Crowdsourced Learning Research Articles

A visual fingerprint update algorithm based on crowdsourced localization and deep learning for smart IoV

Recently, deep learning and vision-based technologies have shown their great significance for the prospective development of smart Internet of Vehicle (IoV). When the smart vehicle enters the indoor parking of a shopping mall, the vision-based localization technology can provide reliable parking service. As known, the vision-based technique relies on a visual map without a change in the position of the reference object. Although, some researchers have proposed a few automatic visual fingerprinting (AVF) methods, which are aiming at reducing the cost of building the visual map database. However, the AVF method still costs too much under such a situation, since it is impossible to determine the specific location of the displaced object. Given the smart IoV and the development of deep learning approach, we propose an algorithm for solving the problem based on crowdsourcing and deep learning in this paper. Firstly, we propose a Region-based Fully Convolutional Network (R-FCN) based method with the feedback of crowdsourced images to locate the specific displaced object in the visual map database. Secondly, we propose a method based on quadratic programming (QP) for solving the translation vector of the displaced objects, which finally solves the problem of updating the visual map database. The simulation results show that our method can provide a higher detection sensitivity and correction accuracy as well as the relocation results. It means that our proposed algorithm outperforms the compared one, which is verified by both synthetic and real data simulation.

Data-Oriented State Space Discretization for Crowdsourced Robot Learning of Physical Skills

AbstractThis work discusses a crowdsourced learning scheme for robot physical intelligence. Using a large amount of data from crowdsourced mentors, the scheme allows robots to synthesize new physical skills that are never demonstrated or only partially demonstrated without heavy re-training. The learning scheme features a data management method to sustainably manage continuously collected data and a growing knowledge library. The method is validated using a simulated challenge of solving a bottle puzzle. The learning scheme aims at realizing ubiquitous robot learning of physical skills and has the potential of automating many demanding tasks that are currently hard to robotize.

Robotic dismembered pyeloplasty surgical simulation using a 3D-printed silicone-based model: development, face validation and crowdsourced learning outcomes assessment.

Ureteropelvic junction obstruction (UPJO) is an uncommonly encountered pathology, posing a challenge for resident training. We describe the development and face validation of a robotic pyeloplasty simulation using a 3D-printed silicone-based model of UPJO for surgical training, in combination with crowdsourced scoring to objectively assess performance and learning outcomes. The organs were created using 3D modeling software and printed using a silicone-based material by Lazarus 3D, LLC. They were secured in a laparoscopic box trainer and the robotic system was docked. Eight residents and three faculty each performed two robotic-assisted right dismembered pyeloplasties on separate occaisions. Face validity was evaluated on a 5-point Likert scale. Crowd-Sourced Assessment of Technical Skills (C-SATS Inc.) scored surgical performance using the Global Evaluative Assessment of Robotic Skills (GEARS) criteria, based on video review of each simulation. All participants completed the simulation twice with fully patent anastomoses. Average time to complete the first and second trials was 44.4min and 43.2min, respectively. The average GEARS score was 17.1 and 17.6 for the first and second trials respectively. Participants improved on average in all 5 GEARS categories, with significant improvement in depth perception (p = 0.006). The model received mean scores (out of 5) of 4.36 for aesthetics, 4.18 for overall feel, 3.55 for realism, 4.72 for usability, and 4.72 for suturability. Residents had a significant increase in confidence between initial and final surveys on a 5-point Likert Scale: 1.63 vs. 2.38 (p = 0.03). Using 3D-printed silicone-based models, participants completed robotic-assisted dismembered pyeloplasties for training and skill acquisition. We demonstrated face validity of the simulation, which was also found to improve participant speed and significantly improve resident confidence. Crowdsourced assessment demonstrated significant improvement in depth perception.

RiPPLE: A Crowdsourced Adaptive Platform for Recommendation of Learning Activities

This paper presents a platform called RiPPLE (Recommendation in Personalised Peer-Learning Environments) that recommends personalized learning activities to students based on their knowledge state from a pool of crowdsourced learning activities that are generated by educators and the students themselves. RiPPLE integrates insights from crowdsourcing, learning sciences, and adaptive learning, aiming to narrow the gap between these large bodies of research while providing a practical platform-based implementation that instructors can easily use in their courses. This paper provides a design overview of RiPPLE, which can be employed as a standalone tool or embedded into any learning management system (LMS) or online platform that supports the Learning Tools Interoperability (LTI) standard. The platform has been evaluated based on a pilot in an introductory course with 453 students at The University of Queensland. Initial results suggest that the use of the RiPPLE platform led to measurable learning gains and that students perceived the platform as beneficially supporting their learning.

Studying in the ‘Bazaar’: An Exploratory Study of Crowdsourced Learning in GitHub

In recent years, GitHub, the most popular social coding site, has been increasingly employed for managing learning content, sharing knowledge, imparting experience, and requesting and contributing learning resources in a crowdsourced way. We thus refer to this type of e-learning practices in which people perform learning-related tasks with open calls in an online community to gain knowledge or skills in specific areas as crowdsourced learning (CL). To understand this emerging phenomenon in GitHub, we investigate the popularity of the learning projects and learners' CL activities, first on the extracted GHTorrent learning projects and then on a selected sample of 105 popular learning projects. We then conduct an online survey of 301 learners to qualitatively understand their practices and perceptions of CL in GitHub. Our main findings reveal that the learners' CL practices show some different characteristics from those of open source development, e.g., the learning projects have very few long-term contributors (less than 5% of all contributors). Moreover, although the learners benefit from conducting personalized learning with the high-quality content and contributions made by the community, they encounter challenges in maintaining initiatives to conduct continuous unsupervised learning and in ensuring the quality of content and contributions. Based on the findings, we discuss the reasons behind the growth of CL and provide implications on the CL platform design.

Learnsourcing: Is it working or failing, and where to go from here?

ABSTRACTOnline learning has taken great leaps forward with the emergence of Web 2.0 and has paved the path for learning technologies that can be conveniently accessed by users provided they are connected to the Internet. One of the most profitable learning techniques, “Learning from the crowd,” or augmenting “collective intelligence,” has attracted the attention of several learners who seek information in terms of knowledge access. This provides an admirable environment for learning with peers rather than learning alone. However, having access to these online learning technologies does not simply give the users the chance to optimize learning. In addition, researchers have expressed ethical concerns related to issues like accessibility, quality, authenticity, and reliability of content. Taking into consideration the pluses and minuses of “crowdsourced learning,” each of the panelists in this program will shed light on how to use the “wisdom of crowds” or “collective intelligence” to better articulate the process of learning in context of such ethical issues. Each of the panelists comes from a different field of expertise (Information Science, Library Science, e‐Learning, Learning Analytics, and Education). This diversity of experience will be reflected in the different perspectives presented in the area of crowdsourcing in education. The session will conclude with the panelists' presentation of significant comments that would emerge in the discussion.

Understanding Learning and Learning Design in MOOCs: A Measurement-Based Interpretation

The paper describes empirical investigations of how participants in a MOOC learn, and the implications for MOOC design. A learner capability to generate higher order learning in MOOCs — called crowd-sourced learning (C-SL) capability — was defined from learning science literature. The capability comprised a complex yet interrelated array of attitudes, beliefs, and understandings about learning that participants bring to a MOOC and which shape their behaviour and explain why individuals differ in their ability to generate higher order learning. The capability was formulated as a developmental progression describing behaviours associated with five levels, from novice to expert, charting the degree to which learners regulate their own learning, effectively exploit the scale and diversity of MOOCs, and harness opportunities for distributed teaching. Item response theory was applied to log stream data in two MOOCs to construct empirically validated measures of this capability, enabling each MOOC learner to be assessed for learning capability on a scale from novice to expert. The majority of participants did not behave in ways conducive to the generation of higher order learning, but the C-SL progression suggested principles to guide MOOC design to make them more efficacious, which, when empirically investigated, were found to be efficacious.

Mobile and Accessible Learning for MOOCs

Many modern web-based systems provide a ‘responsive’ design that allows material and services to be accessed on mobile and desktop devices, with the aim of providing ubiquitous access. Besides offering access to learning materials such as podcasts and videos across multiple locations, mobile, wearable and ubiquitous technologies have some additional affordances that may enable new forms of learning on MOOCs. We can divide these into two categories: firstly, context-sensitive features including delivery of content for a specific location, enabling a seamless continuity of learning across settings, and linking people in a location with others in a virtual representation of that place; secondly social learning opportunities to connect people as they move within and across locations, to enable crowd-sourced learning. In this paper we explore these aspects of mobile and accessible learning for MOOCs, drawing on examples from MOOC courses, mobile toolkits, and crowd-sourced learning sites.