scholarly journals ROSITA: Enhancing Vision-and-Language Semantic Alignments via Cross- and Intra-modal Knowledge Integration

2021 ◽  
Author(s):  
Yuhao Cui ◽  
Zhou Yu ◽  
Chunqi Wang ◽  
Zhongzhou Zhao ◽  
Ji Zhang ◽  
...  
Author(s):  
Honghai LI ◽  
Jun CAI

The transformation of China's design innovation industry has highlighted the importance of design research. The design research process in practice can be regarded as the process of knowledge production. The design 3.0 mode based on knowledge production MODE2 has been shown in the Chinese design innovation industry. On this cognition, this paper establishes a map with two dimensions of how knowledge integration occurs in practice based design research, which are the design knowledge transfer and contextual transformation of design knowledge. We use this map to carry out the analysis of design research cases. Through the analysis, we define four typical practice based design research models from the viewpoint of knowledge integration. This method and the proposed model can provide a theoretical basis and a path for better management design research projects.


Fachsprache ◽  
2017 ◽  
Vol 32 (3-4) ◽  
pp. 100-121
Author(s):  
Friederike Prassl

This article focuses on the decision-making processes involved in research and knowledge integration in translation processes. First, the relevance of decision taking intranslation is discussed. Second, the psychology of decision making as seen by Jungermann et al. (2005) is introduced, who propose a categorization of decision-making processes intofour types: “routinized”, “stereotype”, “reflected” and “constructed”. This classification is then applied to the translations by five professional translators and five novices of five segments occurring in a popular-science text. The analysis reveals that the decision-making types are distributed differently among students and professional translators, which also has to be seen against the background of whether the decisions made were successful or not. The preliminary results of this study show that students resort to reflected decisions in most cases, but with a low success rate. Professionals achieve a higher success rate when making reflected decisions. As expected, they also make more routinized decisions than students. The professionals’ success rates improve with increasing cognitive involvement, while their failure rates are relatively high when making routinized decisions, an aspect worthwhile considering in translation didactics.


Author(s):  
Libby Gerard ◽  
Erika Tate ◽  
Jennifer Chiu ◽  
Stephanie Corliss ◽  
Marcia Linn

Sensors ◽  
2021 ◽  
Vol 21 (3) ◽  
pp. 1012
Author(s):  
Jisu Hwang ◽  
Incheol Kim

Due to the development of computer vision and natural language processing technologies in recent years, there has been a growing interest in multimodal intelligent tasks that require the ability to concurrently understand various forms of input data such as images and text. Vision-and-language navigation (VLN) require the alignment and grounding of multimodal input data to enable real-time perception of the task status on panoramic images and natural language instruction. This study proposes a novel deep neural network model (JMEBS), with joint multimodal embedding and backtracking search for VLN tasks. The proposed JMEBS model uses a transformer-based joint multimodal embedding module. JMEBS uses both multimodal context and temporal context. It also employs backtracking-enabled greedy local search (BGLS), a novel algorithm with a backtracking feature designed to improve the task success rate and optimize the navigation path, based on the local and global scores related to candidate actions. A novel global scoring method is also used for performance improvement by comparing the partial trajectories searched thus far with a plurality of natural language instructions. The performance of the proposed model on various operations was then experimentally demonstrated and compared with other models using the Matterport3D Simulator and room-to-room (R2R) benchmark datasets.


Sign in / Sign up

Export Citation Format

Share Document