scholarly journals Reduced Simulation: Real-to-Sim Approach toward Collision Detection in Narrowly Confined Environments

Robotics ◽  
2021 ◽  
Vol 10 (4) ◽  
pp. 131
Author(s):  
Yusuke Takayama ◽  
Photchara Ratsamee ◽  
Tomohiro Mashita
Keyword(s):  
Deep Learning ◽  
Real World ◽  
Collision Detection ◽  
User Study ◽  
Cost Effective ◽  
Experimental Result ◽  
Simulation Design ◽  
High Quality ◽  
Simulated Environments ◽  
Simulated Environment

Recently, several deep-learning based navigation methods have been achieved because of a high quality dataset collected from high-quality simulated environments. However, the cost of creating high-quality simulated environments is high. In this paper, we present a concept of the reduced simulation, which can serve as a simplified version of a simulated environment yet be efficient enough for training deep-learning based UAV collision avoidance approaches. Our approach deals with the reality gap between a reduced simulation dataset and real world dataset and can provide a clear guideline for reduced simulation design. Our experimental result confirmed that the reduction in visual features provided by textures and lighting does not affect operating performance with the user study. Moreover, by conducting collision detection experiments, we verified that our reduced simulation outperforms the conventional cost-effective simulations in adaptation capability with respect to realistic simulation and real-world scenario.

Understanding the complementary nature of paid and volunteer crowds for content creation

2020 ◽  
pp. 37
Author(s):  
Claudia Flores-Saviaga ◽  
Ricardo Granados ◽  
Liliana Savage ◽  
Lizbeth Escobedo ◽  
Saiph Savage
Keyword(s):  
Real World ◽  
Limited Resource ◽  
Cost Effective ◽  
High Quality ◽  
Power Cost ◽  
Non Profit ◽  
Content Creation ◽  
Online Study ◽  
Complementary Nature ◽  
Different Populations

Crowdsourced content creation like articles or slogans can be powered by crowds of volunteers or workers from paid task markets. Volunteers often have expertise and are intrinsically motivated, but are a limited resource, and are not always reliably available. On the other hand, paid crowd workers are reliably available, can be guided to produce high-quality content, but cost money. How can these different populations of crowd workers be leveraged together to power cost-effective yet high-quality crowd-powered content-creation systems? To answer this question, we need to understand the strengths and weaknesses of each. We conducted an online study where we hired paid crowd workers and recruited volunteers from social media to complete three content creation tasks for three real-world non-profit organizations that focus on empowering women. These tasks ranged in complexity from simply generating keywords or slogans to creating a draft biographical article. Our results show that paid crowds completed work and structured content following editorial guidelines more effectively. However, volunteer crowds provide content that is more original. Based on the findings, we suggest that crowd-powered content-creation systems could gain the best of both worlds by leveraging volunteers to scaffold the direction that original content should take; while having paid crowd workers structure content and prepare it for real world use.

NeckFace

2021 ◽  
Vol 5 (2) ◽  
pp. 1-31
Author(s):  
Tuochao Chen ◽  
Yaxuan Li ◽  
Songyun Tao ◽  
Hyunchul Lim ◽  
Mose Sakashita ◽  
...  
Keyword(s):  
Deep Learning ◽  
Facial Expressions ◽  
Real World ◽  
User Study ◽  
Form Factors ◽  
Physical Activities ◽  
Sensing Technology ◽  
Study Results ◽  
Real World Applications ◽  
Challenges And Opportunities

Facial expressions are highly informative for computers to understand and interpret a person's mental and physical activities. However, continuously tracking facial expressions, especially when the user is in motion, is challenging. This paper presents NeckFace, a wearable sensing technology that can continuously track the full facial expressions using a neck-piece embedded with infrared (IR) cameras. A customized deep learning pipeline called NeckNet based on Resnet34 is developed to learn the captured infrared (IR) images of the chin and face and output 52 parameters representing the facial expressions. We demonstrated NeckFace on two common neck-mounted form factors: a necklace and a neckband (e.g., neck-mounted headphones), which was evaluated in a user study with 13 participants. The study results showed that NeckFace worked well when the participants were sitting, walking, or after remounting the device. We discuss the challenges and opportunities of using NeckFace in real-world applications.

Real–Sim–Real Transfer for Real-World Robot Control Policy Learning with Deep Reinforcement Learning

2020 ◽  
Vol 10 (5) ◽  
pp. 1555
Author(s):  
Naijun Liu ◽  
Yinghao Cai ◽  
Tao Lu ◽  
Rui Wang ◽  
Shuo Wang
Keyword(s):  
Reinforcement Learning ◽  
Real World ◽  
Robot Control ◽  
Training Phase ◽  
Control Policies ◽  
Promising Alternative ◽  
The Real ◽  
Training Costs ◽  
Simulated Environments ◽  
Simulated Environment

Compared to traditional data-driven learning methods, recently developed deep reinforcement learning (DRL) approaches can be employed to train robot agents to obtain control policies with appealing performance. However, learning control policies for real-world robots through DRL is costly and cumbersome. A promising alternative is to train policies in simulated environments and transfer the learned policies to real-world scenarios. Unfortunately, due to the reality gap between simulated and real-world environments, the policies learned in simulated environments often cannot be generalized well to the real world. Bridging the reality gap is still a challenging problem. In this paper, we propose a novel real–sim–real (RSR) transfer method that includes a real-to-sim training phase and a sim-to-real inference phase. In the real-to-sim training phase, a task-relevant simulated environment is constructed based on semantic information of the real-world scenario and coordinate transformation, and then a policy is trained with the DRL method in the built simulated environment. In the sim-to-real inference phase, the learned policy is directly applied to control the robot in real-world scenarios without any real-world data. Experimental results in two different robot control tasks show that the proposed RSR method can train skill policies with high generalization performance and significantly low training costs.

scholarly journals On the Reproducibility and Replicability of Deep Learning in Software Engineering

2022 ◽  
Vol 31 (1) ◽  
pp. 1-46
Author(s):  
Chao Liu ◽  
Cuiyun Gao ◽  
Xin Xia ◽  
David Lo ◽  
John Grundy ◽  
...  
Keyword(s):  
Deep Learning ◽  
Software Engineering ◽  
Source Code ◽  
Experimental Results ◽  
Supervised Machine Learning ◽  
Optimization Process ◽  
Experimental Result ◽  
Experimental Setup ◽  
High Quality ◽  
Two Factors

Context: Deep learning (DL) techniques have gained significant popularity among software engineering (SE) researchers in recent years. This is because they can often solve many SE challenges without enormous manual feature engineering effort and complex domain knowledge. Objective: Although many DL studies have reported substantial advantages over other state-of-the-art models on effectiveness, they often ignore two factors: (1) reproducibility —whether the reported experimental results can be obtained by other researchers using authors’ artifacts (i.e., source code and datasets) with the same experimental setup; and (2) replicability —whether the reported experimental result can be obtained by other researchers using their re-implemented artifacts with a different experimental setup. We observed that DL studies commonly overlook these two factors and declare them as minor threats or leave them for future work. This is mainly due to high model complexity with many manually set parameters and the time-consuming optimization process, unlike classical supervised machine learning (ML) methods (e.g., random forest). This study aims to investigate the urgency and importance of reproducibility and replicability for DL studies on SE tasks. Method: In this study, we conducted a literature review on 147 DL studies recently published in 20 SE venues and 20 AI (Artificial Intelligence) venues to investigate these issues. We also re-ran four representative DL models in SE to investigate important factors that may strongly affect the reproducibility and replicability of a study. Results: Our statistics show the urgency of investigating these two factors in SE, where only 10.2% of the studies investigate any research question to show that their models can address at least one issue of replicability and/or reproducibility. More than 62.6% of the studies do not even share high-quality source code or complete data to support the reproducibility of their complex models. Meanwhile, our experimental results show the importance of reproducibility and replicability, where the reported performance of a DL model could not be reproduced for an unstable optimization process. Replicability could be substantially compromised if the model training is not convergent, or if performance is sensitive to the size of vocabulary and testing data. Conclusion: It is urgent for the SE community to provide a long-lasting link to a high-quality reproduction package, enhance DL-based solution stability and convergence, and avoid performance sensitivity on different sampled data.

Weakly Supervised Spatial Deep Learning for Earth Image Segmentation Based on Imperfect Polyline Labels

2022 ◽  
Vol 13 (2) ◽  
pp. 1-20
Author(s):  
Zhe Jiang ◽  
Wenchong He ◽  
Marcus Stephen Kirby ◽  
Arpan Man Sainju ◽  
Shaowen Wang ◽  
...  
Keyword(s):  
Image Segmentation ◽  
Deep Learning ◽  
Real World ◽  
Large Scale ◽  
Model Parameters ◽  
Vector Representation ◽  
High Quality ◽  
Geometric Properties ◽  
Location Errors ◽  
Weakly Supervised

In recent years, deep learning has achieved tremendous success in image segmentation for computer vision applications. The performance of these models heavily relies on the availability of large-scale high-quality training labels (e.g., PASCAL VOC 2012). Unfortunately, such large-scale high-quality training data are often unavailable in many real-world spatial or spatiotemporal problems in earth science and remote sensing (e.g., mapping the nationwide river streams for water resource management). Although extensive efforts have been made to reduce the reliance on labeled data (e.g., semi-supervised or unsupervised learning, few-shot learning), the complex nature of geographic data such as spatial heterogeneity still requires sufficient training labels when transferring a pre-trained model from one region to another. On the other hand, it is often much easier to collect lower-quality training labels with imperfect alignment with earth imagery pixels (e.g., through interpreting coarse imagery by non-expert volunteers). However, directly training a deep neural network on imperfect labels with geometric annotation errors could significantly impact model performance. Existing research that overcomes imperfect training labels either focuses on errors in label class semantics or characterizes label location errors at the pixel level. These methods do not fully incorporate the geometric properties of label location errors in the vector representation. To fill the gap, this article proposes a weakly supervised learning framework to simultaneously update deep learning model parameters and infer hidden true vector label locations. Specifically, we model label location errors in the vector representation to partially reserve geometric properties (e.g., spatial contiguity within line segments). Evaluations on real-world datasets in the National Hydrography Dataset (NHD) refinement application illustrate that the proposed framework outperforms baseline methods in classification accuracy.

The Formability and Elongation of Aluminum Alloys AA5083 and AA3003 for Micro-Truss Sandwiches Manufacturing

2020 ◽  
Vol 38 (9A) ◽  
pp. 1396-1405
Author(s):  
Arwa F. Tawfeeq ◽  
Matthew R. Barnett
Keyword(s):  
Aluminum Alloy ◽  
Strain Rate ◽  
Cost Effective ◽  
Tensile Tests ◽  
Truss Structures ◽  
High Quality ◽  
Trade Off ◽  
Clad Layer ◽  
Higher Temperature ◽  
Different Shapes

The development in the manufacturing of micro-truss structures has demonstrated the effectiveness of brazing for assembling these sandwiches, which opens new opportunities for cost-effective and high-quality truss manufacturing. An evolving idea in micro-truss manufacturing is the possibility of forming these structures in different shapes with the aid of elevated temperature. This work investigates the formability and elongation of aluminum alloy sheets typically used for micro-truss manufacturing, namely AA5083 and AA3003. Tensile tests were performed at a temperature in the range of 25-500 ○C and strain rate in the range of 2x10-4 -10-2 s-1. The results showed that the clad layer in AA3003 exhibited an insignificant effect on the formability and elongation of AA3003. The formability of the two alloys was improved significantly with values of m as high as 0.4 and 0.13 for AA5083 and AA3003 at 500 °C. While the elongation of both AA5083 and AA3003 was improved at a higher temperature, the elongation of AA5083 was inversely related to strain rate. It was concluded that the higher the temperature is the better the formability and elongation of the two alloys but at the expense of work hardening. This suggests a trade-off situation between formability and strength. 

scholarly journals Auto-Colorization of Historical Images Using Deep Convolutional Neural Networks

Mathematics ◽  
2020 ◽  
Vol 8 (12) ◽  
pp. 2258
Author(s):  
Madhab Raj Joshi ◽  
Lewis Nkenyereye ◽  
Gyanendra Prasad Joshi ◽  
S. M. Riazul Islam ◽  
Mohammad Abdullah-Al-Wadud ◽  
...  
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
User Study ◽  
Mean Squared Error ◽  
Color Image ◽  
Machine Learning Techniques ◽  
Global Features ◽  
Black And White ◽  
Historical Images ◽  
Learning Techniques

Enhancement of Cultural Heritage such as historical images is very crucial to safeguard the diversity of cultures. Automated colorization of black and white images has been subject to extensive research through computer vision and machine learning techniques. Our research addresses the problem of generating a plausible colored photograph of ancient, historically black, and white images of Nepal using deep learning techniques without direct human intervention. Motivated by the recent success of deep learning techniques in image processing, a feed-forward, deep Convolutional Neural Network (CNN) in combination with Inception- ResnetV2 is being trained by sets of sample images using back-propagation to recognize the pattern in RGB and grayscale values. The trained neural network is then used to predict two a* and b* chroma channels given grayscale, L channel of test images. CNN vividly colorizes images with the help of the fusion layer accounting for local features as well as global features. Two objective functions, namely, Mean Squared Error (MSE) and Peak Signal-to-Noise Ratio (PSNR), are employed for objective quality assessment between the estimated color image and its ground truth. The model is trained on the dataset created by ourselves with 1.2 K historical images comprised of old and ancient photographs of Nepal, each having 256 × 256 resolution. The loss i.e., MSE, PSNR, and accuracy of the model are found to be 6.08%, 34.65 dB, and 75.23%, respectively. Other than presenting the training results, the public acceptance or subjective validation of the generated images is assessed by means of a user study where the model shows 41.71% of naturalness while evaluating colorization results.

scholarly journals Deep Learning Classification of Canine Behavior Using a Single Collar-Mounted Accelerometer: Real-World Validation

Animals ◽  
2021 ◽  
Vol 11 (6) ◽  
pp. 1549
Author(s):  
Robert D. Chambers ◽  
Nathanael C. Yoder ◽  
Aletha B. Carson ◽  
Christian Junge ◽  
David E. Allen ◽  
...  
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Real World ◽  
Learning Algorithm ◽  
Drinking Behavior ◽  
True Positive Rate ◽  
Training Dataset ◽  
Activity Levels ◽  
Accelerometer Data ◽  
Activity Monitors

Collar-mounted canine activity monitors can use accelerometer data to estimate dog activity levels, step counts, and distance traveled. With recent advances in machine learning and embedded computing, much more nuanced and accurate behavior classification has become possible, giving these affordable consumer devices the potential to improve the efficiency and effectiveness of pet healthcare. Here, we describe a novel deep learning algorithm that classifies dog behavior at sub-second resolution using commercial pet activity monitors. We built machine learning training databases from more than 5000 videos of more than 2500 dogs and ran the algorithms in production on more than 11 million days of device data. We then surveyed project participants representing 10,550 dogs, which provided 163,110 event responses to validate real-world detection of eating and drinking behavior. The resultant algorithm displayed a sensitivity and specificity for detecting drinking behavior (0.949 and 0.999, respectively) and eating behavior (0.988, 0.983). We also demonstrated detection of licking (0.772, 0.990), petting (0.305, 0.991), rubbing (0.729, 0.996), scratching (0.870, 0.997), and sniffing (0.610, 0.968). We show that the devices’ position on the collar had no measurable impact on performance. In production, users reported a true positive rate of 95.3% for eating (among 1514 users), and of 94.9% for drinking (among 1491 users). The study demonstrates the accurate detection of important health-related canine behaviors using a collar-mounted accelerometer. We trained and validated our algorithms on a large and realistic training dataset, and we assessed and confirmed accuracy in production via user validation.

scholarly journals 2D-3D integration of hexagonal boron nitride and a high-κ dielectric for ultrafast graphene-based electro-absorption modulators

2021 ◽  
Vol 12 (1) ◽  
Author(s):  
Hitesh Agarwal ◽  
Bernat Terrés ◽  
Lorenzo Orsini ◽  
Alberto Montanaro ◽  
Vito Sorianello ◽  
...  
Keyword(s):  
Boron Nitride ◽  
Hafnium Oxide ◽  
High Speed ◽  
Hexagonal Boron Nitride ◽  
Temperature Stability ◽  
Cost Effective ◽  
Fold Increase ◽  
Building Blocks ◽  
High Quality ◽  
New Device

AbstractElectro-absorption (EA) waveguide-coupled modulators are essential building blocks for on-chip optical communications. Compared to state-of-the-art silicon (Si) devices, graphene-based EA modulators promise smaller footprints, larger temperature stability, cost-effective integration and high speeds. However, combining high speed and large modulation efficiencies in a single graphene-based device has remained elusive so far. In this work, we overcome this fundamental trade-off by demonstrating the 2D-3D dielectric integration in a high-quality encapsulated graphene device. We integrated hafnium oxide (HfO2) and two-dimensional hexagonal boron nitride (hBN) within the insulating section of a double-layer (DL) graphene EA modulator. This combination of materials allows for a high-quality modulator device with high performances: a ~39 GHz bandwidth (BW) with a three-fold increase in modulation efficiency compared to previously reported high-speed modulators. This 2D-3D dielectric integration paves the way to a plethora of electronic and opto-electronic devices with enhanced performance and stability, while expanding the freedom for new device designs.

Sign in / Sign up

Export Citation Format

Share Document