Filters








57,159 Hits in 7.2 sec

How Many Workers to Ask? Adaptive Exploration for Collecting High Quality Labels [article]

Ittai Abraham, Omar Alonso, Vasilis Kandylas, Rajesh Patel, Steven Shelford, Aleksandrs Slivkins
2016 arXiv   pre-print
We conduct a data analysis on an industrial crowdsourcing platform, and use the observations from this analysis to design new stopping rules that use the workers' quality scores in a non-trivial manner  ...  In this paper we investigate how to devise better stopping rules given such quality scores.  ...  We investigate how these quality scores can help in building better stopping rules.  ... 
arXiv:1411.0149v3 fatcat:ydtzhv7vjzhmva2uwejqpntamq

How Many Workers to Ask?

Ittai Abraham, Omar Alonso, Vasilis Kandylas, Rajesh Patel, Steven Shelford, Aleksandrs Slivkins
2016 Proceedings of the 39th International ACM SIGIR conference on Research and Development in Information Retrieval - SIGIR '16  
We conduct a data analysis on an industrial crowdsourcing platform, and use the observations from this analysis to design new stopping rules that use the workers' quality scores in a non-trivial manner  ...  The difficulty of the HIT decides how much weight to give to each worker's answer. In this paper we investigate how to devise better stopping rules given workers' performance quality scores.  ...  We investigate how these quality scores can help in building better stopping rules.  ... 
doi:10.1145/2911451.2911514 dblp:conf/sigir/AbrahamAKPSS16 fatcat:kjiw4asb4rexhe5xauxrxayvhq

Tell Me Something New: A New Framework for Asynchronous Parallel Learning [article]

Julaiti Alafate, Yoav Freund
2018 arXiv   pre-print
This approach involves a set of independent workers that use broadcast to update each other when they observe "something new".  ...  We demonstrate the utility of TMSN by applying it to learning boosted trees. We show that our implementation is 10 times faster than XGBoost and LightGBM on the splice-site prediction problem.  ...  The worker computes a "performance score" z t+1 which is an upper bound on the Z-score the strong rule by adding the weak rule to it. The pair (H t+1 , z t+1 ) is broadcast to the other workers (i).  ... 
arXiv:1805.07483v2 fatcat:vf4doqbdird77mdwrq5hl24foa

Toward Effective Automated Content Analysis via Crowdsourcing [article]

Jiele Wu, Chau-Wai Wong, Xinyan Zhao, Xianpeng Liu
2021 arXiv   pre-print
Many computer scientists use the aggregated answers of online workers to represent ground truth.  ...  We observe that with timely feedback on workers' performance quantified by quality scores, better informed online workers can maintain the quality of their labeling throughout an extended period of time  ...  Each input tweet is cleaned using standard NLP procedures to remove stop words and punctuation.  ... 
arXiv:2101.04615v2 fatcat:r3k4qcxw35epdfffpze4tck7va

Becoming a high-performance work organization: the role of security, employee involvement and training

Clair Brown, Michael Reich, David Stern
1993 International Journal of Human Resource Management  
We find that SET systems are difficult to implement in a gradual and partial manner.  ...  We find that SET systems are difficult to implement in a gradual and partial manner.  ...  The union and company have an adversarial relationship in which the union uses the grievance system, work rules, and work stoppage as its basis of power, while the company uses worker discipline, automation  ... 
doi:10.1080/09585199300000015 fatcat:rs7ojqtg6vacrcuj3ml4dncieu

UNDERSTANDING CONSTRUCTION SITE SAFETY HAZARDS THROUGH OPEN DATA: TEXT MINING APPROACH

Neththi Kumara Appuhamilage Heshani Rupasinghe, Kriengsak Panuwatwanich
2021 ASEAN Engineering Journal  
The finding using the developed tool identified the worker factor as the highest contributor to construction site accidents followed by technological factor, surrounding activities, organisational factor  ...  This paper presents a methodology to develop a rule-based extraction tool by providing full details of lexicon building, devising extraction rules and the iterative process of testing and validation.  ...  The F1 score obtained through rule-based model was compared with the existing statistical classifiers by using the final rule-based model.  ... 
doi:10.11113/aej.v11.17871 fatcat:yboymmpk5ncfzeh6cwugceptuu

Crowdsourced Data Management: A Survey

Guoliang Li, Jiannan Wang, Yudian Zheng, Michael J. Franklin
2016 IEEE Transactions on Knowledge and Data Engineering  
We identify three important problems in crowdsourced data management. (1) Quality Control: Workers may return noisy or incorrect results so effective techniques are required to achieve high quality; (2  ...  Based on this analysis we then outline key factors that need to be considered to improve crowdsourced data management.  ...  They estimate how much result quality can be improved by asking the crowd to do more tasks. If the improved quality is not satisfactory, it will stop sending tasks to the crowd, thus saving the cost.  ... 
doi:10.1109/tkde.2016.2535242 fatcat:sit3comyvra4djqkl4xre3kj24

Does Confidence Reporting from the Crowd Benefit Crowdsourcing Performance?

Qunwei Li, Pramod K. Varshney
2017 Proceedings of the 2nd International Workshop on Social Sensing - SocialSens'17  
We present an aggregation approach using a weighted majority voting rule, where each worker's response is assigned an optimized weight to maximize crowd's classification performance.  ...  We consider the scenario where the workers have a reject option so that they are allowed to skip microtasks when they are unable to or choose not to respond to binary microtasks.  ...  con dence scores to improve the quality of crowdsourced labels was investigated in [11] .  ... 
doi:10.1145/3055601.3055607 dblp:conf/cpsweek/LiV17 fatcat:44rp5d4l6vei5enqqdt56kxfh4

Construction Workers Perceptions Toward Safety Culture

Andi Andi
2008 Civil Engineering Dimension  
Further analysis indicates that workers in the three projects have different safety culture perceptions, especially on factors of top management commitment, safety rules and procedures, communication,  ...  Data were obtained through questionnaire survey to three large construction projects in Surabaya. Two hundreds and seven sets of questionnaires were gathered and used for subsequent analyses.  ...  Acknowledgment The author wishes to thank Aditya Chandra for the permission to use the data from his thesis [16] for analyses in this paper.  ... 
doaj:7e3d723848344418822824b21933e5c1 fatcat:kp36nbobfjg5layogtvdutwzfu

Determinants of Sense of Coherence in Male Japanese Workers: a Cross-Sectional Study

SAYAKO KOMATSU, HIROAKI ITOH, KAYOKO URAKAWA, KAZUHITO YOKOYAMA
2016 Juntendo Medical Journal  
Workers in their 40s and 50s had 2-3 points higher adjusted mean SOC scores than did workers aged 20-39 years. However, the most influential factor for SOC was sleep quality.  ...  The objective of the present study is to investigate the determinants of SOC, including age, in male Japanese workers.  ...  Second, we cannot rule out the possibility of measurement error due to the use of self-reported SOC, which may have influenced the lack of observed associations in several cases.  ... 
doi:10.14789/jmj.62.318 fatcat:wqa54pqhhrcc5pg3d7lecy7m64

Detecting low-quality crowdtesting workers

Ricky K. P. Mok, Weichao Li, Rocky K. C. Chang
2015 2015 IEEE 23rd International Symposium on Quality of Service (IWQoS)  
In this paper, we propose to use finer-grained cursor trajectory analysis, including submovement analysis, to identify low quality workers.  ...  A number of methods, including question design, instructions, and human inspections, are used to label workers into three categories.  ...  Violating these "hard" rules can lead to a rejection of his work or stopping him from proceeding to the next assessment.  ... 
doi:10.1109/iwqos.2015.7404734 dblp:conf/iwqos/MokLC15 fatcat:6odnxyti6re7hntbmrwpendz7y

Mystery Shopping and Well-being of Service Workers in South Korea

Heeju Shin
2019 Safety and Health at Work  
Mystery shopping is a method in which a company monitors quality of service and employee conduct and compliance with regulations using an evaluator posing as a customer.  ...  Data from semistructured interviews with 15 workers were collected from January to April 2019 to obtain information on service worker experiences with mystery shopping.  ...  Conflicts of interest The research entitled "Mystery Shopping and Health of Service Workers in South Korea" was supported by 2018 Korean Institute of Labor Safety and Health Research Fund.  ... 
doi:10.1016/j.shaw.2019.10.001 pmid:31890331 pmcid:PMC6933262 fatcat:h2sfouv4hbhknf4lp6x6mitvsq

Active Learning with Efficient Feature Weighting Methods for Improving Data Quality and Classification Accuracy

Justin Martineau, Lu Chen, Doreen Cheng, Amit Sheth
2014 Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)  
to significantly improve annotation quality at low cost.  ...  In this paper we study a large, low quality annotated dataset, created quickly and cheaply using Amazon Mechanical Turk to crowdsource annotations.  ...  We applied substantial quality control to our AMT workers to improve the initial quality of annotation following the common practice of crowdsourcing.  ... 
doi:10.3115/v1/p14-1104 dblp:conf/acl/MartineauCCS14 fatcat:qmsrglgvqngepic7eipojldnby

The Quality of Tuberculosis Services in Health Care Centres in a Rural District in Uganda: The Providers' and Clients' Perspective

Lilian Bulage, Juliet Sekandi, Omar Kigenyi, Ezekiel Mupere
2014 Tuberculosis Research and Treatment  
A high number 150 (87.21%) of TB patients were not aware of all the signs to stop TB medication, and 100 (25.51%) patients received laboratory results after a period of 3–5 working days.  ...  Data was obtained using face-to-face interviews, an observation guide, a check list, and record review of the TB unit and laboratory registers.  ...  A similar approach to quality improvement was used in a TB program in Cape Town, South Africa [17] .  ... 
doi:10.1155/2014/685982 pmid:25276424 pmcid:PMC4170836 fatcat:r252eq7tm5bwhdlfqvbnivvzmq

Behavior-Based Safety and Working Alone

Ryan Olson, John Austin
2001 Journal of organizational behavior management  
It would also be useful to study lone workers for more extended periods of time.  ...  These selfreport measures corresponded to actual improvements where complete stop realized the largest improvement of 20.8 percent (Likert score, 4.25) and bus stopping position was the least improved  ...  The goal is to actively work on improving the behaviors on the checklist. Run number: -------- COMPLETE STOP Wheels of the bus completely stationary at all stops signs and red traffic lights.  ... 
doi:10.1300/j075v21n03_02 fatcat:57uxxuy2mzehnkkrpyr27nupgi
« Previous Showing results 1 — 15 out of 57,159 results