In recent years, the research and development work that has been underway at Bolton College has led to the creation of a number of AIED products and services; notably Ada, the College's campus digital assistant for students, teachers and campus support teams; and FirstPass which offers real-time feedback to students as they respond to open-ended questions. In time, other campus services such as Signal (the College's communications platform), Moodle, student report cards, adaptive online tutorials and more will start to assimilate AIED services to enhance the offer to our students. The assets that are developed from these services will invariably cascade to other parts of the campus and to the student life cycle as a whole. If this trend continues, one can postulate that over a period of time, cognitive computing will come to pervade every digital service that is present in our campuses and it will be present in all digital services yet to come.
If cognitive computing is present in every digital service that students or teachers touch; and if we see technology as a human activity, the digital service will itself be shaped and altered by these interactions. This is one of the defining traits of AIED services; and one that sets them apart from legacy EdTech services. For example, as teachers label sentences in FirstPass it enhances the platforms ability to classify or label student responses. Likewise, as students respond to open-ended questions in FirstPass their responses can be added to the library of classified text; furthering the ability for FirstPass to offer real-time and contextualised feedback to the students who make use of the service.
AIED services are nurtured through communal practice. In the case of FirstPass, Business Studies teachers across the UK and the US could train a classifier about good customer care. As hundreds or thousands of students in the UK or the US respond to a question about the merits of good customer care, their answers will be labelled by FirstPass and their responses could be added to the respective classifier. Over a period of time as teachers and students continue to interact with this classifier, the more capable it becomes in supporting the formative assessment process.
The emergence of Ada, FirstPass and the services that underpin them have come about through the acquisition of new knowledge and through a participatory model. Both are required to stimulate learning at an institutional level within an education setting. In natural language processing there is a requirement to use a collection of annotated text to train machine learning models. In the case of Bolton College’s Ada service, subject specialists are required to author high-quality question and answer pairs. If Ada is to respond in a reliable and consistent manner to student and teacher questions there is a dependency on having a large volume of well-trained question and answer pairs across a wide set of knowledge domains about the institution, campus services and subject topics. With regard to Bolton College’s FirstPass service, there is a dependency on having high quality labelled data to support the formative assessment of student work.
The participants in projects such as Ada and FirstPass require an appropriate form of motivation to achieve a desired set of goals. For colleagues at Bolton College and elsewhere, these will invariably be to improve the student experience, to support student wellbeing, to support students as they journey through their studies, to raise attainment levels and so forth. These motivations are well placed to support the participatory model that sustain the Ada and FirstPass projects.
A participatory process is important when crowdsourcing labelled data; however, project teams need to guard against deteriorating model accuracy as a larger group of individuals are involved in the project. However, repeated labelling by multiple teachers and campus support teams leads to an improvement in model accuracy; especially with regard to label inference. Label inference takes place when a natural language processing or classification model has acquired a sufficient volume of annotated data that allows the model to infer a label to unseen unlabelled text. In the case of Bolton College’s Ada service, the digital assistant can respond correctly to unseen questions posed by students or staff simply because there is a high volume of labelled training data for the knowledge domain at hand. The same holds true for FirstPass when the service is able to infer a label to a sentence that the natural language classification model has not seen before.
The means of production for education services such as Ada, Ada Goes To School and FirstPass are dependent on leveraging the strengths that are inherent in large numbers of people who fulfil tasks that were previously carried out by the few. The participatory model lends itself particularly well to the education sector were the larger group are motivated towards shared goals. If this is the case, AIED services can indeed be developed with a communal spirit. I personally welcome this spirit of cooperation, participation and collaboration.