Report for Brainstorming Session at Higher Education & Research Advisory Council –
April 6th 2017
Editors: Silke Jakobi and Jean-Christophe Pazzaglia
Domain experts: Philip Miseldin and Smitha Rayala
Brainstormers: the HERAC community
“You’ll never manage to do that in one hour”
Latest trends across all industry are discussing how to deal with explosion of (big) data, how to connect people and devices, how to organize processes across geo zones, systems and organizations and how to optimize process using machine learning and/or artificial intelligence capabilities. The answer to these trends is a solution portfolio for Internet of things (IoT), artificial intelligence (AI) and machine learning (ML).
This is a trend which is valid for higher education & research as well but are universities and research institutes having their own use cases?
It was the aim of this Higher Education & Research Advisory Council (HERAC) 2017 interactive brainstorming session to find out about use cases for IoT, ML and AI especially in higher education & research institutions. To find out about their needs and their ideas how to support the daily business with Iot, ML and AI we asked the following questions:
- What are the business challenges that may be addressed with IoT technology?
- Think about your institution in 2022: which use cases will be deployed leveraging IoT capabilities?
- What are the business challenges that may be addressed with AI/ML technology?
- Think about your institution in 2022: which use cases will be deployed leveraging AI/ML capabilities?
We are looking for use cases which support the university in the term of how to run the university. Of course, some of these institutions do utilize IoT, AI, ML use cases in their research departments or they execute research projects in relation to IoT, AI and ML; but, as of today, there are only a few cases where these technologies do support the administration of education and research and the university in general.
To foster the creativity of our HER guests and to give the opportunity for everybody to contribute, we implement a variation of the Carousel Brainstorm process. Chart paper containing the questions to be addressed were posted in four different rooms.
Groups brainstormed for 10 minutes at one station and then rotated to the next position where they add additional comments. As new thoughts and ideas emerge, the list grows. When the carousel “stops” the coach took a picture and the summary of findings will be presented in a webinar and within this report.
Figure 1: The four stations
The Carousel Brainstorm provides an opportunity to use the group’s collective prior knowledge to further individual understanding. It is an active, users centered method for generating and sharing large amounts of data.
Figure 2: Evolution of ideas and clusters at the Monaco station
To implement efficiently the method, 4 groups were pre-determined and a colored sticker identified the group membership (eg. blue, green, red and white). Groups mixed genders and continents, people from the same organization were in different groups. A guide was guiding the different groups from station to station. Each station was equipped with a whiteboard, postit and the topic of the station. After the first iteration, the coach explained the different clusters identified and the ideas developed during the previous iterations. Then a 3 minutes silent brainstorm was used to collect new ideas that complemented the initial ideas.
As previously stated, the four station topics were :
- What are the business challenges that may be adressed with IoT technology?
- Think about your institution in 2022: which use cases will be deployed using IoT?
- What are the business challenges that may be adressed with AI/ML technology?
- Think about your institution in 2022: which use cases will be deployed levraging AI/ML capabilities?
These topics were related to the two preceding presentations:
- Designing for an AI & ML future – Philip Miseldine, SAP Global Design, Frontrunner Applications
- SAP Leonardo – Internet of Things platform Smitha Rayala, Product Manager for SAP Leonardo
What are the business challenges that may be addressed with IoT technology?
Figure 3: Final picture of the challenges generated at the D-Shop station
Six main clusters emerged from the brainstorming:
- People connectivity
- Research facilities management
- Assets Management
- Urban Campus
In these clusters, our participants do see the challenges to be addressed by IoT in connected student & staff – just in time lecture management, automation of attendance tracking or evacuation strategies, connected infrastructure and assets, connected fleet and transportation, smart city – smart campus and research (lab) facilities. E.g. there is the need to track student’s attendance automatically and prevent deception. Or the need to utilize all rooms and resources dynamically and sustainable. There is the need to track assets and inventory predictive and proactively. Also, researchers need to be able to monitor the lab equipment & goods for expiration & explosion and/or find underutilized equipment and make it available for others.
Think about your institution in 2022: which use cases will be deployed using IoT?
Figure 4: Final picture of the ideas generated at the Nice station
With the outlook in 2022 the participants do see quite many use cases which could be deployed leveraging IoT solutions. They defined use cases for such as
- Connected infrastructure: Universities and research institutes should know utilization of all rooms to plan dynamically and real time timetables. Instructors and researchers should be able to book a room ad hoc and have this insight every time and everywhere. Connected rooms and infrastructure should also support administration to work on security and evacuation plan.
- Energy saving and utilization: University staff would like analytics of room utilization to optimize energy usage and save costs. Especially facility management would like to organize the energy assignment for the rooms accordingly to utilization to make sure resources are used sustainable.
- In relation to connected infrastructure participants came up with further use cases such as garbage and waste control, building access control, etc.
- For connected assets there were use cases listed such as inventory and asset tracking plus predictive & proactive maintenance to avoid additional costs due to missed maintenance.
- Regarding connected fleet & logistics use cases were simple and nothing which cannot be provided today. Use cases are covering requirements related to campus transportation including E2E mobility integrating public transports, intelligent parking management, etc.
- Research lab facilities: research groups should be able to manage inventory, equipment and goods. They need to know every time and everywhere, how this equipment and rooms are used. The researcher needs a way to plan research experiments especially the utilization of equipment and goods. That research itself should always run smoothly and not be hindered by lack of resources.
What are the business challenges that may be addressed with AI/ML technology?
Figure 5: Final picture of the challenges generated at the station 3
Moving on to next question in relation to AI and ML. Seven main clusters emerged from the brainstorming:
- AI Powered Research (including management)
- Natural User Experience
- Innovative Pedagogy
- Proactive student success
- Curriculum management
- Compliance and Security
- Resources Optimization
Our participants do see challenges to be addressed by AI/ML around student performance and progression, research management or scheduling and organization. E.g. student would like to make sure that their selected courses and performance fit to their career plans. Or how can Universities utilize AI/ML capabilities to optimize research proposal processes as well as research collaboration? Universities need to schedule events and always assign automatically the appropriate resource and room to it
Think about your institution in 2022: which use cases will be deployed leveraging AI/ML capabilities?
Figure 5: Final picture of the ideas generated at the station 4
With the outlook in 2022 universities do see quite a lot of use cases which could be deployed leveraging AI/ML. They defined use cases for
- Research area: e.g. to be successful, researchers need to collaborate across research projects, get the optimal collaborators suggested and make sure it is an intelligent, hyperconnected and reciprocity research collaboration
- Curriculum planning: holistic planning and scheduling tool which considers the progression of every student, the availability of teachers and instructors and complete facility circumstances. If a new event is planned, only based on the content of the event, the appropriate room and resource should be suggested.
- Optimization in areas such as budget and planning, facility management and logistics by using predictive analytics and intelligent sharing tools.
- Natural UX like voice based chat consoles (Siiri, Alexa), chat bots integrated with any legacy system. Intelligent support of staff during their daily business and automatically improvement.
- Pedagogy: e.g. AI based questionnaires, utilize student’s notes in real time during the lecture. Use student’s note for intelligent evaluation, find out where students may need further insight or where lecture needs improvement.
- Student Services around admission advises, progression advises, employability advises. Monitoring student’s performance and success and providing personalized progression advise. Offer to the students an intelligent career developer mentor, evaluate hiring probability by industry.
Shortlist of Use Cases
In relation to IoT use cases the outcome shows that participants discussed challenges and use cases which are largely shared across industries and not only in higher education & research industry. These are use cases also discussed in concepts for smart city or in any organization who is dealing with infrastructure and assets.
As our participants were representing university IT organization and management we just may have not uncovered the real specific higher education & research use cases. For example, we have good reasons to believe that the management of large research infrastructures such as a large particle collider or a large array of radio telescope may be particularly demanding in term of reactiveness, number and type of sensors, number of generated events or effective algorithms for predictive maintenance. We also do believe that two 2022-scenarios may exhibit some HER differences and we encourage HERAC members to team to further investigate how IoT technology can be used to solve them:
Inventory management for technological equipment
Universities and research institutes need to administer and manage inventory, equipment and goods. They need to know every time and everywhere, how these equipment and rooms are used. This use case is focusing on researcher in the area of experimental research (e.g. medical, biology, chemistry, etc.)
The researcher needs a way to plan research experiments especially the utilization of equipment and goods. That research itself is always running smoothly and not hindering by lack of resources. Research should be informed if a required equipment and/or goods are running out or not available. E.g. there is expiring date researcher should be informed about it.
Connected users, dynamic room and utilization
Universities and research institutes need to know how all rooms are used to organize time tables and as well being able to book rooms containing specific equipment.
Professors and researchers need to have this insight every time and everywhere, as attendance may vary the tracking of individuals may enable to support ad-hoc optimization but also support administration to handle with emergency and provide optimal evacuation plan.
In relation to ML and AI, discussed challenges and use cases had a stronger focus on education and research and less on generic administration related areas. With the usage of predictive analytics universities are today deploying solutions utilizing AI capabilities. But there is a much bigger field of options for innovations and while AI/ML is becoming more mature some scenarios are still futuristic. We extracted 2 scenarios that we perceived as disruptive while we do believe that the technology and the data are nowadays available:
Individualized, outcome based, Curriculum Planning
Today, the students/their parents invest in their education because they think that a good education is a passport to get a better life. However, some concerns may happen if a student is not gifted for certain topics or if the job market does not provide enough employment in these topics. Historically, the dynamicity and the tension of the job market and the innovation pace were less demanding and curriculum had the time to adapt to fill the demand but today?
Can we use AI/ML technology to anticipate the fact that graduating on a specific topic will be an asset on the job market, can we recommend options and specialization to the student that fit the best his studying profile? To summarize, can we provide to the students an intelligent career developer mentor connected to the job market?
Optimization of your Research Work
To be successful, researchers need to collaborate across research projects while being acknowledge for their unique contribution. However, like in other domains, researchers are swamped by the data tsunami and similarly to what is happening in Social network, researchers may tend to be isolated in their projects or their research community.
How AI/ML can be used to detect optimal existing collaborations? Can automated document processing and social research network exploration enlarge the research network to adjacent fields or related problems where new research collaboration may be reciprocally beneficial? How to evaluate the pertinence to engage with new research partners on their fields based on their collaboration records and openness to cross domain fertilization?
To summarize, the answer to our original question “are universities and research institutes having their own use cases?”, based on the use cases that were identified, depends on the target technology. Higher Education and Research institutions are certainly in the position to leverage IoT technology, but the large majority of use cases appears similar to other industries. Their implementations will enable to offer new services or to optimize resources but they will only marginally change the students or professors experience. AI/ML clearly exhibit a more potential toward full-fledged digital transformation if the technology and the available data can deliver on the vision.
This brainstorming session showed clearly the willingness of institutions to innovate specially to drive their own digital transformation. To move forward three avenues can be envisaged:
- To implement the latest SAP technology to optimize resources or equipment usage, the SAP HER community may benefit here to join their forces to elicit the peculiar requirements and to share some extension to the standard solution;
- To co-innovate together with SAP on our most disruptive technology to develop ad-hoc use cases (eg. usage of e-quill for Professors);
- To work together on top of the SAP Platform – with the technical know-how of SAP – to leverage the users group collective intelligence (eg. in the mentoring of students), to evaluate the feasibility of the use cases based on a larger data set eager to provide better learning capabilities.
The last point will follow the “live what you preach” approach – university administration may be able to engage themselves on the IoT, AI, ML related research projects bringing exciting use cases to their data scientists or live data to their field experts. This midterm co-innovation may benefit to leverage SAP pragmatic innovation capabilities. This may well be an accelerator to their own digital transformation.