AI can already generate short textual summaries from lengthy texts. But that’s just nonsense 🙁 Only if the long text is not 99% water spilled around a few important thoughts 😉 On the other hand, creating summaries in the form of visual schemes is a whole different story! Here’s an example of a platform that you can use as a starting point to create an even more efficient tool for this.
The essence of the project
Algor Education has created a useful learning tool that converts long texts (lessons, books, articles) into visual maps that present the main ideas of the text in a structured way.
The platform’s AI engine can create two types of visual maps:
- Hierarchical: These maps reveal the essence of the topic discussed in the text from top to bottom. For example, if the text is about volcanoes, one chain of objects on the map would reveal the history of volcanic origins, another would reveal the structure of volcanoes, and a third would list the characteristics used to describe volcanoes.
- Mind map: These maps scatter the main concepts and ideas from the text in a seemingly disordered way, but with associative connections linking them.
In addition to the visual maps, Algor Education’s AI engine also creates a text summary of the text.
The automatically created visual maps and text summaries can be edited directly on the platform if the user wants to present the information in a more understandable way.
If the source text is very long, such as a book or textbook, the AI engine will automatically create several visual maps and text summaries, each of which will describe a separate logical block of text, such as one lesson from a textbook or one chapter from a book.
The platform can be used to upload not only text. Audio recordings can also be uploaded, such as self-recorded lectures or audio tracks from found online videos, which the platform’s AI engine will automatically convert to text. It is also possible to take pictures of pages of a book or textbook on your phone and upload them to the platform.
In this case, the AI engine will extract the text from the photos and then create visual maps and summaries.
The free version of the platform has serious functional limitations, but it allows you to get acquainted with the essence of the service. The full-featured versions will cost 5.49 or 7.99 euros per month, depending on the number of “credits” included in the tariff. However, you will have to pay for a full year. Credits are spent on text analysis, and their number depends on the length of the text and the need for its additional transformations (“audio to text” or “image to text”). If the credits included in the tariff are not enough, you can buy them separately.
Algor Education is based in Italy. In 2022, it released the beta version of the platform, and in early 2023, the full version. Currently, it has about 10,000 subscribers, 20% of whom are from abroad. The startup has raised its first serious round of investment of 1.4 million euros, which added to the 180,000 euros it raised at the pre-seed stage.
What’s interesting
Visualizing information is a powerful learning tool that helps you understand the essence of the topic you are studying, remember it better, and recall it more quickly.
One of the pioneers of this method is considered to be the Soviet educator V. F. Shatalov, who used “key schemes” or “key notes” in school education, which visually represented the “key signals” – the main thoughts and concepts of the studied topic. His method turned out to be so successful that his students managed to finish school a few years earlier and almost 100% enter institutes after that.
On an ideological level, today’s Algor Education is a follower of the same method, but its visual maps are more traditional. In my opinion, the full use of the “Shatalov system” can make them even more effective. Although it is possible that the founders of this startup have never heard of Shatalov
And here is another very cool, and at the same time modern method of visualization that can be applied in education – using “memes”. The startup Antimatter, which I wrote about in June, decided to create a platform for this method of learning. The main idea of the startup is: if you cannot express the essence of a concept in the form of a meme, then you do not understand it yourself Therefore, students create their own memes on the platform and offer them to other students in the form of “key notes” or riddles – and this helps them to better master and memorize the studied material. This startup raised $2 million in investment.
If we digress from visualization, the development of AI technologies has made it possible to create different types of tools to support the educational process.
The startup Gizmo, which I wrote about in September, developed an AI machine that automatically creates cards with concepts, facts, formulas, and other things that you need to periodically sort through, reminding yourself of what you want to remember. Such cards are usually used in language learning – when one side of the card says a foreign word and the other side says its translation in your native language. The AI machine automatically creates concept cards from the texts of lectures and articles that the student uploads to the platform. This startup raised $3.5 million in investment.
The startup Atlas Primer, which I wrote about in January, is aimed at people who find it easier to perceive information by ear. Therefore, the platform converts text lessons into audio recordings that can be listened to on the go, instead of sitting and reading them. In addition, the AI machine of the platform automatically creates brief summaries of lessons that can be listened to to remember what they were about. It even knows how to create sets of tests for uploaded lessons that can also be passed in audio mode. This startup raised $850,000 in investment, but since then, quite a few similar applications have appeared.
If we return to the topic of visualizing information for better understanding, then such a thing is needed not only in education.
The startup StructureFlow, which I wrote about at the end of 2022, started by creating an AI machine that can convert corporate legal documents into visual diagrams that more clearly show the structure and relationships of the agreements set forth in them.
Since then, the startup has expanded the scope of application of its platform, which now knows how to turn different types of business documents into visual diagrams. This startup raised $8 million in investment.
Where to Run
Using AI for creating concise summaries from lengthy texts no longer surprises anyone. Such functionality is widespread.
However, transforming a long text into a short one is an obvious loss of a lot of information, unless the lengthy text is just a lot of water poured by the author around a few main ideas 😉
On the other hand, transforming long texts into another format that remains short but retains much more useful information and helps understand it better is a much more promising thing:
Where will more useful information be retained—on the text summary page or on the mind map or summary page? What is quicker to grasp at a glance to remember the essence—a text page or a visual scheme? By the way, tests to check the comprehension of the material can also be created in a visual form! Draw the missing connection on the scheme, insert the missing block with the concept, and so on.
Moreover, creating truly informative summaries is not just about a simple graphic representation of a text summary—what today’s Algor Education essentially does. The methodology for creating summaries simultaneously uses different methods of visualizing information, relies on psychological principles of memorization and recall—such summaries look different and are more informationally rich.
Additionally, now a summary can be not just a static image but a moving picture, a short cartoon that can be viewed on your phone. The static format is an old limitation when summaries could only be drawn on paper.
A similar approach can be applied to create voice summaries. The mentioned Atlas Primer essentially just narrates a traditionally created text summary:
Who said that such a voice summary must be a coherent text—why can’t it be a voice representation of a summary? And why should such a voice summary necessarily be a boring text—why can’t it be a song put to automatically generated music? Rhymed texts are remembered better, and songs, for some reason, are remembered on their own, unlike lectures 😉
So, the main direction of movement is the development of algorithms and platforms for creating more efficient summary formats than ordinary text.
Education and self-education are the most suitable markets for the application of such platforms. Considering that people now have to learn throughout their lives to continue mastering new things without which they cannot remain competitive, this market becomes truly massive, not limited to just schools and universities.
The timing for entering it is also very appropriate—because only at today’s level of technological development can AI be taught to create such unusual but effective summary formats.
About the Company
Algor Education
Website: algoreducation.com/en
Latest funding round: €1.4M, 18.12.2023
Total investments: €1.58M, rounds: 2