scholarly journals Image-to-Image Translation with Multi-Path Consistency Regularization

2019 ◽  
Author(s):  
Jianxin Lin ◽  
Yingce Xia ◽  
Yijun Wang ◽  
Tao Qin ◽  
Zhibo Chen
Keyword(s):  
Machine Learning ◽  
Computer Vision ◽  
The Other ◽  
Natural Images ◽  
Original Image ◽  
Target Domain ◽  
Source Domain ◽  
Face To Face ◽  
Image Translation ◽  
The Difference

Image translation across different domains has attracted much attention in both machine learning and computer vision communities. Taking the translation from a source domain to a target domain as an example, existing algorithms mainly rely on two kinds of loss for training: One is the discrimination loss, which is used to differentiate images generated by the models and natural images; the other is the reconstruction loss, which measures the difference between an original image and the reconstructed version. In this work, we introduce a new kind of loss, multi-path consistency loss, which evaluates the differences between direct translation from source domain to target domain and indirect translation from source domain to an auxiliary domain to target domain, to regularize training. For multi-domain translation (at least, three) which focuses on building translation models between any two domains, at each training iteration, we randomly select three domains, set them respectively as the source, auxiliary and target domains, build the multi-path consistency loss and optimize the network. For two-domain translation, we need to introduce an additional auxiliary domain and construct the multi-path consistency loss. We conduct various experiments to demonstrate the effectiveness of our proposed methods, including face-to-face translation, paint-to-photo translation, and de-raining/de-noising translation.

scholarly journals Hyperbolic markers in modeling hyperbole: a scenario-based account

2021 ◽  
Vol 85 ◽  
pp. 61-71
Author(s):  
Carla Ovejas Ramírez
Keyword(s):  
Cognitive Linguistics ◽  
Language Use ◽  
Meaning Construction ◽  
The Other ◽  
Target Domain ◽  
Real Situation ◽  
Source Domain ◽  
The Real ◽  
Theoretical Context ◽  
Hypothetical Scenario

This article discusses hyperbolic markers in modeling hyperbole from the perspective of a scenario-based account of language use within the framework of Cognitive Linguistics. In this view, hyperbole is seen as a mapping across two conceptual domains (Peña y Ruiz de Mendoza, 2017), a source domain, here relabeled as the magnified scenario, which contains a hypothetical unrealistic situation based on exaggeration, and a target domain or observable scenario which depicts the real situation addressed by the hyperbolic expression. Since the hypothetical scenario is a magnified version of the observable scenario, the mapping contains source-target matches in varying degrees of resemblance. Within this theoretical context, the article explores resources available to speakers for the construction of magnified scenarios leading to hyperbolic interpretation. Among such resources, we find hyperbole markers and the setting up of domains of reference. Finally, the article also discusses hyperbole blockers, which cancel out the activity of the other hyperbolic meaning construction mechanisms.

scholarly journals INITIATION OF INTERACTION AS THE BEGINNING OF SOCIAL PARTICIPATION: CONVERSATION ANALYSIS OF A JAPANESE SENIOR CLUB

2019 ◽  
Vol 3 (Supplement_1) ◽  
pp. S173-S174
Author(s):  
Eri Sakai ◽  
Akihiko Kamesawa ◽  
Riko Nakayama ◽  
Jihoon Kim ◽  
Akizuki Yuri ◽  
...  
Keyword(s):  
Conversation Analysis ◽  
Social Participation ◽  
Social Factors ◽  
Senior Citizens ◽  
Participation Rate ◽  
The Other ◽  
Teaching Materials ◽  
Face To Face ◽  
The Difference

Abstract The rate of social participation of senior citizens in a senior club’s activities is not equal to the rate of desire for the said participation. Earlier studies mainly examined personal and social factors which influence the participation rate, overlooking the practical methods by which senior citizens can overcome barriers to participating in club activities. Our study aims to clarify the features of a club activity as a resource by analyzing the activity’s interactions. Our study is based on data extracted from videotaped recordings of a senior calligraphy club in Kanto, Japan. In September 2018, one lecturer and 11 participants were videotaped for 3 hours, and the video underwent conversation analysis, which elucidates how people organize activities under specific circumstances. We analyzed how a female newcomer to the activity initiated face-to-face interaction, which is considered the first step of social participation. She talked to other participants who were familiar with the exercise several times by inquiring how to read kanjis on teaching materials. These findings suggest that visualization of skill relative to the other creates an environment for initiating face-to-face interaction. In this case, the newcomer utilized the difference in skill denoted by teaching materials and was given the rational reason to talk to the others already engaging in the activity. Therefore, designing teaching materials that assign the learning level of each participant may be effective in promoting social participation in senior study clubs.

scholarly journals Customized Dictionary Learning for Subdatasets with Fine Granularity

2016 ◽  
Vol 2016 ◽  
pp. 1-13
Author(s):  
Lei Ye ◽  
Can Wang ◽  
Xin Xu ◽  
Hui Qian
Keyword(s):  
Machine Learning ◽  
Computer Vision ◽  
Theoretical Analysis ◽  
Improve Performance ◽  
Real World Data ◽  
Sparse Models ◽  
Learned Dictionary ◽  
Wide Range ◽  
Sparsity Constraints ◽  
The Difference

Sparse models have a wide range of applications in machine learning and computer vision. Using a learned dictionary instead of an “off-the-shelf” one can dramatically improve performance on a particular dataset. However, learning a new one for each subdataset (subject) with fine granularity may be unwarranted or impractical, due to restricted availability subdataset samples and tremendous numbers of subjects. To remedy this, we consider the dictionary customization problem, that is, specializing an existing global dictionary corresponding to the total dataset, with the aid of auxiliary samples obtained from the target subdataset. Inspired by observation and then deduced from theoretical analysis, a regularizer is employed penalizing the difference between the global and the customized dictionary. By minimizing the sum of reconstruction errors of the above regularizer under sparsity constraints, we exploit the characteristics of the target subdataset contained in the auxiliary samples while maintaining the basic sketches stored in the global dictionary. An efficient algorithm is presented and validated with experiments on real-world data.

scholarly journals ITERATIVE RE-WEIGHTED INSTANCE TRANSFER FOR DOMAIN ADAPTATION

2016 ◽  
Vol III-3 ◽  
pp. 339-346 ◽  
Author(s):  
A. Paul ◽  
F. Rottensteiner ◽  
C. Heipke
Keyword(s):  
Domain Adaptation ◽  
Training Data ◽  
Aerial Images ◽  
Target Domain ◽  
Data Set ◽  
Source Domain ◽  
The Difference ◽  
A New Technique ◽  
Class Labels ◽  
Decision Boundaries

Domain adaptation techniques in transfer learning try to reduce the amount of training data required for classification by adapting a classifier trained on samples from a source domain to a new data set (target domain) where the features may have different distributions. In this paper, we propose a new technique for domain adaptation based on logistic regression. Starting with a classifier trained on training data from the source domain, we iteratively include target domain samples for which class labels have been obtained from the current state of the classifier, while at the same time removing source domain samples. In each iteration the classifier is re-trained, so that the decision boundaries are slowly transferred to the distribution of the target features. To make the transfer procedure more robust we introduce weights as a function of distance from the decision boundary and a new way of regularisation. Our methodology is evaluated using a benchmark data set consisting of aerial images and digital surface models. The experimental results show that in the majority of cases our domain adaptation approach can lead to an improvement of the classification accuracy without additional training data, but also indicate remaining problems if the difference in the feature distributions becomes too large.

Robot Process Automation (RPA) and Its Future

2022 ◽  
pp. 35-58
Author(s):  
Ozge Doguc
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Computer Vision ◽  
Customer Satisfaction ◽  
Human Behavior ◽  
The Other ◽  
Human Intervention ◽  
Process Automation ◽  
Robotic Process Automation ◽  
Human Agent

Many software automation techniques have been developed in the last decade to cut down cost, improve customer satisfaction, and reduce errors. Robotic process automation (RPA) has become increasingly popular recently. RPA offers software robots (bots) that can mimic human behavior. Attended robots work in tandem with humans and can operate while the human agent is active on the computer. On the other hand, unattended robots operate behind locked screens and are designed to execute automations that don't require any human intervention. RPA robots are equipped with artificial intelligence engines such as computer vision and machine learning, and both robot types can learn automations by recording human actions.

scholarly journals Meditation Practice Improves Short-Term Changes in Heart Rate Variability

2020 ◽  
Vol 17 (6) ◽  
pp. 2128 ◽  
Author(s):  
Kang-Ming Chang ◽  
Miao-Tien Wu Chueh ◽  
Yi-Jung Lai
Keyword(s):  
Blood Pressure ◽  
Heart Rate ◽  
Heart Rate Variability ◽  
Physiological Parameters ◽  
The Other ◽  
Meditation Practice ◽  
Short Term ◽  
Face To Face ◽  
Before And After ◽  
The Difference

Background: It is well known that meditation improves the physical and psychological condition of its practitioners. This study investigated the heart rate variability response of meditation practitioners in two Chan master teaching environments, namely face-to-face and video classes. Methods: Experimental sessions were conducted, one featuring face-to-face classes and the other featuring video classes. The difference in participants’ physiological parameters (blood pressure and heart rate variability) between the two experimental sessions was determined. In the first session, physiological parameters were recorded twice, before and after one teaching course, and the second session took place one month after the first. The first and second sessions had 45 and 27 participants, respectively. Those involved in the first experiment had no experience with meditation, whereas participants in the second experiment had practiced meditation for an average of 9 years (range = 1 to 27 years). Both experiments were conducted once a week, with each session lasting 1.5 h. Results: For both experiments, both heart rate and heart rate variability by age significantly decreased after one teaching course. Conclusions: Chan meditation practitioners benefit from receiving both face-to-face and video class teaching from a Chan master.

“Pour water where it burns”

2016 ◽  
Vol 6 (1) ◽  
pp. 103-133 ◽  
Author(s):  
Mohsen Bakhtiar
Keyword(s):  
The Other ◽  
Islamic Republic ◽  
General Phenomenon ◽  
Target Domain ◽  
The West ◽  
Source Domain ◽  
Supreme Leader ◽  
High Level ◽  
The U.S ◽  
Selection Of

While dysphemism has been extensively studied as a general phenomenon, there are not too many studies on how it is used in political discourse by top officials. This paper aims to examine the ways in which a sample of two high-level Iranian politicians offensively conceptualize their alleged enemies, namely the U.S., Israel, and the West, through conceptual metaphors and metonymies. A cognitive linguistic analysis of the speeches of Iran’s supreme leader and ex-president Mahmoud Ahmadinejad indicate that the selection of the metaphorical dysphemistic source domain is primarily determined by religion, previous discourse (pre-existing conventional dysphemistic metaphors), aspects of the target domain, and anger or hatred toward the enemies. The analysis indicates that most of the pejorative connotations are attributed to Israel as the alleged number one enemy of Iran via Israel is an animal, Israel is a tumor, and Israel is a bastard. The other presumed enemies, that is, the U.S. and the West are characterized via the u.s. is a devil, and the u.s. and the west are criminals. Moreover, the two politicians, while resorting to taboo concepts, remain loyal to the established discursive norms of delegitimizing the actions and thoughts of the enemies of the Islamic Republic.

scholarly journals ITERATIVE RE-WEIGHTED INSTANCE TRANSFER FOR DOMAIN ADAPTATION

2016 ◽  
Vol III-3 ◽  
pp. 339-346 ◽  
Author(s):  
A. Paul ◽  
F. Rottensteiner ◽  
C. Heipke
Keyword(s):  
Domain Adaptation ◽  
Training Data ◽  
Aerial Images ◽  
Target Domain ◽  
Data Set ◽  
Source Domain ◽  
The Difference ◽  
A New Technique ◽  
Class Labels ◽  
Decision Boundaries

Domain adaptation techniques in transfer learning try to reduce the amount of training data required for classification by adapting a classifier trained on samples from a source domain to a new data set (target domain) where the features may have different distributions. In this paper, we propose a new technique for domain adaptation based on logistic regression. Starting with a classifier trained on training data from the source domain, we iteratively include target domain samples for which class labels have been obtained from the current state of the classifier, while at the same time removing source domain samples. In each iteration the classifier is re-trained, so that the decision boundaries are slowly transferred to the distribution of the target features. To make the transfer procedure more robust we introduce weights as a function of distance from the decision boundary and a new way of regularisation. Our methodology is evaluated using a benchmark data set consisting of aerial images and digital surface models. The experimental results show that in the majority of cases our domain adaptation approach can lead to an improvement of the classification accuracy without additional training data, but also indicate remaining problems if the difference in the feature distributions becomes too large.

scholarly journals The Difference the Body Makes: The Teacher’s Presence Online and Offline

2006 ◽  
Vol 2 (2) ◽  
Author(s):  
Carina Henriksson
Keyword(s):  
Online Communication ◽  
The Body ◽  
Virtual Space ◽  
The Other ◽  
Online Student ◽  
Face To Face ◽  
Computer Mediated ◽  
On Line ◽  
The Difference ◽  
The Relationship

This paper examines the question: what is the experience of meeting online and how does it differ from ordinary classroom situations? Drawing from personal experience, the author explores possible experiences of existing in virtual space and time. How do people meet, get to know each other and, interact in a pedagogical situation? Her experience as an online student made her to seriously reflect on the experiential nature of the computer-mediated encounter. But, it was not until she happened to participate in a workshop offered by the same teacher that the contrasts began to take shape for her. If there is a difference between online and offline meetings, what is it that makes the difference? Online communication could, just as face-to-face meetings, create feelings of closeness, and friendship; from the other-as-a-text on the screen, we subjectively create the other-as-an-idea, an idea that might be perceived as the real other. But is it? What reality is for real? What is the nature of the relationship established between body-less persons on line, and what difference does the body make in a face-to-face meeting?

scholarly journals Domain Adaptation for Semantic Parsing

2020 ◽  
Author(s):  
Zechang Li ◽  
Yuxuan Lai ◽  
Yansong Feng ◽  
Dongyan Zhao
Keyword(s):  
Domain Adaptation ◽  
Specific Information ◽  
Semantic Parsing ◽  
Target Domain ◽  
Source Domain ◽  
Domain Specific ◽  
The Difference ◽  
Two Stages ◽  
Coarse To Fine ◽  
Target Data

Recently, semantic parsing has attracted much attention in the community. Although many neural modeling efforts have greatly improved the performance, it still suffers from the data scarcity issue. In this paper, we propose a novel semantic parser for domain adaptation, where we have much fewer annotated data in the target domain compared to the source domain. Our semantic parser benefits from a two-stage coarse-to-fine framework, thus can provide different and accurate treatments for the two stages, i.e., focusing on domain invariant and domain specific information, respectively. In the coarse stage, our novel domain discrimination component and domain relevance attention encourage the model to learn transferable domain general structures. In the fine stage, the model is guided to concentrate on domain related details. Experiments on a benchmark dataset show that our method consistently outperforms several popular domain adaptation strategies. Additionally, we show that our model can well exploit limited target data to capture the difference between the source and target domain, even when the target domain has far fewer training instances.

Sign in / Sign up

Export Citation Format

Share Document