Edge devices rely extensively on machinelearning for intelligent inferences and pattern matching. However, edge devices use a multitude of sensing modalities and are exposed to wide ranging contexts. It is difficult ...
详细信息
ISBN:
(纸本)9781450358378
Edge devices rely extensively on machinelearning for intelligent inferences and pattern matching. However, edge devices use a multitude of sensing modalities and are exposed to wide ranging contexts. It is difficult to develop separate machinelearning models for each scenario as manual labeling is not scalable. To reduce the amount of labeled data and to speed up the training process, we propose to transfer knowledge between edge devices by using unlabeled data. Our approach, called RecycleML, uses cross modal transfer to accelerate the learning of edge devices across different sensing modalities. Using human activity recognition as a case study, over our collected CMActivity dataset, we observe that RecycleML reduces the amount of required labeled data by at least 90% and speeds up the training process by up to 50 times in comparison to training the edge device from scratch.
Two major journals in mechanical engineering field put forward a special note for author(s) of articles discussing structural health monitoring. Only contributing new classifiers or new features without applying engin...
详细信息
ISBN:
(纸本)9781538694220
Two major journals in mechanical engineering field put forward a special note for author(s) of articles discussing structural health monitoring. Only contributing new classifiers or new features without applying engineering knowledge or principle is considered insufficient. This article intends to provide empirical evidence to support the decision. This work shows the classification accuracy of a machinelearning approach strongly depends on the number of training data. A high accuracy can simply be obtained by training the classifier with a large dataset, which is hardly realizable in practice. Meanwhile, the features and classifier that based on a sound engineering principle can provide a level of classification accuracy independent of the number of training data.
The proceedings contain 8 papers. The topics discussed include: multiple evaluation in future population distribution for sustainable city;detecting street signs in cities based on object recognition with machine lear...
ISBN:
(纸本)9781450360395
The proceedings contain 8 papers. The topics discussed include: multiple evaluation in future population distribution for sustainable city;detecting street signs in cities based on object recognition with machinelearning and GIS spatial analysis;D-record: disaster response and relief coordination pipeline;analysis, integration and visualization of urban data from multiple heterogeneous sources;SCOUTS: a smart community centric urban heat monitoring framework;real-time traffic light detection from videos with inertial sensor fusion;and an integrated visual analytics framework for spatiotemporal data.
In this paper we evaluate a new Estimation of Distribution Algorithm (EDA) constructed on top of a very successful Bayesian network learning procedure, Max-Min Hill-Climbing (MMHC). The aim of this paper is to check w...
详细信息
ISBN:
(纸本)9783030011321;9783030011314
In this paper we evaluate a new Estimation of Distribution Algorithm (EDA) constructed on top of a very successful Bayesian network learning procedure, Max-Min Hill-Climbing (MMHC). The aim of this paper is to check whether the excellent properties reported for this algorithm in machinelearning papers, have some impact on the efficiency and efficacy of EDA based optimization. Our experiments show that the proposed algorithm outperform wellknown state of the art EDA like BOA and EBNA in a test bed based on B-functions. On the basis of these results we conclude that the proposed scheme is a promising candidate for challenging real-world applications, specifically, problems related to the areas of datamining, Patter recognition and Artificial Intelligence.
Image recognition is a major area of application of machinelearning - evolving at a rapid pace with a number of programming platforms available to developers. While each platform has its own uniqueness, the methodolo...
详细信息
ISBN:
(纸本)9781538692097
Image recognition is a major area of application of machinelearning - evolving at a rapid pace with a number of programming platforms available to developers. While each platform has its own uniqueness, the methodology of image recognition consists of a sequence of image processing tasks, development of a classifier algorithm, training and testing followed by deployment. This tutorial will delve into the programming aspects of image processing including thresholding, contouring and template matching. In order to provide practical hands on programming this tutorial will closely look at three real life applications of image patternrecognition namely ALPR using Tesseract OCR and will touch upon using CNN for character detection. The tutorial will explain the algorithm, implementation of pseudocode through Python using two major platforms: OpenCV and Tensorflow.
Knowledge Engineering is key to enable knowledge extraction, representation and reasoning, leading to better business insights and decisions. Current advances in machinelearning and new trends in AI are bringing a pl...
详细信息
ISBN:
(纸本)9781538692097
Knowledge Engineering is key to enable knowledge extraction, representation and reasoning, leading to better business insights and decisions. Current advances in machinelearning and new trends in AI are bringing a plethora of algorithms capable of performing advanced patternrecognition and data classification. The ability to link, to organize and to query the outputs of these algorithms as well as the ability to handle huge amounts of data and its multiple sources is crucial to maximize the potential of such advances, specially over large datasets. This paper presents challenges in the context of digital agriculture and our position in moving forward with these capabilities whilst using knowledge engineering techniques.
Deep learning is nowadays a buzzword and is considered a new era of machinelearning which trains the computers in finding the pattern from a massive amount of data. It mainly describes the learning at multiple levels...
详细信息
The proceedings contain 16 papers. The topics discussed include: software product lines to support language rehabilitation therapies: an experience report;formalistic modeling based on patternrecognition applied to t...
ISBN:
(纸本)9781538694596
The proceedings contain 16 papers. The topics discussed include: software product lines to support language rehabilitation therapies: an experience report;formalistic modeling based on patternrecognition applied to the knowledge and human talent sector in Ecuador;relaxation state induction through binaural acoustic stimuli;machinelearning techniques for PM10 levels forecast in Bogotáanalysis of the tertiary sector of Colombia and demand estimation using LEAP;and economic viability of hydrogen generators in transportation vehicles.
Drones are conventionally controlled using joysticks, remote controllers, mobile applications, and embedded computers. A few significant issues with these approaches are that drone control is limited by the range of e...
详细信息
ISBN:
(纸本)9781538657621
Drones are conventionally controlled using joysticks, remote controllers, mobile applications, and embedded computers. A few significant issues with these approaches are that drone control is limited by the range of electromagnetic radiation and susceptible to interference noise. In this study we propose the use of hand gestures as a method to control drones. We investigate the use of computer vision methods to develop an intuitive way of agent-less communication between a drone and its operator. Computer vision-based methods rely on the ability of a drones camera to capture surrounding images and use patternrecognition to translate images to meaningful and/or actionable information. The proposed framework involves a few key parts toward an ultimate action to be taken. They are: image segregation from the video streams of front camera, creating a robust and reliable image recognition based on segregated images, and finally conversion of classified gestures into actionable drone movement, such as takeoff, landing, hovering and so forth. A set of five gestures are studied in this work. Haar feature-based AdaBoost classifier [1] is employed for gesture recognition. We also envisage safety of the operator and drone's action calculating the distance based on computer vision for this task. A series of experiments are conducted to measure gesture recognition accuracies considering the major scene variabilities, illumination, background, and distance. Classification accuracies show that well-lit, clear background, and within 3 ft gestures are recognized correctly over 90%. Limitations of current framework and feasible solutions for better gesture recognition are discussed, too. The software library we developed, and hand gesture datasets are open-sourced at project website.
Inferring causal relationships in observational data is crucial for understanding scientific and social processes. We develop the firststatistical machinelearning approach for automatically discovering regression di...
详细信息
ISBN:
(纸本)9781450355520
Inferring causal relationships in observational data is crucial for understanding scientific and social processes. We develop the firststatistical machinelearning approach for automatically discovering regression discontinuity designs (RDDs), a quasi-experimental setup often used in econometrics. Our method identifies interpretable, localized RDDs in arbitrary dimensional data and can seamlessly compute treatment effects without expert supervision. By applying the technique to a variety of synthetic and real datasets, we demonstrate robust performance under adverse conditions including unobserved variables, substantial noise, and model misspecification.
暂无评论