检索结果-内蒙古大学图书馆

A Unified Approach Towards Active learning and Out-of-Distribution Detection

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Schmidt, Sebastian Schenk, Leonard Schwinn, Leo Günnemann, Stephan BMW Group Munich Germany Technical University of Munich Data Analytics and Machine Learning Group Munich Germany

When applying deep learning models in open-world scenarios, active learning (AL) strategies are crucial for identifying label candidates from a nearly infinite amount of unlabeled data. In this context, robust out-of-distribution (OOD) detection mechanisms are essential for handling data outside the target distribution of the application. However, current works investigate both problems separately. In this work, we introduce SISOM as the first unified solution for both AL and OOD detection. By leveraging feature space distance metrics SISOM combines the strengths of the currently independent tasks to solve both effectively. We conduct extensive experiments showing the problems arising when migrating between both tasks. In these evaluations SISOM underlined its effectiveness by achieving first place in two of the widely used OpenOOD benchmarks and second place in the remaining one. In AL, SISOM outperforms others and delivers top-1 performance in three benchmarks. Copyright © 2024, The Authors. All rights reserved.

关键词： data handling

Stream-based Active learning by Exploiting Temporal Properties in Perception with Temporal Predicted Loss

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Schmidt, Sebastian Günnemann, Stephan BMW Group Munich Germany Technical University of Munich Data Analytics and Machine Learning Group Munich Germany

Active learning (AL) reduces the amount of labeled data needed to train a machine learning model by intelligently choosing which instances to label. Classic pool-based AL requires all data to be present in a datacenter, which can be challenging with the increasing amounts of data needed in deep learning. However, AL on mobile devices and robots, like autonomous cars, can filter the data from perception sensor streams before reaching the datacenter. We exploited the temporal properties for such image streams in our work and proposed the novel temporal predicted loss (TPL)1 method. To evaluate the stream-based setting properly, we introduced the GTA V streets and the A2D2 streets dataset and made both publicly available. Our experiments showed that our approach significantly improves the diversity of the selection while being an uncertainty-based method. As pool-based approaches are more common in perception applications, we derived a concept for comparing pool-based and stream-based AL, where TPL outperformed state-of-the-art pool- or stream-based approaches for different models. TPL demonstrated a gain of 2.5 precept points (pp) less required data while being significantly faster than pool-based methods. Copyright © 2023, The Authors. All rights reserved.

关键词： Deep learning

Design and Implementation of a data Governance Framework and Platform: A Case Study of a National Research Organization of Thailand 20

学校读者我要写书评

暂无评论

Design and Implementation of a Data Governance Framework and...

20th International Joint Conference on Computer Science and Software Engineering, JCSSE 2023

作者： Chanyachatchawan, Sapa Nasingkun, Krich Tumsangthong, Patipat Chata, Porntiwa Buranarach, Marut Socharoentum, Monsak National Electronics and Computer Technology Center Leveraging Technology Solutions Section Bangkok Thailand National Electronics and Computer Technology Center Strategic Analytics Networks with Machine Learning and Ai Research Bangkok Thailand National Electronics and Computer Technology Center Data Science and Analytics Research Group Bangkok Thailand Digital Government Development Agency Bangkok Thailand

ISBN: (纸本)9798350300505

In the current era of extensive data usage across industries, data collection, preservation, utilization, and organization has become more challenging and nuanced because it is necessary to consider critical concerns such as data security, privacy, and legal issues, apart from efficiency issues. As a result, Thai government initiated the idea and effort to implement data governance throughout the government agency. This paper showcases the implementation of data governance in a governmental research organization with highly diverse structured and unstructured data. The implementation follows international standards and the guidelines of the Digital Government Development Agency (DGA). The executives set up the working body, including the data Governance Council and data Stewards, responsible for setting up and deploying policies and regulations. Creating awareness and the necessary infrastructure are the main focuses in the first-year phase. The metadata was designed to extend DGA's version and match the organization's unique requirements. A data catalog platform was developed accordingly. We organized activities to boost employee awareness and participation, including advertising and data catalog platform training. By the end of the first year of implementation, every organization unit had registered at least one data record into the data catalog. © 2023 IEEE.

关键词： Metadata

Predicting Influential Higher-Order Patterns in Temporal Network data 14

学校读者我要写书评

暂无评论

Predicting Influential Higher-Order Patterns in Temporal Net...

14th IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, ASONAM 2022

作者： Gote, Christoph Perri, Vincenzo Scholtes, Ingo Data Analytics Group University of Zurich Zurich Switzerland Eth Zurich Systems Design Switzerland Julius-Maximilians-Universität Würzburg Chair of Machine Learning for Complex Networks Würzburg Germany

ISBN: (纸本)9781665456616

Networks are frequently used to model complex systems comprised of interacting elements. While edges capture the topology of direct interactions, the true complexity of many systems originates from higher-order patterns in paths by which nodes can indirectly influence each other. Path data, representing ordered sequences of consecutive direct interactions, can be used to model these patterns. On the one hand, to avoid overfitting, such models should only consider those higher-order patterns for which the data provide sufficient statistical evidence. On the other hand, we hypothesise that network models, which capture only direct interactions, underfit higher-order patterns present in data. Consequently, both approaches are likely to misidentify influential nodes in complex networks. We contribute to this issue by proposing five centrality measures based on MOGen, a multi-order generative model that accounts for all indirect influences up to a maximum distance but disregards influences at higher distances. We compare MOGen-based centralities to equivalent measures for network models and path data in a prediction experiment where we aim to identify influential nodes in out-of-sample data. Our results show strong evidence supporting our hypothesis. MOGen consistently outperforms both the network model and path-based prediction. We further show that the performance difference between MOGen and the path-based approach disappears if we have sufficient observations, confirming that the error is due to overfitting. © 2022 IEEE.

关键词： Forecasting

p-value adjustment for monotonous, unbiased, and fast clustering comparison 23

学校读者我要写书评

暂无评论

p-value adjustment for monotonous, unbiased, and fast cluste...

Proceedings of the 37th International Conference on Neural Information Processing Systems

作者： Kai Klede Thomas Altstidl Dario Zanca Björn Eskofier Machine Learning and Data Analytics (MaD) Lab Friedrich-Alexander Universität Erlangen-Nürnberg Machine Learning and Data Analytics (MaD) Lab Friedrich-Alexander Universität Erlangen-Nürnberg and Translational Digital Health Group Institute of AI for Health Helmholtz Zentrum München

Popular metrics for clustering comparison, like the Adjusted Rand Index and the Adjusted Mutual Information, are type II biased. The Standardized Mutual Information removes this bias but suffers from counterintuitive non-monotonicity and poor computational efficiency. We introduce the p-value adjusted Rand Index (PMI2), the first cluster comparison method that is type II unbiased and provably monotonous. The PMI2 has fast approximations that outperform the Standardized Mutual information. We demonstrate its unbiased clustering selection, approximation quality, and runtime efficiency on synthetic benchmarks. In experiments on image and social network datasets, we show how the PMI2 can help practitioners choose better clustering and community detection algorithms.

关键词：

Beyond the Known: Adversarial Autoencoders in Novelty Detection

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Asad, Muhammad Ullah, Ihsan Sistu, Ganesh Madden, Michael G. Machine Learning Research Group School of Computer Science University of Galway Ireland Insight SFI Research Centre for Data Analytics University of Galway Ireland Valeo Vision Systems Tuam Ireland

In novelty detection, the goal is to decide if a new data point should be categorized as an inlier or an outlier, given a training dataset that primarily captures the inlier distribution. Recent approaches typically use deep encoder and decoder network frameworks to derive a reconstruction error, and employ this error either to determine a novelty score, or as the basis for a one-class classifier. In this research, we use a similar framework but with a lightweight deep network, and we adopt a probabilistic score with reconstruction error. Our methodology calculates the probability of whether the sample comes from the inlier distribution or not. This work makes two key contributions. The first is that we compute the novelty probability by linearizing the manifold that holds the structure of the inlier distribution. This allows us to interpret how the probability is distributed and can be determined in relation to the local coordinates of the manifold tangent space. The second contribution is that we improve the training protocol for the network. Our results indicate that our approach is effective at learning the target class, and it outperforms recent state-of-the-art methods on several benchmark datasets. Copyright © 2024, The Authors. All rights reserved.

关键词： Anomaly detection

Depth Estimation using Weighted-loss and Transfer learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Hafeez, Muhammad Adeel Madden, Michael G. Sistu, Ganesh Ullah, Ihsan Machine Learning Research Group School of Computer Science University of Galway Ireland Insight SFI Research Centre for Data Analytics University of Galway Ireland Valeo Vision Systems Tuam Ireland

Depth estimation from 2D images is a common computer vision task that has applications in many fields including autonomous vehicles, scene understanding and robotics. The accuracy of a supervised depth estimation method mainly relies on the chosen loss function, the model architecture, quality of data and performance metrics. In this study, we propose a simplified and adaptable approach to improve depth estimation accuracy using transfer learning and an optimized loss function. The optimized loss function is a combination of weighted losses to which enhance robustness and generalization: Mean Absolute Error (MAE), Edge Loss and Structural Similarity Index (SSIM). We use a grid search and a random search method to find optimized weights for the losses, which leads to an improved model. We explore multiple encoder-decoder-based models including DenseNet121, DenseNet169, DenseNet201, and EfficientNet for the supervised depth estimation model on NYU Depth dataset v2. We observe that the EfficientNet model, pre-trained on ImageNet for classification when used as an encoder, with a simple upsampling decoder, gives the best results in terms of RSME, REL and log10: 0.386, 0.113 and 0.049, respectively. We also perform a qualitative analysis which illustrates that our model produces depth maps that closely resemble ground truth, even in cases where the ground truth is flawed. The results indicate significant improvements in accuracy and robustness, with EfficientNet being the most successful architecture. Copyright © 2024, The Authors. All rights reserved.

关键词： Signal encoding

Deep learning algorithms to classify Fitzpatrick skin types for smartphone based NIRS imaging device 15

学校读者我要写书评

暂无评论

Deep learning algorithms to classify Fitzpatrick skin types ...

Next-Generation Spectroscopic Technologies XV 2023

作者： Leizaola, Daniela Sobhan, Masrur Kaile, Kacie Mondal, Ananda Mohan Godavarty, Anuradha Optical Imaging Laboratory Department of Biomedical Engineering Florida International University 10555 W Flagler St MiamiFL33174 United States Machine Learning and Data Analytics Group Knight Foundation School Computing and Information Sciences Florida International University 11200 SW 8th St MiamiFL33199 United States

ISBN: (数字)9781510661479

ISBN: (纸本)9781510661462

Non-contact imaging modalities for monitoring wound health could supplement the current standard which is a visual inspection by clinicians. Recently, a smartphone oxygenation tool (SPOT) has been developed for physiological imaging of tissue oxygenation changes in response to treatment. However, upon visual inspection of the wound bed and surrounding area, there are variations of pigmentation. Melanin concentration is a highly absorbing chromophore that can impact spatial oxygenation measurements. The objective of this study was to classify the six Fitzpatrick Skin Types (FST) by applying deep learning techniques prior to correcting SPOT’s oxygenation maps. In this IRB-approved study, control subjects were imaged on seven skin locations of varying FST (I-VI) under three different lighting conditions using SPOT device’s camera. A benchmark dataset with samples of 28 × 28 pixel images of human subjects’ feet was developed in three color spaces (R-G-B, Y-Cb-Cr, L-a-b). A deep learning algorithm, specifically a convolutional neural network (CNN), was used to classify skin into six FST (classes). Preliminary results showed the skin types on control subjects’ feet could be classified using deep learning with hyperparameter tuning with accuracies of > 82%. Our ongoing efforts are focused on extensive in-vivo studies on control subjects of FST I-VI on feet towards future implementation of the technology for diabetic wounds and oxygenation mapping using the SPOT device. © 2023 SPIE.

关键词： Smartphones

A Two-Stage Minimum Cost Multicut Approach to Self-supervised Multiple Person Tracking 15th

学校读者我要写书评

暂无评论

A Two-Stage Minimum Cost Multicut Approach to Self-supervise...

15th Asian Conference on Computer Vision, ACCV 2020

作者： Ho, Kalun Kardoost, Amirhossein Pfreundt, Franz-Josef Keuper, Janis Keuper, Margret Fraunhofer Center Machine Learning Sankt Augustin Germany CC-HPC Fraunhofer ITWM Kaiserslautern Germany Data and Web Science Group University of Mannheim Mannheim Germany Institute for Machine Learning and Analytics Offenburg University Offenburg Germany

ISBN: (纸本)9783030695316

Multiple Object Tracking (MOT) is a long-standing task in computer vision. Current approaches based on the tracking by detection paradigm either require some sort of domain knowledge or supervision to associate data correctly into tracks. In this work, we present a self-supervised multiple object tracking approach based on visual features and minimum cost lifted multicuts. Our method is based on straight-forward spatio-temporal cues that can be extracted from neighboring frames in an image sequences without supervision. Clustering based on these cues enables us to learn the required appearance invariances for the tracking task at hand and train an AutoEncoder to generate suitable latent representations. Thus, the resulting latent representations can serve as robust appearance cues for tracking even over large temporal distances where no reliable spatio-temporal features can be extracted. We show that, despite being trained without using the provided annotations, our model provides competitive results on the challenging MOT Benchmark for pedestrian tracking. © 2021, Springer Nature Switzerland AG.

关键词： Computer vision