-
Investigation of Federated Learning Algorithms for Retinal Optical Coherence Tomography Image Classification with Statistical Heterogeneity
Authors:
Sanskar Amgain,
Prashant Shrestha,
Sophia Bano,
Ignacio del Valle Torres,
Michael Cunniffe,
Victor Hernandez,
Phil Beales,
Binod Bhattarai
Abstract:
Purpose: We apply federated learning to train an OCT image classifier simulating a realistic scenario with multiple clients and statistical heterogeneous data distribution where data in the clients lack samples of some categories entirely.
Methods: We investigate the effectiveness of FedAvg and FedProx to train an OCT image classification model in a decentralized fashion, addressing privacy conc…
▽ More
Purpose: We apply federated learning to train an OCT image classifier simulating a realistic scenario with multiple clients and statistical heterogeneous data distribution where data in the clients lack samples of some categories entirely.
Methods: We investigate the effectiveness of FedAvg and FedProx to train an OCT image classification model in a decentralized fashion, addressing privacy concerns associated with centralizing data. We partitioned a publicly available OCT dataset across multiple clients under IID and Non-IID settings and conducted local training on the subsets for each client. We evaluated two federated learning methods, FedAvg and FedProx for these settings.
Results: Our experiments on the dataset suggest that under IID settings, both methods perform on par with training on a central data pool. However, the performance of both algorithms declines as we increase the statistical heterogeneity across the client data, while FedProx consistently performs better than FedAvg in the increased heterogeneity settings.
Conclusion: Despite the effectiveness of federated learning in the utilization of private data across multiple medical institutions, the large number of clients and heterogeneous distribution of labels deteriorate the performance of both algorithms. Notably, FedProx appears to be more robust to the increased heterogeneity.
△ Less
Submitted 15 February, 2024;
originally announced February 2024.
-
Occupancy Grid Map to Pose Graph-based Map: Robust BIM-based 2D-LiDAR Localization for Lifelong Indoor Navigation in Changing and Dynamic Environments
Authors:
Miguel Arturo Vega Torres,
Alexander Braun,
André Borrmann
Abstract:
Several studies rely on the de facto standard Adaptive Monte Carlo Localization (AMCL) method to localize a robot in an Occupancy Grid Map (OGM) extracted from a building information model (BIM model). However, most of these studies assume that the BIM model precisely represents the real world, which is rarely true. Discrepancies between the reference BIM model and the real world (Scan-BIM deviati…
▽ More
Several studies rely on the de facto standard Adaptive Monte Carlo Localization (AMCL) method to localize a robot in an Occupancy Grid Map (OGM) extracted from a building information model (BIM model). However, most of these studies assume that the BIM model precisely represents the real world, which is rarely true. Discrepancies between the reference BIM model and the real world (Scan-BIM deviations) are not only due to furniture or clutter but also the usual as-planned and as-built deviations that exist with any model created in the design phase. These deviations affect the accuracy of AMCL drastically. This paper proposes an open-source method to generate appropriate Pose Graph-based maps from BIM models for robust 2D-LiDAR localization in changing and dynamic environments. First, 2D OGMs are automatically generated from complex BIM models. These OGMs only represent structural elements allowing indoor autonomous robot navigation. Then, an efficient technique converts these 2D OGMs into Pose Graph-based maps enabling more accurate robot pose tracking. Finally, we leverage the different map representations for accurate, robust localization with a combination of state-of-the-art algorithms. Moreover, we provide a quantitative comparison of various state-of-the-art localization algorithms in three simulated scenarios with varying levels of Scan-BIM deviations and dynamic agents. More precisely, we compare two Particle Filter (PF) algorithms: AMCL and General Monte Carlo Localization (GMCL); and two Graph-based Localization (GBL) methods: Google's Cartographer and SLAM Toolbox, solving the global localization and pose tracking problems. The numerous experiments demonstrate that the proposed method contributes to a robust localization with an as-designed BIM model or a sparse OGM in changing and dynamic environments, outperforming the conventional AMCL in accuracy and robustness.
△ Less
Submitted 10 August, 2023;
originally announced August 2023.
-
Detection of Tomato Ripening Stages using Yolov3-tiny
Authors:
Gerardo Antonio Alvarez Hernández,
Juan Carlos Olguin,
Juan Irving Vasquez,
Abril Valeria Uriarte,
Maria Claudia Villicaña Torres
Abstract:
One of the most important agricultural products in Mexico is the tomato (Solanum lycopersicum), which occupies the 4th place national most produced product . Therefore, it is necessary to improve its production, building automatic detection system that detect, classify an keep tacks of the fruits is one way to archieve it. So, in this paper, we address the design of a computer vision system to det…
▽ More
One of the most important agricultural products in Mexico is the tomato (Solanum lycopersicum), which occupies the 4th place national most produced product . Therefore, it is necessary to improve its production, building automatic detection system that detect, classify an keep tacks of the fruits is one way to archieve it. So, in this paper, we address the design of a computer vision system to detect tomatoes at different ripening stages. To solve the problem, we use a neural network-based model for tomato classification and detection. Specifically, we use the YOLOv3-tiny model because it is one of the lightest current deep neural networks. To train it, we perform two grid searches testing several combinations of hyperparameters. Our experiments showed an f1-score of 90.0% in the localization and classification of ripening stages in a custom dataset.
△ Less
Submitted 31 January, 2023;
originally announced February 2023.
-
The Internet-of-Things Meets Business Process Management: A Manifesto
Authors:
Christian Janiesch,
Agnes Koschmider,
Massimo Mecella,
Barbara Weber,
Andrea Burattin,
Claudio Di Ciccio,
Giancarlo Fortino,
Avigdor Gal,
Udo Kannengiesser,
Francesco Leotta,
Felix Mannhardt,
Andrea Marrella,
Jan Mendling,
Andreas Oberweis,
Manfred Reichert,
Stefanie Rinderle-Ma,
Estefania Serral,
WenZhan Song,
Jianwen Su,
Victoria Torres,
Matthias Weidlich,
Mathias Weske,
Liang Zhang
Abstract:
The Internet of Things (IoT) refers to a network of connected devices collecting and exchanging data over the Internet. These things can be artificial or natural, and interact as autonomous agents forming a complex system. In turn, Business Process Management (BPM) was established to analyze, discover, design, implement, execute, monitor and evolve collaborative business processes within and acros…
▽ More
The Internet of Things (IoT) refers to a network of connected devices collecting and exchanging data over the Internet. These things can be artificial or natural, and interact as autonomous agents forming a complex system. In turn, Business Process Management (BPM) was established to analyze, discover, design, implement, execute, monitor and evolve collaborative business processes within and across organizations. While the IoT and BPM have been regarded as separate topics in research and practice, we strongly believe that the management of IoT applications will strongly benefit from BPM concepts, methods and technologies on the one hand; on the other one, the IoT poses challenges that will require enhancements and extensions of the current state-of-the-art in the BPM field. In this paper, we question to what extent these two paradigms can be combined and we discuss the emerging challenges.
△ Less
Submitted 28 October, 2020; v1 submitted 11 September, 2017;
originally announced September 2017.
-
Change Patterns for Model Creation: Investigating the Role of Nesting Depth
Authors:
Barbara Weber,
Jakob **gera,
Victoria Torres,
Manfred Reichert
Abstract:
Process model quality has been an area of considerable research efforts. In this context, the correctness-by-construction principle of change patterns offers a promising perspective. However, using change patterns for model creation imposes a more structured way of modeling. While the process of process modeling (PPM) based on change primitives has been investigated, little is known about this pro…
▽ More
Process model quality has been an area of considerable research efforts. In this context, the correctness-by-construction principle of change patterns offers a promising perspective. However, using change patterns for model creation imposes a more structured way of modeling. While the process of process modeling (PPM) based on change primitives has been investigated, little is known about this process based on change patterns and factors that impact the cognitive complexity of pattern usage. Insights from the field of cognitive psychology as well as observations from a pilot study suggest that the nesting depth of the model to be created has a significant impact on cognitive complexity. This paper proposes a research design to test the impact of nesting depth on the cognitive complexity of change pattern usage in an experiment.
△ Less
Submitted 11 November, 2015;
originally announced November 2015.
-
How Advanced Change Patterns Impact the Process of Process Modeling
Authors:
Barbara Weber,
Sarah Zeitelhofer,
Jakob **gera,
Victoria Torres,
Manfred Reichert
Abstract:
Process model quality has been an area of considerable research efforts. In this context, correctness-by-construction as enabled by change patterns provides promising perspectives. While the process of process modeling (PPM) based on change primitives has been thoroughly investigated, only little is known about the PPM based on change patterns. In particular, it is unclear what set of change patte…
▽ More
Process model quality has been an area of considerable research efforts. In this context, correctness-by-construction as enabled by change patterns provides promising perspectives. While the process of process modeling (PPM) based on change primitives has been thoroughly investigated, only little is known about the PPM based on change patterns. In particular, it is unclear what set of change patterns should be provided and how the available change pattern set impacts the PPM. To obtain a better understanding of the latter as well as the (subjective) perceptions of process modelers, the arising challenges, and the pros and cons of different change pattern sets we conduct a controlled experiment. Our results indicate that process modelers face similar challenges irrespective of the used change pattern set (core pattern set versus extended pattern set, which adds two advanced change patterns to the core patterns set). An extended change pattern set, however, is perceived as more difficult to use, yielding a higher mental effort. Moreover, our results indicate that more advanced patterns were only used to a limited extent and frequently applied incorrectly, thus, lowering the potential benefits of an extended pattern set.
△ Less
Submitted 11 November, 2015;
originally announced November 2015.
-
Change Patterns in Use: A Critical Evaluation
Authors:
Barbara Weber,
Jakob **gera,
Victoria Torres,
Manfred Reichert
Abstract:
Process model quality has been an area of considerable research efforts. In this context, the correctness-by-construction principle of change patterns provides promising perspectives. However, using change patterns for model creation imposes a more structured way of modeling. While the process of process modeling (PPM) based on change primitives has been investigated, little is known about this pr…
▽ More
Process model quality has been an area of considerable research efforts. In this context, the correctness-by-construction principle of change patterns provides promising perspectives. However, using change patterns for model creation imposes a more structured way of modeling. While the process of process modeling (PPM) based on change primitives has been investigated, little is known about this process based on change patterns. To obtain a better understanding of the PPM when using change patterns, the arising challenges, and the subjective perceptions of process designers, we conduct an exploratory study. The results indicate that process designers face little problems as long as control-flow is simple, but have considerable problems with the usage of change patterns when complex, nested models have to be created. Finally, we outline how effective tool support for change patterns should be realized.
△ Less
Submitted 11 November, 2015;
originally announced November 2015.
-
How to Apply Assignment Methods that were Developed for Vehicular Traffic to Pedestrian Microsimulations
Authors:
Vidal Roca,
Vicente Torres,
Tobias Kretz,
Karsten Lehmann,
Ingmar Hofsäß
Abstract:
Applying assignment methods to compute user-equilibrium route choice is very common in traffic planning. It is common sense that vehicular traffic arranges in a user-equilibrium based on generalized costs in which travel time is a major factor. Surprisingly travel time has not received much attention for the route choice of pedestrians. In microscopic simulations of pedestrians the vastly dominati…
▽ More
Applying assignment methods to compute user-equilibrium route choice is very common in traffic planning. It is common sense that vehicular traffic arranges in a user-equilibrium based on generalized costs in which travel time is a major factor. Surprisingly travel time has not received much attention for the route choice of pedestrians. In microscopic simulations of pedestrians the vastly dominating paradigm for the computation of the preferred walking direction is set into the direction of the (spatially) shortest path. For situations where pedestrians have travel time as primary determinant for their walking behavior it would be desirable to also have an assignment method in pedestrian simulations. To apply existing (road traffic) assignment methods with simulations of pedestrians one has to reduce the nondenumerably many possible pedestrian trajectories to a small subset of routes which represent the main, relevant, and significantly distinguished routing alternatives. All except one of these routes will mark detours, i.e. not the shortest connection between origin and destination. The proposed assignment method is intended to work with common operational models of pedestrian dynamics. These - as mentioned before - usually send pedestrians into the direction of the spatially shortest path. Thus, all detouring routes have to be equipped with intermediate destinations, such that pedestrians can do a detour as a piecewise connection of segments on which they walk into the direction of the shortest path. One has then to take care that the transgression from one segment to the following one no artifacts are introduced into the pedestrian trajectory.
△ Less
Submitted 7 February, 2014;
originally announced February 2014.
-
Development of Biofeedback Mechanisms in a Procedural Environment Using Biometric Sensors
Authors:
Vasco Pereira Torres
Abstract:
Before the computer age, games were played in the physical world where players would have to interact with real objects and each other, triggering a series of emotions. Nowadays, the computer games have become one of the most popular forms of entertainment due to their high-level of attraction and accessibility. However, the game industry is always trying to find new ways of making games more inte…
▽ More
Before the computer age, games were played in the physical world where players would have to interact with real objects and each other, triggering a series of emotions. Nowadays, the computer games have become one of the most popular forms of entertainment due to their high-level of attraction and accessibility. However, the game industry is always trying to find new ways of making games more interactive and exciting in order to attract new players, and one of the recent trends on the area of human-computer interaction is Biofeedback.
The goal of this dissertation is to study different approaches on the use of indirect biofeedback within videogames, with the purpose of creating a better human-computer interaction, and provide a more appealing and immersive user experience. For this, we focused on the development of a framework capable of testing different indirect biofeedback models within a specified game, in order to assess the effect of each of these variations on the user experience. This framework is game independent, with the intention of being used on further studies.
△ Less
Submitted 8 October, 2013;
originally announced October 2013.
-
BP Variability Case Studies Development using different Modeling Approaches
Authors:
Clara Ayora,
Victoria Torres,
Vicente Pelechano
Abstract:
Variability in Business Process modeling has already been faced by different authors from the literature. Depending on the context in which each author faces the modeling problem, we find different approaches (C-EPC, C-YAWL, FEATURE-EPC, PESOA, PROVOP, or WORKLETS). In this report we present four of the most representative approaches (C-EPC, PESOA, PROVOP and WORKLETS) which are presented by means…
▽ More
Variability in Business Process modeling has already been faced by different authors from the literature. Depending on the context in which each author faces the modeling problem, we find different approaches (C-EPC, C-YAWL, FEATURE-EPC, PESOA, PROVOP, or WORKLETS). In this report we present four of the most representative approaches (C-EPC, PESOA, PROVOP and WORKLETS) which are presented by means of the different case studies found in the literature.
△ Less
Submitted 28 February, 2011; v1 submitted 22 February, 2011;
originally announced February 2011.
