graphs. We improve their performance by proposing an attention block to explore channel-wise inter-dependencies. The proposed method performs the unprojection of the 2D points (image) onto their 3D version. We use it to estimate 3d human pose from 2D images. Both 2D and 3D human poses can be represented as structured graphs, exploring their particularities in this context. The attention layer improves the accuracy of skeleton estimation using 58% fewer parameters than state-of-the-art.

This work investigates the use of neural networks admitting high-order derivatives for modeling d... more This work investigates the use of neural networks admitting high-order derivatives for modeling dynamic variations of smooth implicit surfaces. For this purpose, it extends the representation of differentiable neural implicit surfaces to higher dimensions, which opens up mechanisms that allow to exploit geometric transformations in many settings, from animation and surface evolution to shape morphing and design galleries. The problem is modeled by a kkk-parameter family of surfaces ScS_cSc, specified as a neural network function f:mathbbR3timesmathbbRkrightarrowmathbbRf : \mathbb{R}^3 \times \mathbb{R}^k \rightarrow \mathbb{R}f:mathbbR3timesmathbbRkrightarrowmathbbR, where ScS_cSc is the zero-level set of the implicit function f(cdot,c):mathbbR3rightarrowmathbbRf(\cdot, c) : \mathbb{R}^3 \rightarrow \mathbb{R} f(cdot,c):mathbbR3rightarrowmathbbR, with cinmathbbRkc \in \mathbb{R}^kcinmathbbRk, with variations induced by the control variable ccc. In that context, restricted to each coordinate of mathbbRk\mathbb{R}^kmathbbRk, the underlying representation is a neural homotopy which is the solution of a general partial differential equation.

We introduce a neural implicit framework that bridges discrete differential geometry of triangle ... more We introduce a neural implicit framework that bridges discrete differential geometry of triangle meshes and continuous differential geometry of neural implicit surfaces. It exploits the differentiable properties of neural networks and the discrete geometry of triangle meshes to approximate them as the zero-level sets of neural implicit functions. To train a neural implicit function, we propose a loss function that allows terms with high-order derivatives, such as the alignment between the principal directions, to learn more geometric details. During training, we consider a non-uniform sampling strategy based on the discrete curvatures of the triangle mesh to access points with more geometric details. This sampling implies faster learning while preserving geometric accuracy. We present the analytical differential geometry formulas for neural surfaces, such as normal vectors and curvatures. We use them to render the surfaces using sphere tracing. Additionally, we propose a network opt...

2020 33rd SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI), 2020

Pose estimation is a challenging task in computer vision that has many applications, as for examp... more Pose estimation is a challenging task in computer vision that has many applications, as for example: in motion capture, in medical analysis, in human posture monitoring, and in robotics. In other words, it is a main tool to enable machines do understand human patterns in videos or images. Performing this task in real-time while maintaining accuracy and precision is critical for many of these applications. Several papers propose real time approaches considering deep neural networks for pose estimation. However, in most cases they fail when considering run-time performance or do not achieve the precision needed. In this work, we propose a new model for real-time pose estimation considering attention modules for convolutional neural networks (CNNs). We introduce a two-dimensional relative attention mechanism for feature extraction in pose machines leading to improvements in accuracy. We create a single shot architecture where both operations to infer keypoints and part affinity fields share the information. Also, for each stage, we use tensor decompositions to not only reduce dimensionality, but also to improve performance. This allows us to factorize each convolution and drastically reduce the number of parameters in our network. Our experiments show that, with this factorized approach, it is possible to achieve state-of-art performance in terms of run-time while we have a small reduction in accuracy.

The amount of recorded video in the world is increasing a lot due not only to the humans interest... more The amount of recorded video in the world is increasing a lot due not only to the humans interests and habits regarding this kind of media, but also the diversity of devices used to create them. However, there is a lack of information about video content because generating video meta-data is complex. It requires too much time to be performed by humans, and from the technology perspective, it is not easy to overcome obstacles regarding the huge amount and diversity of video frames. The manuscript proposes an automated face recognition system to detect soap opera characters within videos, called Globo Face Stream. It was developed to recognize characters, in order to increase video meta-data. It combines standard computer vision techniques to improved accuracy by processing existing models output data in a complementary manner. The model performed accurately using a real life dataset from a large media company.

Silva, Luiz José Schirmer; Lopes, Hélio Côrtes Vieira (Advisor); Barbosa, Simone Diniz Junqueira ... more Silva, Luiz José Schirmer; Lopes, Hélio Côrtes Vieira (Advisor); Barbosa, Simone Diniz Junqueira (Co-advisor). CrimeVis:An Interactive Visualization System for Analyzing Criminal Data in the State of Rio de Janeiro. Rio de Janeiro, 2016. 53p. MsC Thesis — Departamento de Informática, Pontifícia Universidade Católica do Rio de Janeiro. This work presents the development of an interactive graphic visualization system for analyzing criminal data in the State of Rio de Janeiro, provided by the Public Safety Institute from the State of Rio de Janeiro (ISP-RJ, Instituto de Segurança Pública). The system presents to the user a set of integrated tools that support visualizing and analyzing statistical data on crimes, which make it possible to infer relevant information regarding government policies on public safety and their effects. The tools allow us to visualize multidimensional data, spatiotemporal data, and multivariate data in an integrated manner using brushing and linking techniques...

Proceedings of the ACM Symposium on Document Engineering 2018, 2018

Finding concepts considering their meaning and semantic relations in a document corpus is an impo... more Finding concepts considering their meaning and semantic relations in a document corpus is an important and challenging task. In this paper, we present our contributions on how to understand unstructured data present in one or multiple documents. Generally, the current literature concentrates efforts in structuring knowledge by identifying semantic entities in the data. In this paper, we test our hypothesis that hyperknowledge specifications are capable of defining rich relations among documents and extracted facts. The main evidence supporting this hypothesis is the fact that hyperknowledge was built on top of hypermedia fundamentals, easing the specification of rich relationships between different multimodal components (i.e. multimedia content and knowledge entities). The key challenge tackled in this paper is how to structure and correlate these components considering their meaning and semantic relations.

Enterprise Information Systems, 2019

Dealing with average-sized event logs is considered a challenging task in process mining, in orde... more Dealing with average-sized event logs is considered a challenging task in process mining, in order to give value to event log data created by a wide variety of systems. An event log consists of a sequence of events for every case that was handled by the system. Discovery algorithms proposed in the literature work well in specific cases, but they usually fail in generic ones. Furthermore, there is no evidence that those existing strategies can handle logs with a large number of variants. We lack a generic approach to allow experts to explore event log data and decompose information into a series of smaller problems, to identify not only outliers, but also relations between the analyzed cases. In this chapter we propose a visual approach for filtering processes based on a low dimensionality representation of cases, a dissimilarity function based on both case attributes and case paths, and the use of entropy and silhouette to evaluate the uncertainty and quality, respectively, of each subset of cases. For each subset of cases, it is possible to reconstruct and evaluate each process model. Those contributions can be combined in an interactive tool to support process discovery. To demonstrate our tool, we use the event log from BPI Challenge 2017.

SPE Latin American and Caribbean Petroleum Engineering Conference, 2020

The objective of this work is to develop and train feedforward artificial neural networks (ANNs) ... more The objective of this work is to develop and train feedforward artificial neural networks (ANNs) on the forecasting of layer permeability in heterogeneous reservoirs. The results are validated by comparing the model outputs with permeability curves computed from production logging data. Production logs are used as targets to train the model. A flow-profile interpretation method is used to compute continuous permeability curves free of wellbore skin effects. In addition, segmentation techniques are applied to high-resolution ultrasonic image logs. These logs provide not only the image of the mega-and giga-pore system but can also identify the permeable facies along the reservoir. The image segmentation jointly with other borehole logs provides the necessary features for the network training process. The proposed neural network focuses on delivering reliable and validated permeability curves. Its development accounts for formation skin factor, as well as nongeological noise usually found in ultrasonic image logs. The procedure is tested on both synthetic and field data sets. The estimations presented herein demonstrate the model's ability to learn nonlinear relationships between geological input variables and reservoir dynamic data even if the actual physical relationship is complex and not known a priori. Although the preprocessing stages of the procedure involve some expertise in data interpretation, the neural-network structure can be easily coded in any programming language, requiring no assumptions on physics in advance. For the case studies presented in this work, the proposed procedure provides more accurate permeability curves than the ones obtained from conventional methods, which usually fail to predict the permeability measured on drill-stem tests conducted in dualporosity reservoirs. The novelty of this work is to incorporate dynamic production-logging (PL) data into the permeabilityestimation workflow.

Cornell University - arXiv, Jan 22, 2022

We introduce a novel approach for rendering static and dynamic 3D neural signed distance function... more We introduce a novel approach for rendering static and dynamic 3D neural signed distance functions (SDF) in realtime. We rely on nested neighborhoods of zero-level sets of neural SDFs, and mappings between them. This framework supports animations and achieves real-time performance without the use of spatial data-structures. It consists of three uncoupled algorithms representing the rendering steps. The multiscale sphere tracing focuses on minimizing iteration time by using coarse approximations on earlier iterations. The neural normal mapping transfers details from a fine neural SDF to a surface nested on a neighborhood of its zero-level set. It is smooth and it does not depend on surface parametrizations. As a result, it can be used to fetch smooth normals for discrete surfaces such as meshes and to skip later iterations when sphere tracing level sets. Finally, we propose an algorithm for analytic normal calculation for MLPs and describe ways to obtain sequences of neural SDFs to use with the algorithms.

This work investigates the use of neural networks admitting high-order derivatives for modeling d... more This work investigates the use of neural networks admitting high-order derivatives for modeling dynamic variations of smooth implicit surfaces. For this purpose, it extends the representation of differentiable neural implicit surfaces to higher dimensions, which opens up mechanisms that allow to exploit geometric transformations in many settings, from animation and surface evolution to shape morphing and design galleries. The problem is modeled by a kkk-parameter family of surfaces ScS_cSc, specified as a neural network function f:mathbbR3timesmathbbRkrightarrowmathbbRf : \mathbb{R}^3 \times \mathbb{R}^k \rightarrow \mathbb{R}f:mathbbR3timesmathbbRkrightarrowmathbbR, where ScS_cSc is the zero-level set of the implicit function f(cdot,c):mathbbR3rightarrowmathbbRf(\cdot, c) : \mathbb{R}^3 \rightarrow \mathbb{R} f(cdot,c):mathbbR3rightarrowmathbbR, with cinmathbbRkc \in \mathbb{R}^kcinmathbbRk, with variations induced by the control variable ccc. In that context, restricted to each coordinate of mathbbRk\mathbb{R}^kmathbbRk, the underlying representation is a neural homotopy which is the solution of a general partial differential equation.

SPE Latin American and Caribbean Petroleum Engineering Conference

SPE Journal

Summary The objective of this work is to develop and train feedforward artificial neural networks... more Summary The objective of this work is to develop and train feedforward artificial neural networks (ANNs) on the forecasting of layer permeability in heterogeneous reservoirs. The results are validated by comparing the model outputs with permeability curves computed from production logging data. Production logs are used as targets to train the model. A flow-profile interpretation method is used to compute continuous permeability curves free of wellbore skin effects. In addition, segmentation techniques are applied to high-resolution ultrasonic image logs. These logs provide not only the image of the mega- and giga-pore system but can also identify the permeable facies along the reservoir. The image segmentation jointly with other borehole logs provides the necessary features for the network training process. The proposed neural network focuses on delivering reliable and validated permeability curves. Its development accounts for formation skin factor, as well as nongeological noise u...

Working with average-sized event logs is still a major task in process mining, where the main goa... more Working with average-sized event logs is still a major task in process mining, where the main goal is to gain process-related insights based on event logs created by a wide variety of systems. An event log contains a sequence of events for every case that was handled by the system. Several discovery algorithms have been proposed and work well in specific cases but fail to be generic strategies. Moreover, there is no evidence that the existing strategies can handle events with a large number of variants. For this reason, a generic approach is needed to allow experts to explore event log data and decompose information into a series of smaller problems, to identify outliers and relations between the analyzed cases. In this paper we present a visual filtering approach for event logs that makes process analysis tasks more feasible and tractable. To evaluate our approach, we have developed a visual filtering tool and used it with the event log from BPI

Proceedings of the 20th International Conference on Enterprise Information Systems, 2018

Working with average-sized event logs is still a major task in process mining, where the main goa... more Working with average-sized event logs is still a major task in process mining, where the main goal is to gain process-related insights based on event logs created by a wide variety of systems. An event log contains a sequence of events for every case that was handled by the system. Several discovery algorithms have been proposed and work well in specific cases but fail to be generic strategies. Moreover, there is no evidence that the existing strategies can handle events with a large number of variants. For this reason, a generic approach is needed to allow experts to explore event log data and decompose information into a series of smaller problems, to identify outliers and relations between the analyzed cases. In this paper we present a visual filtering approach for event logs that makes process analysis tasks more feasible and tractable. To evaluate our approach, we have developed a visual filtering tool and used it with the event log from BPI Challenge 2017.

Cornell University - arXiv, Aug 24, 2022

We present MR-Net, a general architecture for multiresolution neural networks, and a framework fo... more We present MR-Net, a general architecture for multiresolution neural networks, and a framework for imaging applications based on this architecture. Our coordinate-based networks are continuous both in space and in scale as they are composed of multiple stages that progressively add finer details. Besides that, they are a compact and efficient representation. We show examples of multiresolution image representation and applications to texture magnification, minification, and antialiasing. This document is the extended version of the paper [PNS + 22]. It includes additional material that would not fit the page limitations of the conference track for publication.

Computers & Graphics

We introduce a neural implicit framework that exploits the differentiable properties of neural ne... more We introduce a neural implicit framework that exploits the differentiable properties of neural networks and the discrete geometry of point-sampled surfaces to approximate them as the level sets of neural implicit functions. To train a neural implicit function, we propose a loss functional that approximates a signed distance function, and allows terms with high-order derivatives, such as the alignment between the principal directions of curvature, to learn more geometric details. During training, we consider a non-uniform sampling strategy based on the curvatures of the point-sampled surface to prioritize points with more geometric details. This sampling implies faster learning while preserving geometric accuracy when compared with previous approaches. We also use the analytical derivatives of a neural implicit function to estimate the differential measures of the underlying point-sampled surface.

Anais Estendidos da XXXIV Conference on Graphics, Patterns and Images (SIBRAPI Estendido 2021)

This thesis proposes new architectures for deep neural networks with attention enhancement and mu... more This thesis proposes new architectures for deep neural networks with attention enhancement and multilinear algebra methods to increase their performance. We also explore graph convolutions and their particularities. We focus here on the problems related to real-time human pose estimation. We explore different architectures to reduce computational complexity, and, as a result, we propose two novel neural network models for 2D and 3D pose estimation. We also introduce a new architecture for Graph attention networks called Semantic Graph Attention.

2021 34th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI)

This survey presents methods that use neural networks for implicit representations of 3D geometry... more This survey presents methods that use neural networks for implicit representations of 3D geometry — neural implicit functions. We explore the different aspects of neural implicit functions for shape modeling and synthesis. We aim to provide a theoretical analysis of 3D shape reconstruction using deep neural networks and introduce a discussion between researchers interested in this research field.

2021 34th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI)

We propose a novel gating mechanism applied to Semantic Graph Convolutions for 3D applications, n... more We propose a novel gating mechanism applied to Semantic Graph Convolutions for 3D applications, named Semantic Graph Attention. Semantic Graph Convolutions learn to capture semantic information such as local and global node relationships, not explicitly represented in graphs. We improve their performance by proposing an attention block to explore channel-wise inter-dependencies. The proposed method performs the unprojection of the 2D points (image) onto their 3D version. We use it to estimate 3d human pose from 2D images. Both 2D and 3D human poses can be represented as structured graphs, exploring their particularities in this context. The attention layer improves the accuracy of skeleton estimation using 58% fewer parameters than state-of-the-art.

This work investigates the use of neural networks admitting high-order derivatives for modeling d... more This work investigates the use of neural networks admitting high-order derivatives for modeling dynamic variations of smooth implicit surfaces. For this purpose, it extends the representation of differentiable neural implicit surfaces to higher dimensions, which opens up mechanisms that allow to exploit geometric transformations in many settings, from animation and surface evolution to shape morphing and design galleries. The problem is modeled by a kkk-parameter family of surfaces ScS_cSc, specified as a neural network function f:mathbbR3timesmathbbRkrightarrowmathbbRf : \mathbb{R}^3 \times \mathbb{R}^k \rightarrow \mathbb{R}f:mathbbR3timesmathbbRkrightarrowmathbbR, where ScS_cSc is the zero-level set of the implicit function f(cdot,c):mathbbR3rightarrowmathbbRf(\cdot, c) : \mathbb{R}^3 \rightarrow \mathbb{R} f(cdot,c):mathbbR3rightarrowmathbbR, with cinmathbbRkc \in \mathbb{R}^kcinmathbbRk, with variations induced by the control variable ccc. In that context, restricted to each coordinate of mathbbRk\mathbb{R}^kmathbbRk, the underlying representation is a neural homotopy which is the solution of a general partial differential equation.

We introduce a neural implicit framework that bridges discrete differential geometry of triangle ... more We introduce a neural implicit framework that bridges discrete differential geometry of triangle meshes and continuous differential geometry of neural implicit surfaces. It exploits the differentiable properties of neural networks and the discrete geometry of triangle meshes to approximate them as the zero-level sets of neural implicit functions. To train a neural implicit function, we propose a loss function that allows terms with high-order derivatives, such as the alignment between the principal directions, to learn more geometric details. During training, we consider a non-uniform sampling strategy based on the discrete curvatures of the triangle mesh to access points with more geometric details. This sampling implies faster learning while preserving geometric accuracy. We present the analytical differential geometry formulas for neural surfaces, such as normal vectors and curvatures. We use them to render the surfaces using sphere tracing. Additionally, we propose a network opt...

2020 33rd SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI), 2020

Pose estimation is a challenging task in computer vision that has many applications, as for examp... more Pose estimation is a challenging task in computer vision that has many applications, as for example: in motion capture, in medical analysis, in human posture monitoring, and in robotics. In other words, it is a main tool to enable machines do understand human patterns in videos or images. Performing this task in real-time while maintaining accuracy and precision is critical for many of these applications. Several papers propose real time approaches considering deep neural networks for pose estimation. However, in most cases they fail when considering run-time performance or do not achieve the precision needed. In this work, we propose a new model for real-time pose estimation considering attention modules for convolutional neural networks (CNNs). We introduce a two-dimensional relative attention mechanism for feature extraction in pose machines leading to improvements in accuracy. We create a single shot architecture where both operations to infer keypoints and part affinity fields share the information. Also, for each stage, we use tensor decompositions to not only reduce dimensionality, but also to improve performance. This allows us to factorize each convolution and drastically reduce the number of parameters in our network. Our experiments show that, with this factorized approach, it is possible to achieve state-of-art performance in terms of run-time while we have a small reduction in accuracy.

The amount of recorded video in the world is increasing a lot due not only to the humans interest... more The amount of recorded video in the world is increasing a lot due not only to the humans interests and habits regarding this kind of media, but also the diversity of devices used to create them. However, there is a lack of information about video content because generating video meta-data is complex. It requires too much time to be performed by humans, and from the technology perspective, it is not easy to overcome obstacles regarding the huge amount and diversity of video frames. The manuscript proposes an automated face recognition system to detect soap opera characters within videos, called Globo Face Stream. It was developed to recognize characters, in order to increase video meta-data. It combines standard computer vision techniques to improved accuracy by processing existing models output data in a complementary manner. The model performed accurately using a real life dataset from a large media company.

Silva, Luiz José Schirmer; Lopes, Hélio Côrtes Vieira (Advisor); Barbosa, Simone Diniz Junqueira ... more Silva, Luiz José Schirmer; Lopes, Hélio Côrtes Vieira (Advisor); Barbosa, Simone Diniz Junqueira (Co-advisor). CrimeVis:An Interactive Visualization System for Analyzing Criminal Data in the State of Rio de Janeiro. Rio de Janeiro, 2016. 53p. MsC Thesis — Departamento de Informática, Pontifícia Universidade Católica do Rio de Janeiro. This work presents the development of an interactive graphic visualization system for analyzing criminal data in the State of Rio de Janeiro, provided by the Public Safety Institute from the State of Rio de Janeiro (ISP-RJ, Instituto de Segurança Pública). The system presents to the user a set of integrated tools that support visualizing and analyzing statistical data on crimes, which make it possible to infer relevant information regarding government policies on public safety and their effects. The tools allow us to visualize multidimensional data, spatiotemporal data, and multivariate data in an integrated manner using brushing and linking techniques...

Proceedings of the ACM Symposium on Document Engineering 2018, 2018

Finding concepts considering their meaning and semantic relations in a document corpus is an impo... more Finding concepts considering their meaning and semantic relations in a document corpus is an important and challenging task. In this paper, we present our contributions on how to understand unstructured data present in one or multiple documents. Generally, the current literature concentrates efforts in structuring knowledge by identifying semantic entities in the data. In this paper, we test our hypothesis that hyperknowledge specifications are capable of defining rich relations among documents and extracted facts. The main evidence supporting this hypothesis is the fact that hyperknowledge was built on top of hypermedia fundamentals, easing the specification of rich relationships between different multimodal components (i.e. multimedia content and knowledge entities). The key challenge tackled in this paper is how to structure and correlate these components considering their meaning and semantic relations.

Enterprise Information Systems, 2019

Dealing with average-sized event logs is considered a challenging task in process mining, in orde... more Dealing with average-sized event logs is considered a challenging task in process mining, in order to give value to event log data created by a wide variety of systems. An event log consists of a sequence of events for every case that was handled by the system. Discovery algorithms proposed in the literature work well in specific cases, but they usually fail in generic ones. Furthermore, there is no evidence that those existing strategies can handle logs with a large number of variants. We lack a generic approach to allow experts to explore event log data and decompose information into a series of smaller problems, to identify not only outliers, but also relations between the analyzed cases. In this chapter we propose a visual approach for filtering processes based on a low dimensionality representation of cases, a dissimilarity function based on both case attributes and case paths, and the use of entropy and silhouette to evaluate the uncertainty and quality, respectively, of each subset of cases. For each subset of cases, it is possible to reconstruct and evaluate each process model. Those contributions can be combined in an interactive tool to support process discovery. To demonstrate our tool, we use the event log from BPI Challenge 2017.

SPE Latin American and Caribbean Petroleum Engineering Conference, 2020

The objective of this work is to develop and train feedforward artificial neural networks (ANNs) ... more The objective of this work is to develop and train feedforward artificial neural networks (ANNs) on the forecasting of layer permeability in heterogeneous reservoirs. The results are validated by comparing the model outputs with permeability curves computed from production logging data. Production logs are used as targets to train the model. A flow-profile interpretation method is used to compute continuous permeability curves free of wellbore skin effects. In addition, segmentation techniques are applied to high-resolution ultrasonic image logs. These logs provide not only the image of the mega-and giga-pore system but can also identify the permeable facies along the reservoir. The image segmentation jointly with other borehole logs provides the necessary features for the network training process. The proposed neural network focuses on delivering reliable and validated permeability curves. Its development accounts for formation skin factor, as well as nongeological noise usually found in ultrasonic image logs. The procedure is tested on both synthetic and field data sets. The estimations presented herein demonstrate the model's ability to learn nonlinear relationships between geological input variables and reservoir dynamic data even if the actual physical relationship is complex and not known a priori. Although the preprocessing stages of the procedure involve some expertise in data interpretation, the neural-network structure can be easily coded in any programming language, requiring no assumptions on physics in advance. For the case studies presented in this work, the proposed procedure provides more accurate permeability curves than the ones obtained from conventional methods, which usually fail to predict the permeability measured on drill-stem tests conducted in dualporosity reservoirs. The novelty of this work is to incorporate dynamic production-logging (PL) data into the permeabilityestimation workflow.

Cornell University - arXiv, Jan 22, 2022

We introduce a novel approach for rendering static and dynamic 3D neural signed distance function... more We introduce a novel approach for rendering static and dynamic 3D neural signed distance functions (SDF) in realtime. We rely on nested neighborhoods of zero-level sets of neural SDFs, and mappings between them. This framework supports animations and achieves real-time performance without the use of spatial data-structures. It consists of three uncoupled algorithms representing the rendering steps. The multiscale sphere tracing focuses on minimizing iteration time by using coarse approximations on earlier iterations. The neural normal mapping transfers details from a fine neural SDF to a surface nested on a neighborhood of its zero-level set. It is smooth and it does not depend on surface parametrizations. As a result, it can be used to fetch smooth normals for discrete surfaces such as meshes and to skip later iterations when sphere tracing level sets. Finally, we propose an algorithm for analytic normal calculation for MLPs and describe ways to obtain sequences of neural SDFs to use with the algorithms.

This work investigates the use of neural networks admitting high-order derivatives for modeling d... more This work investigates the use of neural networks admitting high-order derivatives for modeling dynamic variations of smooth implicit surfaces. For this purpose, it extends the representation of differentiable neural implicit surfaces to higher dimensions, which opens up mechanisms that allow to exploit geometric transformations in many settings, from animation and surface evolution to shape morphing and design galleries. The problem is modeled by a kkk-parameter family of surfaces ScS_cSc, specified as a neural network function f:mathbbR3timesmathbbRkrightarrowmathbbRf : \mathbb{R}^3 \times \mathbb{R}^k \rightarrow \mathbb{R}f:mathbbR3timesmathbbRkrightarrowmathbbR, where ScS_cSc is the zero-level set of the implicit function f(cdot,c):mathbbR3rightarrowmathbbRf(\cdot, c) : \mathbb{R}^3 \rightarrow \mathbb{R} f(cdot,c):mathbbR3rightarrowmathbbR, with cinmathbbRkc \in \mathbb{R}^kcinmathbbRk, with variations induced by the control variable ccc. In that context, restricted to each coordinate of mathbbRk\mathbb{R}^kmathbbRk, the underlying representation is a neural homotopy which is the solution of a general partial differential equation.

SPE Latin American and Caribbean Petroleum Engineering Conference

SPE Journal

Summary The objective of this work is to develop and train feedforward artificial neural networks... more Summary The objective of this work is to develop and train feedforward artificial neural networks (ANNs) on the forecasting of layer permeability in heterogeneous reservoirs. The results are validated by comparing the model outputs with permeability curves computed from production logging data. Production logs are used as targets to train the model. A flow-profile interpretation method is used to compute continuous permeability curves free of wellbore skin effects. In addition, segmentation techniques are applied to high-resolution ultrasonic image logs. These logs provide not only the image of the mega- and giga-pore system but can also identify the permeable facies along the reservoir. The image segmentation jointly with other borehole logs provides the necessary features for the network training process. The proposed neural network focuses on delivering reliable and validated permeability curves. Its development accounts for formation skin factor, as well as nongeological noise u...

Working with average-sized event logs is still a major task in process mining, where the main goa... more Working with average-sized event logs is still a major task in process mining, where the main goal is to gain process-related insights based on event logs created by a wide variety of systems. An event log contains a sequence of events for every case that was handled by the system. Several discovery algorithms have been proposed and work well in specific cases but fail to be generic strategies. Moreover, there is no evidence that the existing strategies can handle events with a large number of variants. For this reason, a generic approach is needed to allow experts to explore event log data and decompose information into a series of smaller problems, to identify outliers and relations between the analyzed cases. In this paper we present a visual filtering approach for event logs that makes process analysis tasks more feasible and tractable. To evaluate our approach, we have developed a visual filtering tool and used it with the event log from BPI