18. november 2018

Robot Assisted Teaching

I did a short presentation and had a following, interesting discussion at SLATE the other day.  I began with the robot introducing itself through a video we use to recruit students to test and write about the use of the robot.

Through the project Robot Assisted Teaching (Robotassistert undervisning) we try to develop and test various robot assisted scenarios related to physical robots. The scenarios will be developed with teacher students and in cooperation with schools.

The project want to contribute to a better understanding of how robotics can be used to support coding in schools and prepare future teachers for didactical use of this type of technology. The main goal is that students should be able to work with methods and tools that they can later use in their own, future teaching practices.

The core idea of the project can be narrowed down to two premises: 1. coding will become part of education in primary schools, and 2. to an increasing degree we will communicate with machines through speech-interfaces. Our conclusion: pupils should learn to program speech interfaces, to better understand how these technologies work and influence on their ability to take part in meaningful discussion about technology.


Robots that can be programmed to talk, by pupils, seems to be an appropriate way of teaching some skills for the future. We therefore work with school robots from the Spanish company Aisoy. The reason is that these robots are cheap (about the same price range as Lego Mindstorms), but also programmable in Scratch and able to perform text to speech (TTS) synthesis without being online. The ability to use the robots without sending any information out of the classroom is something I consider extremely important, and we are really satisfied with Aisoy who implemented this feature.

We follow what one have already experienced in Denmark: robots are not necessarily useful when it comes to learn ordinary pupils about a topic, but they can be really useful in the intersection between various topics and coding. Hardware that schools can afford is important in this respect, given that we want pupils to to interact with robots, both through software and physically.

In the discussion about the revision and renewal and the learning plans (fagfornyelsen) coding is instroduced, but for reasons I can't figure out it seems like this will be mainly in maths and science. Which might work fine, but from my perspective coding has much more to offer in topics like language learning, arts and crafts and music. This because I see coding as communication, and a way to think logically about communication. To implement logic in communication this you will need math and a structured way of thinking. To me the focus on coding in maths seems to approach this the wrong way around, instead of coming from other topics to math, through coding.

There is also a lot of talk about "algoritmisk tenking", which might be a rather poor translation of computational thinking (in the tradition following Seymour Papert's Mindstorms). Hopefully not because someone has looked it up on Wikipedia. Anyways this way of algorithmic and/or computational thinking involves decomposition, pattern recognition / data representation, generalization/abstraction, and finally algorithms (a video from Google explains this view on CT).  In short the ability to break down complex problems, into smaller and more feasible tasks that one is able to solve.

With Seymour Papert's words:


I am not by any means against algorithmic thinking, but I am a little bit worried if this becomes the dominant approach to coding. It is a difference between solving a problem and creating a problem. The two are obviously connected. I come from a slightly other tradition where I like to approach a problem through a narrative logic. In this way I can tinker with problems by expression them in words, drawing sketches, making physical models etc. In my world thing are created through trial and failure, where one often construct problems that build up towards solutions. This is not in opposition to computational thinking, and pretty much along the lins that Papert describes in the video above.

At this point I will quote Andy J. Ko, who write about Mindstorms: what did Papert argue and what does it mean for learning and education?: "Papert argued that instead of teaching our most “correct” forms of knowledge to children—Newton’s F=ma—we should be crafting representations of those ideas that are more easily linked to children’s prior knowledge and more easily appropriated by children to engage their interests."

The essence is that we learn and construct knowledge in different ways, which is why we want to approach coding as communication. That means that pupils should have access to various, affordable programmable artefacts, from micro:bits to affordable robots. We believe that we can achieve a lot with a robot like the Aisoy, not being able to dance the macarena (which is only really useful when you program the dance yourself). However, the Aisoy will make pupils and teachers able to get some substantial insight into some of the logic behind speech interfaces (chatbots, etc).

Not being similar products, but in schools you can do pretty much the same with Lego (best for programmed movement) and Aisoy (best for programmed communication). For the cost of 1.000 euros you can have 2-3 robots. Compare this to Pepper, at the cost of more that 20.000 Euros, or Nao, about 8.000 Euros.
We were discussing the differences between the cheaper robots, the Lego and the AIsoy. The latter can be programmed to speak, which interests me. Verbalisation is important, also in math and science. The Lego, taking its name for Paperts book Mindstorms, is more about programming movement.



The following discussion was really interesting, with people explaining their own experiences and visions for how robots can be used. At this point it is quite easy to use robots to impress people, that is to make robots appear as they are intelligent. Most are, however, more like mechanical turks.

The story of Sophia (or scam) comes up. Others believe that AI will never really happen. However, what we can safely see coming is robots performing more or less limited tasks, better than humans. Because computers are fast these tasks can be combined and easily come close to something that resembles intelligence.

Robots are good for repetitions. This can be really useful in a lot of special cases. Aisoy have had really good results with the use of their robots  with kids with autism. We also discussed the potential for the use of robots to assist people with various cognitive disabilities. I also hope to pull off a project where we look into how a robot can be a reading companion, in order to give children better reading experiences.

And, yes, the presentation:

Ingen kommentarer:

Legg inn en kommentar

Kommentarer er svært velkommen. På grunn av problemer med spam må jeg imidlertid godkjenne kommentarer før de publiseres. Vanligvis skjer dette i løpet av noen timer.