scholarly journals Progressive Boundary Refinement Network for Temporal Action Detection

2020 ◽  
Vol 34 (07) ◽  
pp. 11612-11619
Author(s):  
Qinying Liu ◽  
Zilei Wang
Keyword(s):  
State Of The Art ◽  
The State ◽  
Action Detection ◽  
Fine Grained ◽  
The Third ◽  
One Stage ◽  
The Family ◽  
Main Components ◽  
Temporal Action ◽  
Boundary Refinement

Temporal action detection is a challenging task due to vagueness of action boundaries. To tackle this issue, we propose an end-to-end progressive boundary refinement network (PBRNet) in this paper. PBRNet belongs to the family of one-stage detectors and is equipped with three cascaded detection modules for localizing action boundary more and more precisely. Specifically, PBRNet mainly consists of coarse pyramidal detection, refined pyramidal detection, and fine-grained detection. The first two modules build two feature pyramids to perform the anchor-based detection, and the third one explores the frame-level features to refine the boundaries of each action instance. In the fined-grained detection module, three frame-level classification branches are proposed to augment the frame-level features and update the confidence scores of action instances. Evidently, PBRNet integrates the anchor-based and frame-level methods. We experimentally evaluate the proposed PBRNet and comprehensively investigate the effect of the main components. The results show PBRNet achieves the state-of-the-art detection performances on two popular benchmarks: THUMOS'14 and ActivityNet, and meanwhile possesses a high inference speed.

Health Informatics Education in the Third World

1989 ◽  
Vol 28 (04) ◽  
pp. 270-272 ◽  
Author(s):  
O. Rienhoff
Keyword(s):  
Developing Countries ◽  
Health Informatics ◽  
Third World ◽  
State Of The Art ◽  
The State ◽  
Collaboration Network ◽  
Educational Tools ◽  
The Third ◽  
The Third World

Abstract:The state of the art is summarized showing many efforts but only few results which can serve as demonstration examples for developing countries. Education in health informatics in developing countries is still mainly dealing with the type of health informatics known from the industrialized world. Educational tools or curricula geared to the matter of development are rarely to be found. Some WHO activities suggest that it is time for a collaboration network to derive tools and curricula within the next decade.

scholarly journals Representation Learning for Fine-Grained Change Detection

Sensors ◽  
2021 ◽  
Vol 21 (13) ◽  
pp. 4486
Author(s):  
Niall O’Mahony ◽  
Sean Campbell ◽  
Lenka Krpalkova ◽  
Anderson Carvalho ◽  
Joseph Walsh ◽  
...  
Keyword(s):  
Deep Learning ◽  
Change Detection ◽  
Model Calibration ◽  
State Of The Art ◽  
Representation Learning ◽  
Machine Intelligence ◽  
The State ◽  
Sensor Data ◽  
Fine Grained ◽  
Learning Techniques

Fine-grained change detection in sensor data is very challenging for artificial intelligence though it is critically important in practice. It is the process of identifying differences in the state of an object or phenomenon where the differences are class-specific and are difficult to generalise. As a result, many recent technologies that leverage big data and deep learning struggle with this task. This review focuses on the state-of-the-art methods, applications, and challenges of representation learning for fine-grained change detection. Our research focuses on methods of harnessing the latent metric space of representation learning techniques as an interim output for hybrid human-machine intelligence. We review methods for transforming and projecting embedding space such that significant changes can be communicated more effectively and a more comprehensive interpretation of underlying relationships in sensor data is facilitated. We conduct this research in our work towards developing a method for aligning the axes of latent embedding space with meaningful real-world metrics so that the reasoning behind the detection of change in relation to past observations may be revealed and adjusted. This is an important topic in many fields concerned with producing more meaningful and explainable outputs from deep learning and also for providing means for knowledge injection and model calibration in order to maintain user confidence.

Researching Telework

2000 ◽  
pp. 255-276 ◽  
Author(s):  
Nicole B. Ellison
Keyword(s):  
Organizational Culture ◽  
State Of The Art ◽  
The State ◽  
Future Research ◽  
Scholarly Literature ◽  
Home Based ◽  
The Family ◽  
The Individual ◽  
The Impact

This chapter examines the state of the art in telework research. The author reviews the most central scholarly literature examining the phenomenon of telework (also called home-based work or telecommuting) and develops a framework for organizing this body of work. She organizes previous research on telework into six major thematic concerns relating to the definition, measurement, and scope of telework; management of teleworkers; travel-related impacts of telework; organizational culture and employee isolation; boundaries between “home” and “work” and the impact of telework on the individual and the family. Areas for future research are suggested.

scholarly journals Thiruvenkata Rajendra Seshadri, 3 February 1900 - 27 September 1975

1979 ◽  
Vol 25 ◽  
pp. 504-533
Keyword(s):  
Tamil Nadu ◽  
South India ◽  
Small Town ◽  
The State ◽  
Local School ◽  
Only Children ◽  
The Third ◽  
The Family ◽  
Personal Attitudes

Thiruvenkata Rajendra Seshadri was born on 3 February 1900 in the small town of Kulitalai lying on the bank of the Kaveri, one of the seven sacred rivers of India, and situated in the Tiruchy district of South India. This district formed a part of the Madras Presidency of pre-independent India and is now a part of the state of Tamil Nadu of the Indian Republic. His father, Thiruvengadatha Iyengar, was a teacher in a local school; his mother was Namagiri Ammal, and T. R. Seshadri was the third of five sons who were the only children of the marriage. The family was deeply religious, and this influence was dominant throughout T. R. Seshadri’s life, not only in his personal attitudes but also in his complete dedication to his work.

scholarly journals Relational Prototypical Network for Weakly Supervised Temporal Action Localization

2020 ◽  
Vol 34 (07) ◽  
pp. 11053-11060
Author(s):  
Linjiang Huang ◽  
Yan Huang ◽  
Wanli Ouyang ◽  
Liang Wang
Keyword(s):  
State Of The Art ◽  
Previous Method ◽  
The State ◽  
Localization Method ◽  
Action Localization ◽  
Art Methods ◽  
Prototype Learning ◽  
Weakly Supervised ◽  
Temporal Action

In this paper, we propose a weakly supervised temporal action localization method on untrimmed videos based on prototypical networks. We observe two challenges posed by weakly supervision, namely action-background separation and action relation construction. Unlike the previous method, we propose to achieve action-background separation only by the original videos. To achieve this, a clustering loss is adopted to separate actions from backgrounds and learn intra-compact features, which helps in detecting complete action instances. Besides, a similarity weighting module is devised to further separate actions from backgrounds. To effectively identify actions, we propose to construct relations among actions for prototype learning. A GCN-based prototype embedding module is introduced to generate relational prototypes. Experiments on THUMOS14 and ActivityNet1.2 datasets show that our method outperforms the state-of-the-art methods.

Some Graeco-Phoenician Shrines

1911 ◽  
Vol 31 ◽  
pp. 56-64 ◽  
Author(s):  
G. F. Hill
Keyword(s):  
Roman Empire ◽  
The State ◽  
Fifth Century ◽  
Third Century ◽  
The Third ◽  
The Family ◽  
Roman Culture ◽  
Roman Age

With but two exceptions, no trace now remains of the shrines with which this paper deals, or at least no trace has been revealed by excavation. Practically the sole record of these buildings is to be found on the coins struck in the district during the period of the Roman Empire, and more especially during the third century of our era. The earlier coins, from the beginning of the coinage towards the end of the fifth century B.C., tell us something about the cults, but little of their furniture. But in the Roman age, especially during the time of the family of Severus and Elagabalus, there was a considerable outburst of coinage, which, in its types, reveals certain details interesting to the student of the fringe of Greek and Roman culture.The evidence thus provided is necessarily disjointed, and concerns only the external, official aspects of the Phoenician religion. The inner truth of these things, it is safe to say, is hidden for ever: even the development from the primitive religion to the weird syncretistic systems of the Roman age is hopelessly obscure. One can only see dimly what was the state of things during the period illustrated by the monuments.

scholarly journals Beyond the “Third Wave of Positive Psychology”: Challenges and Opportunities for Future Research

2022 ◽  
Vol 12 ◽  
Author(s):  
Marié P. Wissing
Keyword(s):  
Positive Psychology ◽  
State Of The Art ◽  
Well Being ◽  
The State ◽  
Future Research ◽  
Complex Nature ◽  
Third Wave ◽  
Health And Wellness ◽  
The Third ◽  
Challenges And Opportunities

The positive psychology (PP) landscape is changing, and its initial identity is being challenged. Moving beyond the “third wave of PP,” two roads for future research and practice in well-being studies are discerned: The first is the state of the art PP trajectory that will (for the near future) continue as a scientific (sub)discipline in/next to psychology (because of its popular brand name). The second trajectory (main focus of this manuscript) links to pointers described as part of the so-called third wave of PP, which will be argued as actually being the beginning of a new domain of inter- or transdisciplinary well-being studies in its own right. It has a broader scope than the state of the art in PP, but is more delineated than in planetary well-being studies. It is in particular suitable to understand the complex nature of bio-psycho-social-ecological well-being, and to promote health and wellness in times of enormous challenges and changes. A unique cohering focus for this post-disciplinary well-being research domain is proposed. In both trajectories, future research will have to increase cognizance of metatheoretical assumptions, develop more encompassing theories to bridge the conceptual fragmentation in the field, and implement methodological reforms, while keeping context and the interwovenness of the various levels of the scientific text in mind. Opportunities are indicated to contribute to the discourse on the identity and development of scientific knowledge in mainstream positive psychology and the evolving post-disciplinary domain of well-being studies.

scholarly journals Framework for Generating Configurable SAT Solvers

2011 ◽  
Vol 6 (1) ◽  
pp. 50-59
Author(s):  
Bernardo C. Vieira ◽  
Fabrício V. Andrade ◽  
Antônio O. Fernandes
Keyword(s):  
State Of The Art ◽  
The State ◽  
Equivalence Checking ◽  
Decision Heuristics ◽  
Sat Solvers ◽  
The Third ◽  
Sat Solver ◽  
Run Time

The state-of-the-art SAT solvers usually share the same core techniques, for instance: the watched literals structure, conflict clause recording and non-chronological backtracking. Nevertheless, they might differ in the elimination of learnt clauses, as well as in the decision heuristic. This article presents a framework for generating configurable SAT solvers. The proposed framework is composed of the following components: a Base SAT Solver, a Perl Preprocessor, XML files (Solver Description and Heuristics Description files) to describe each heuristic as well as the set of heuristics that the generated solver uses. This solvers may use several techniques and heuristics such as those implemented in BerkMin, and in Equivalence Checking of Dissimilar Circuits, and also in Minisat. In order to demonstrate the effectiveness of the proposed framework, this article also presents three distinct SAT solver instances generated by the framework to address a complex and challenging industry problem: the Combinational Equivalence Checking problem (CEC).The first instance is a SAT solver that uses BerkMin and Dissimilar Circuits core techniques except the learnt clause elimination heuristic that has been adapted from Minisat; the second is another solver that combines BerkMin and Minisat decision heuristics at run-time; and the third is yet another SAT solver that changes the database reducing heuristic at run-time. The experiments demonstrate that the first SAT solver generated is a faster solver than state-of-the-art SAT solver BerkMin for several instances as well as for Minisat in almost every instance.

Sign in / Sign up

Export Citation Format

Share Document