skip to main content

Title: BiGBERT: Classifying Educational Web Resources for Kindergarten-12𝑡ℎ Grades
In this paper, we present BiGBERT, a deep learning model that simultaneously examines URLs and snippets from web resources to determine their alignment with children's educational standards. Preliminary results inferred from ablation studies and comparison with baselines and state-of-the-art counterparts, reveal that leveraging domain knowledge to learn domain-aligned contextual nuances from limited input data leads to improved identification of educational web resources.
Authors:
Award ID(s):
1763649
Publication Date:
NSF-PAR ID:
10337092
Journal Name:
European Conference on Information Retrieval (ECIR)
Page Range or eLocation-ID:
176-184
Sponsoring Org:
National Science Foundation
More Like this
  1. The DeepLearningEpilepsyDetectionChallenge: design, implementation, andtestofanewcrowd-sourced AIchallengeecosystem Isabell Kiral*, Subhrajit Roy*, Todd Mummert*, Alan Braz*, Jason Tsay, Jianbin Tang, Umar Asif, Thomas Schaffter, Eren Mehmet, The IBM Epilepsy Consortium◊ , Joseph Picone, Iyad Obeid, Bruno De Assis Marques, Stefan Maetschke, Rania Khalaf†, Michal Rosen-Zvi† , Gustavo Stolovitzky† , Mahtab Mirmomeni† , Stefan Harrer† * These authors contributed equally to this work † Corresponding authors: rkhalaf@us.ibm.com, rosen@il.ibm.com, gustavo@us.ibm.com, mahtabm@au1.ibm.com, sharrer@au.ibm.com ◊ Members of the IBM Epilepsy Consortium are listed in the Acknowledgements section J. Picone and I. Obeid are with Temple University, USA. T. Schaffter is with Sage Bionetworks, USA. E. Mehmetmore »is with the University of Illinois at Urbana-Champaign, USA. All other authors are with IBM Research in USA, Israel and Australia. Introduction This decade has seen an ever-growing number of scientific fields benefitting from the advances in machine learning technology and tooling. More recently, this trend reached the medical domain, with applications reaching from cancer diagnosis [1] to the development of brain-machine-interfaces [2]. While Kaggle has pioneered the crowd-sourcing of machine learning challenges to incentivise data scientists from around the world to advance algorithm and model design, the increasing complexity of problem statements demands of participants to be expert data scientists, deeply knowledgeable in at least one other scientific domain, and competent software engineers with access to large compute resources. People who match this description are few and far between, unfortunately leading to a shrinking pool of possible participants and a loss of experts dedicating their time to solving important problems. Participation is even further restricted in the context of any challenge run on confidential use cases or with sensitive data. Recently, we designed and ran a deep learning challenge to crowd-source the development of an automated labelling system for brain recordings, aiming to advance epilepsy research. A focus of this challenge, run internally in IBM, was the development of a platform that lowers the barrier of entry and therefore mitigates the risk of excluding interested parties from participating. The challenge: enabling wide participation With the goal to run a challenge that mobilises the largest possible pool of participants from IBM (global), we designed a use case around previous work in epileptic seizure prediction [3]. In this “Deep Learning Epilepsy Detection Challenge”, participants were asked to develop an automatic labelling system to reduce the time a clinician would need to diagnose patients with epilepsy. Labelled training and blind validation data for the challenge were generously provided by Temple University Hospital (TUH) [4]. TUH also devised a novel scoring metric for the detection of seizures that was used as basis for algorithm evaluation [5]. In order to provide an experience with a low barrier of entry, we designed a generalisable challenge platform under the following principles: 1. No participant should need to have in-depth knowledge of the specific domain. (i.e. no participant should need to be a neuroscientist or epileptologist.) 2. No participant should need to be an expert data scientist. 3. No participant should need more than basic programming knowledge. (i.e. no participant should need to learn how to process fringe data formats and stream data efficiently.) 4. No participant should need to provide their own computing resources. In addition to the above, our platform should further • guide participants through the entire process from sign-up to model submission, • facilitate collaboration, and • provide instant feedback to the participants through data visualisation and intermediate online leaderboards. The platform The architecture of the platform that was designed and developed is shown in Figure 1. The entire system consists of a number of interacting components. (1) A web portal serves as the entry point to challenge participation, providing challenge information, such as timelines and challenge rules, and scientific background. The portal also facilitated the formation of teams and provided participants with an intermediate leaderboard of submitted results and a final leaderboard at the end of the challenge. (2) IBM Watson Studio [6] is the umbrella term for a number of services offered by IBM. Upon creation of a user account through the web portal, an IBM Watson Studio account was automatically created for each participant that allowed users access to IBM's Data Science Experience (DSX), the analytics engine Watson Machine Learning (WML), and IBM's Cloud Object Storage (COS) [7], all of which will be described in more detail in further sections. (3) The user interface and starter kit were hosted on IBM's Data Science Experience platform (DSX) and formed the main component for designing and testing models during the challenge. DSX allows for real-time collaboration on shared notebooks between team members. A starter kit in the form of a Python notebook, supporting the popular deep learning libraries TensorFLow [8] and PyTorch [9], was provided to all teams to guide them through the challenge process. Upon instantiation, the starter kit loaded necessary python libraries and custom functions for the invisible integration with COS and WML. In dedicated spots in the notebook, participants could write custom pre-processing code, machine learning models, and post-processing algorithms. The starter kit provided instant feedback about participants' custom routines through data visualisations. Using the notebook only, teams were able to run the code on WML, making use of a compute cluster of IBM's resources. The starter kit also enabled submission of the final code to a data storage to which only the challenge team had access. (4) Watson Machine Learning provided access to shared compute resources (GPUs). Code was bundled up automatically in the starter kit and deployed to and run on WML. WML in turn had access to shared storage from which it requested recorded data and to which it stored the participant's code and trained models. (5) IBM's Cloud Object Storage held the data for this challenge. Using the starter kit, participants could investigate their results as well as data samples in order to better design custom algorithms. (6) Utility Functions were loaded into the starter kit at instantiation. This set of functions included code to pre-process data into a more common format, to optimise streaming through the use of the NutsFlow and NutsML libraries [10], and to provide seamless access to the all IBM services used. Not captured in the diagram is the final code evaluation, which was conducted in an automated way as soon as code was submitted though the starter kit, minimising the burden on the challenge organising team. Figure 1: High-level architecture of the challenge platform Measuring success The competitive phase of the "Deep Learning Epilepsy Detection Challenge" ran for 6 months. Twenty-five teams, with a total number of 87 scientists and software engineers from 14 global locations participated. All participants made use of the starter kit we provided and ran algorithms on IBM's infrastructure WML. Seven teams persisted until the end of the challenge and submitted final solutions. The best performing solutions reached seizure detection performances which allow to reduce hundred-fold the time eliptologists need to annotate continuous EEG recordings. Thus, we expect the developed algorithms to aid in the diagnosis of epilepsy by significantly shortening manual labelling time. Detailed results are currently in preparation for publication. Equally important to solving the scientific challenge, however, was to understand whether we managed to encourage participation from non-expert data scientists. Figure 2: Primary occupation as reported by challenge participants Out of the 40 participants for whom we have occupational information, 23 reported Data Science or AI as their main job description, 11 reported being a Software Engineer, and 2 people had expertise in Neuroscience. Figure 2 shows that participants had a variety of specialisations, including some that are in no way related to data science, software engineering, or neuroscience. No participant had deep knowledge and experience in data science, software engineering and neuroscience. Conclusion Given the growing complexity of data science problems and increasing dataset sizes, in order to solve these problems, it is imperative to enable collaboration between people with differences in expertise with a focus on inclusiveness and having a low barrier of entry. We designed, implemented, and tested a challenge platform to address exactly this. Using our platform, we ran a deep-learning challenge for epileptic seizure detection. 87 IBM employees from several business units including but not limited to IBM Research with a variety of skills, including sales and design, participated in this highly technical challenge.« less
  2. Society’s increased reliance on technology has simultaneously increased the demand for people who can develop and design these new advancements. This has led to an influx of students looking to learn how to code and gain the technological skill set that is currently among the most marketable. Learning to code is challenging; without the right tools, resources, and assistance, it can be tough to build the foundation needed to understand key computer science fundamentals. The existing web platforms focused on assisting K-12 learners are competitive from an educational and technical perspective. There is a huge lack of virtual educational platformsmore »that can deliver resources to students with disabilities through innovative accessible features and provide guidance to K-12 teachers that are trying to support this area. This lack of guidance is especially evident when examining resources available to teachers about increasing access and engagement of struggling learners including students with disabilities. The motivation of this paper is to introduce a prototype of a centralized portal, Accessible Virtual Learning, that implements user experience strategies and accessible usability principles aiming to be accessible to any student and also educators who need guidance on finding suitable materials. The success of this portal relies heavily on its ability to allow teachers and self-directed learners to facilitate curriculums effectively while maximizing student engagement, ease of learning, and digital assistance for students at various ages with different learning abilities, both physical and cognitive.« less
  3. Over the past two decades, I have been actively involved in teaching astronomy and astrophysics to Chicago Public School (CPS) students and their teachers, in collaboration with various groups as well as by myself. Valuable resources that we have created for schools include the Multiwavelength Astronomy Website, with modules for infrared, optical, ultraviolet, X-ray and gamma-ray astronomy. The content of each lesson is derived from interviews with scientists, archived oral histories, and/or memoirs. Lessons are evaluated by a science educator and at least one subject matter expert before being produced for the web. They are supplemented by NASA media, archivalmore »material from the University of Chicago Library and other archives, and participant contributed photographs, light curves, and spectra. Summer programs provided training to CPS teachers to use the resources in their classrooms. Currently, I lead the Chicago Area Research Mentoring (CHARM) initiative. In the past academic year I worked with a class of 17 diverse 11th grade honors students at the University of Chicago Charter School, Woodlawn. Through frequent lectures (∼ every 4 weeks), these students were exposed to astrophysical topics and concepts not normally covered in a school curriculum. CHARM aims to develop the student's critical thinking, introduce them to astrophysical research methods and techniques, and prepare them for a career in science, technology, engineering and mathematics (STEM), particularly a research-oriented one. In this article, I highlight some projects, educational resources, results achieved, and lessons learned along the way.« less
  4. Miami Dade College, based on demands in the workforce, has focused on creating emerging technology education, particularly Cloud technology. Business Cloud migration has accelerated over the last several years, and companies around the world are investing in their future with the cloud. With the increased demand for cloud-skilled professionals the last four years, we launched a cloud literacy initiative to meet cloud talent needs. This initiative aims to provide our students in the computing/IT fields with the knowledge, abilities and skills needed to accelerate their cloud-related learning. With the support of NSF ATE, we collaborated with Amazon Web Services (AWS)more »to create a new pathway for the next generation of cloud computing professionals. The course sequence was designed in conjunction with an AWS Educate team assisting in the design of course sequencing and degree plans to leverage their educational experience in teaching cloud technologies. The core curriculum designed for the cloud literacy initiative leveraged an existing pathway for an associate degree in networking technology, and then partially pulled classes from the design of the bachelor’s in information systems technology degree. The classes identified, used current offerings across our programs and included a focus on the supporting infrastructure of cloud systems: Databases, Linux OS, and Networking. With these three foundational classes that were cloudified, three cloud-focused courses based on industry certifications were developed: Cloud Essentials for AWS Cloud Practitioner and Cloud Infrastructure and Services for AWS Solutions Architect certifications with a capstone class completing the academic pathway. This new curriculum includes pedagogical changes to utilize project-based learning by incorporating resources and learning from multiple sources to best mimic real-world application, data, and design attributes. In addition to the associate degree, a college credit certificate in cloud computing was created to strengthen (re-)entering students in the workforce and dual enrollment students with credentials and employability skills by using high impact educational practices. Our cloud curriculum incorporates project-based learning approach, a real-world experience using the cloud technology. This poster shares strategies and pedagogical tools for teaching a cloud-focused curriculum for broader impact and student success.« less
  5. National Science Foundation (NSF) funded Engineering Research Centers (ERC) must complement their technical research with various education and outreach opportunities to: 1) improve and promote engineering education, both within the center and to the local community; 2) encourage and include the underrepresented populations to participate in Engineering activities; and 3) advocate communication and collaboration between industry and academia. ERCs ought to perform an adequate evaluation of their educational and outreach programs to ensure that beneficial goals are met. Each ERC has complete autonomy in conducting and reporting such evaluation. Evaluation tools used by individual ERCs are quite similar, but eachmore »ERC has designed their evaluation processes in isolation, including evaluation tools such as survey instruments, interview protocols, focus group protocols, and/or observation protocols. These isolated efforts resulted in redundant resources spent and lacking outcome comparability across ERCs. Leaders from three different ERCs led and initiated a collaborative effort to address the above issue by building a suite of common evaluation instruments that all current and future ERCs can use. This leading group consists of education directors and external evaluators from all three partners ERCs and engineering education researchers, who have worked together for two years. The project intends to address the four ERC program clusters: Broadening Participation in Engineering, Centers and Networks, Engineering Education, and Engineering Workforce Development. The instruments developed will pay attention to culture of inclusion, outreach activities, mentoring experience, and sustained interest in engineering. The project will deliver best practices in education program evaluation, which will not only support existing ERCs, but will also serve as immediate tools for brand new ERCs and similar large-scale research centers. Expanding the research beyond TEEC and sharing the developed instruments with NSF as well as other ERCs will also promote and encourage continual cross-ERC collaboration and research. Further, the joint evaluation will increase the evaluation consistency across all ERC education programs. Embedded instrumental feedback loops will lead to continual improvement to ERC education performance and support the growth of an inclusive and innovative engineering workforce. Four major deliveries are planned. First, develop a common quantitative assessment instrument, named Multi-ERC Instrument Inventory (MERCII). Second, develop a set of qualitative instruments to complement MERCII. Third, create a web-based evaluation platform for MERCII. Fourth, update the NSF ERC education program evaluation best practice manual. These deliveries together will become part of and supplemented by an ERC evaluator toolbox. This project strives to significantly impact how ERCs evaluate their educational and outreach programs. Single ERC based studies lack the sample size to truly test the validity of any evaluation instruments or measures. A common suite of instruments across ERCs would provide an opportunity for a large scale assessment study. The online platform will further provide an easy-to-use tool for all ERCs to facilitate evaluation, share data, and reporting impacts.« less