A common problem in human-object interaction (HOI) detection task is that numerous HOI classes have only a small number of labeled examples, resulting in training sets with a long-tailed distribution. The lack of positive labels can lead to low classification accuracy for these classes. Towards addressing this issue, we observe that there exist natural correlations and anti-correlations among human-object interactions. In this paper, we model the correlations as action co-occurrence matrices and present techniques to learn these priors and leverage them for more effective training, especially in rare classes. The utility of our approach is demonstrated experimentally, where the performance of our approach exceeds the state-of-the-art methods on both of the two leading HOI detection benchmark datasets, HICO-Det and V-COCO.
This review summarizes the fundamentals of the phenomenon of electron transfer (ET) reactions occurring in redox enzymes that were widely employed for the development of electroanalytical devices, like biosensors, and enzymatic fuel cells (EFCs). A brief introduction on the ET observed in proteins/enzymes and its paradigms (e.g., classification of ET mechanisms, maximal distance at which is observed direct electron transfer, etc.) are given. Moreover, the theoretical aspects related to direct electron transfer (DET) are resumed as a guideline for newcomers to the field. Snapshots on the ET theory formulated by Rudolph A. Marcus and on the mathematical model used to calculate the ET rate constant formulated by Laviron are provided. Particular attention is devoted to the case of glucose oxidase (GOx) that has been erroneously classified as an enzyme able to transfer electrons directly. Thereafter, all tools available to investigate ET issues are reported addressing the discussions toward the development of new methodology to tackle ET issues. In conclusion, the trends toward upcoming practical applications are suggested as well as some directions in fundamental studies of bioelectrochemistry.
Due to the expansion of distributed renewable energy resources, peer to peer energy trading (P2P DET) is expected to be one of the key elements of next generation power systems. P2P DET can provide various benefits such as creating a competitive energy market, reducing power outages, increasing overall efficiency of power systems and supplementing alternative sources of energy according to user preferences. Because of these promising advantages, P2P DET has attracted the attention of several researchers. Current research related to P2P DET include demand response optimization, power routing, network communication, security and privacy. This paper presents a review of the main research topics revolving around P2P DET. Particularly, we present a comprehensive survey of existing demand response optimization models, power routing devices and power routing algorithms. We also identify some key challenges faced in realizing P2P DET. Furthermore, we discuss state of the art enabling technologies such as Energy Internet, Blockchain and Software Defined Networking (SDN) and we provide insights into future research directions.
Bioelectrochemical conversion of CO2 to CH4 is a promising way to increase the calorific value of biogas produced during anaerobic digestion. There are two groups of methanogens enriched in these systems, hydrogenotrophs and acetoclastic methanogens that can also directly accept electrons from an electrode or another microorganism. In this study, a microbial electrolysis cell (MEC) poised at -500 mV (vs. SHE) was operated for biogas upgrading. Methane content in the biogas increased from 71% to >90%, and 8.2% of the CO2 was converted to methane. Methanothrix, an acetoclastic methanogen that can participate in direct electron transfer (DET), and Azonexus, an acetate-oxidizing electrogen, were enriched on the cathode. Transcriptomics revealed that Methanothrix on the cathode were using the CO2 reduction pathway, while Methanothrix in the bulk sludge were using the acetate decarboxylation pathway for production of methane. These results show that stimulation of DET in MEC enhances biogas-upgrading processes.
Surface defect inspection is a key step to ensure the quality of the hot rolled steel surface. However, current advanced detection (DET) methods have high precision but low detection speed, which hinders the application of the detector in actual production. In this work, a real-time detection network (RDN) focusing on both speed and accuracy is proposed to solve the problem of steel surface defect detection. RDN takes ResNet-dcn, a modular encoding, and decoding network with light weights, as the basic convolutional architecture whose backbone is pretrained on ImageNet. To improve the detection accuracy, a skip layer connection module (SCM) and a pyramid feature fusion module (PFM) are involved into RDN. On the standard dataset NEU-DET, the proposed method can achieve the state-of-the-art recognition speed of 64 frames per second (FPS) and the mean average precision of 80.0% on a single GPU, which fully meets the requirements of the detection accuracy and speed in the actual production line.
We explore long-term temporal visual correspondence-based optimization for 3D video object detection in this work. Visual correspondence refers to one-to-one mappings for pixels across multiple images. Correspondence-based optimization is the cornerstone for 3D scene reconstruction but is less studied in 3D video object detection, because moving objects violate multi-view geometry constraints and are treated as outliers during scene reconstruction. We address this issue by treating objects as first-class citizens during correspondence-based optimization. In this work, we propose BA-Det, an end-to-end optimizable object detector with object-centric temporal correspondence learning and featuremetric object bundle adjustment. Empirically, we verify the effectiveness and efficiency of BA-Det for multiple baseline 3D detectors under various setups. Our BA-Det achieves SOTA performance on the large-scale Waymo Open Dataset (WOD) with only marginal computation cost. Our code is available at https://github.com/jiaweihe1996/BA-Det.
Kristian Firing, Anne Torhild Klomsten, Frode Moen
Sammendrag I studier av veiledning er det ofte gjort et skille mellom den akademiske og den personlige dimensjon. Det anerkjennes videre at veiledning har betydning for studenters mestring. Med bakgrunn i denne faglige argumentasjonen, utviklet vi folgende problemstilling: Hvordan opplever studenter sitt mote med veileder i forhold til mestring under veiledning pa masteroppgaver? Intervjuer viste at studentene opplevde veilederne i spekteret fra svak til sterk innenfor bade den akademiske og den personlige dimensjon. Den pafolgende analysen resulterte i folgende fire kategorier: (1) Ikke-moting, (2) Rosemoting, (3) Prestasjonsmoting og (4) Mestringsmoting. Videre ble de ulike motene diskutert i forhold til mestring. Motene og veiledningens betydning for mestring illustreres gjennom ytringen ”det er veiledningen som gjor at en foler at en mestrer”. Abstract Within the field of counseling; academic and personal dimensions are often treated separately from each other in research. Counseling is recognized as being influential on a student's feeling of mastery. Based on this argument, the following research question was developed: What experiences do students have from their meetings with their counselors during counseling related to their Master Degrees, and how do these relate to their feeling of mastery? Our findings show that the students experienced their counselor as being from weak to strong within both academic and personal dimensions. The analysis revealed four categories: 1) Not Meeting, 2) Praise Meeting, 3) Performance Meeting, and 4) Mastery Meeting. The different categories were discussed relating to the student's feeling of mastery. The different categories of meetings and the counselor's importance regarding the student's feeling of mastery are illustrated by the statement “it is the counseling that affects my feeling of mastery”. (Published: 1 August 2013) Citation: Uniped 2013, 36 : 21515 - http://dx.doi.org/10.3402/uniped.v36i2.21515