This site requires Javascript to be turned on. Please enable Javascript and reload the page.

AI for Teachers, An Open Textbook: Edition 1

X5GON

Searching for educational resources is an important activity for teachers: when preparing a new lecture, exploring a new field or subject, or preparing some activity. This material (courseware) can be just for documentation, but sometimes the teacher might want to build new courseware from it; And it is tempting and intellectually legitimate to not reinvent the wheel and to use an intelligent form of copy-paste. Of course this is usually not legal in most countries as it breaks copyright laws.

When the authors of the resources license their work with Creative Common licenses, the resources become Open Educational Resources and the teacher can reuse, transform, remix and redistribute them freely. The only obligation, usually, is to quote the original author correctly. It is therefore important to identify when a resource is open or not.

There are some collections of well licensed open educational resources and your ministry probably has one. But what about finding these resources anywhere on the web? Can we make use of a specific search engine for this?
Project X5-GON was funded by the European Union to find and index OER, to use artificial intelligence to curate these OER and propose AI tools, typically search tools, allowing users to better find OER.

Where does the AI appear in such a process? It will appear in all stages:

During the ingestion stage, robots will scrap the web and return OER: this is a complex process as it means recognizing the OER and therefore the licenses. Part of it can be seen as a typical classification task (a common AI task).
When the robot has found a resource, the text from this resource has to be recovered. When the resource is an audio or a video file, this means using transcription (more about this in Chapter 7).

The 5 in X5-GON refers to the 5 barriers or dimensions the project wished to address: one of these being language. So the next step of the process is to use automatic translation tools to obtain text versions in each of the chosen languages. Again a typical AI tool.

At this point you may wonder: Automatic transcription and translation are fast growing technologies. But they still make serious mistakes. Isn’t it dangerous to rely on these? The answer is that search and recommendation algorithms don’t need the exact text. They need to position the document in a space; next to keywords, to other documents.

Think about when you have a box full of old papers that you need to organize. Ideally you would want to have a preset organization and file each paper in the right place. But we usually don’t have this pre-existing filing system and end up putting the papers close to each other when they have things in common, with unwritten rules of all sorts: these two papers go together because they are from the same year, these two because they are related to cars, these two because they have the same size, and so forth…The key term here is “next to”: we will discuss this later in the book.

Once raw texts have been extracted, models can be built.Documents will become vectors in some high dimensional space and comparing vectors will allow us to answer the questions: which are the 10 documents most similar to this one? Which are the 5 documents which best match this keyword?

Richer models can be obtained through training. They can answer more complex questions:

How difficult is this course? The answer can perhaps be somewhere in the course description, or in the meta-data, the data hidden to the viewer but which are supposed to give information about a document. More likely, they can be obtained through automatically analysing the document: the lengths of the sentences and the words, the words themselves are strong indicators of the age for which a course was intended
Should I look at this courseware before this other one? This is the prequel to be able to have a full course built automatically from given courseware.
What is the quality of the course? This is for AI the difficult question. It is also possibly a wrong question where AI can do more harm than good. Nevertheless, being able to find out if the facts in a course are correct makes a lot of sense. After the fake news, will we have fake courses?

Some links
X5-Discover (https://discovery.x5gon.org/) is a search engine
The learning platform X5-Learn (http://x5learn.org/) allows to choose one's courseware and get the AI to organize it in the best order. In this case, a recommendation engine is used.
More X5-GON tools (an API for developers, a version to be installed in Moodle) can be found here.

The X5GON project was funded by the European Union’s Horizon 2020 research and innovation programme grant No 761758

This page is referenced by:

Open or closed?
Open Educational Resources (OER) and their history
Educational resources refer to any material, nowadays mostly digital, which will play a part in education: textbooks, slides, curricula, exams,... They will be open when they can be freely shared with others (but a more exact definition will be given in a moment).

Even if education has been open in many aspects in several moments of history, the actual terms were better understoodT. he following definitions of OER and Open License were revised in connection with the Recommendation on November 25, 2019³:
1. Open educational resources (OER) are learning, teaching, and research materials in any format and medium that reside in the public domain or are under copyright that has been released under an open license and that permit no-cost access, reuse, repurpose, adaptation, and redistribution by others.
2. An open license is a license that respects the intellectual property rights of the copyright owner and provides permissions granting the public the right to access, reuse, repurpose, adapt, and redistribute educational materials.

The terms open content and OER refer to any copyrightable work (traditionally excluding software, which is described by other terms such as open source) that is licensed to grant the following rights (also known as the 5 Rs)^1,2:
• to Retain - the right to make, possess, and control copies of the content (e.g., download, reproduce, store, and manage).
• to Reuse - the right to use the content in a variety of ways (e.g., in class, in a study group, on a website, in a video).
• to Revise - the right to adapt, adjust, modify, or alter the content itself (e.g., translate the content into another language).
• to Remix - the right to combine the original or revised content with other material to create something new (e.g., to embed the content).
• to Redistribute - the right to distribute copies of the original content, revisions, or their combination to others.

It should be noted that these rights are non trivial: for example, the third right is essential for teachers: to be allowed to take someone's learning material and adapt to one's own purpose, to the duration and level of one's classroom, perhaps to geographic and cultural specificities.
Why AI wants open data
On the other hand, as demonstrated in different parts of this book, and also by the financial investments of the industry, Education can be seen as a market. And as machine learning if the principal force driving Artificial Intelligence, it is fair to deduce that for AI to thrive, AI for Education will need data.
The difference between user data and knowledge data
The sort of data AI for education will be needing is of two types.
Data about the users. How do they learn? What triggers good learning? What allows to better learn? As Daphne Koller once put it: ‘Let's make education science into a data science!’

This data can only be produced by the users themselves. It is therefore essential for companies to own platforms with which Users will be asked to interact. This has been the key to success of many AI companies and will be the key for success in education.

The second type of data concerns knowledge. In education, courseware represents a large chunk of this knowledge. This data is or isn't shared: in most cases knowledge creators or collectors may know little about licenses and the material they have produced will be hidden in University repositories, on strange blogs, or shared inside specific groups on social networks. Some of this knowledge is of course behind paywalls and some is on sites whose business model involves offering the knowledge for free, but in a setting in which one has to view adverts and unwanted publicity to get or maintain access.
User data has to be protected
In the first case the data -the user data- has to be protected. More so if this data belongs to under age pupils. Which means that the school or the teacher should not share this data with platforms unless they are explicitly allowed to do so. Even when the platform does offer some interesting service. In a similar way, it is never a good idea to register the names and addresses of one's pupils in order to participate in some activity. Please refer to the video about data and reidentification for details.

The European Union has provided a robust framework to protect its citizens, their privacy, their digital rights. This is called the GDPR. The GDPR protects by giving the citizens rights that have to be granted by the platforms, whether they are for education or not.
Knowledge data should be shared
On the other hand, knowledge can be shared. And should be shared. Obviously this is only possible when one has the right to do this, which means understanding how licensing works. Creative commons licenses are usually those that work best for OER.
Once OER are shared, artificial intelligence can be used by many things, such as those present in project X5-GON.

------------------------------------------------------------------------------------------------------
¹Wiley, D., & Hilton, J. (2018). Defining OER-enabled pedagogy. International Review of Research in Open and Distance Learning, 19(4).
²Wiley, D (2014).The Access Compromise and the 5th R.
³UNESCO. (2019). Recommendation on open educational resources (OER).
Search Engines Part 2
Authenticity and Relevance

We have been taught to have confidence in books. Publishers, librarians, professors and subject-matter specialists act as gatekeepers to printed resources. They make sure the resource is authentic and of good quality. How to find, and make sure the students find, appropriate sources when it comes to the web?^1,2

Even printed material from good sources have been proven to be full of errors and biases. Perhaps nurturing the culture of critical evaluation would reap benefits beyond choosing good online sources.²
Digital Repositories
One good way to ensure content stays authentic is to use digital collections from known and trusted sources. These can vary from school level resources to global ones. Over the past two decades, the number of digital libraries have increased dramatically, allowing educators to use data sets, maps and images,and other documents.³

To find information, you can use either search inside the corresponding web site or using site: keyword in a search engine.
Even Google has specific search engines for
- public datasets , where governments, public institutions, and large corporations make their collected data available for the public. It is all public information collected from studies, surveys, and censuses.⁴
- research articles, where millions of papers and books, both open access and fee based, are indexed and ready for search.

Apart from these, opencourseware from universities, Khan Academy and online encyclopedias are all popular sources of information.

Volatile Content
Search engines are constantly testing new algorithms. Mobile search differs from desktop search since it tends to favour results that are tied to current location.⁴ Everyday, new content is indexed, old content is indexed in a different way. New data is created from old data by re-analyzing content.¹ Copyrights and licences change. Laws concerning data change both in time and from place to place. For example, search engines continue to list content to which index has been removed inside the European union because of GDPR. Even maps change depending on where they are accessed from. Language and its usage changes. Both the use of medical categories and their interpretation vary from country to country.¹

Not to forget that search results are ranked according to the history of user activity, their personal information and privacy settings. Thus all of us have access to different content and might not even be able to find the same content twice. All these differences have to be taken into account while setting and grading educational activities.
Other Challenges
Search engines bring in other changes too. Knowledge is easily available. We don't have to know facts anymore. Instead, we try to remember where and how to find it.¹ Programmers cut and paste snippets of code. Engineers run simulators. There are forums for both homework questions and teacher complaints. Skill is becoming more important than knowledge and memory. Even our views on ethics and morality change - How to explain plagiarism to the Generation Copy-paste?

------------------------------------------------------------------------------------------------------
¹Russell, D., What Do You Need to Know to Use a Search Engine? Why We Still Need to Teach Research Skills, AI Magazine, 36(4), 2015
²Marion Walton, M., Archer, A., The Web and information literacy: scaffolding the use of web sources in a project-based curriculum, British Journal of Educational Technology, Vol 35 No 2, 2004
³Land, S., Hannafin, M. J., & Oliver, K. Student-Centered Learning Environments: Foundations, Assumptions and Design. In Jonassen, D. H. & Land, S. (Ed.), Theoretical foundations of learning environments (pp. 3–26), Routledge, 2012
⁴Spencer, Stephan. Google Power Search: The Essential Guide to Finding Anything Online With Google, Koshkonong. Kindle Edition.

This page is referenced by:

This page references: