State of the art - Why using a humanoid like Poppy?


I’m part of the cherry project

In this project my role is to back up the use of Poppy instead of using a tablet. My first idea was to look into empathy. It would be logical that human, and in our case children, tend to personify and feel emotions for a robot more easily than for a tablet.

What do you think ? Do you have any researches that would support or undermine this idea ?

This aspect is not my research speciality, I think @Pierre or @oudeyer will have more good references for you but meanwhile you can look at the work of Cynthia Breazeal and Ben Robins

Also here are some interesting projects exploring the interaction between robots and children:

First why oppose having a robot and using a tablet? I think both approaches are actually quite complementary (no I’m not advertising my own phd here ^^) and are both really interesting in the context of the Cherry Project.

Second, I would say that a very important point when using robots and especially humanoid ones is their physical presence. They are embedded. You can look at them, touch them, see what their doing…

This is as you say very important for empathy. As we are social creature we really like to interact with others but also with creatures. It has been shown that interacting with pets could really improve mental health of elders for instance. Yet, their usually not allowed in hospital and robots could be a really interesting substitution in this case. You can look at the very interesting (and successful!) Paro robot. Even really simple interaction - we are far here from human-human type of interaction - can already be really stimulating.

For your specific question, I think it really depend on what type of information are convey by the robot/tablet. I guess you have to be more specify here. For instance, you can find studies trying to compare gaze of humanoid face versus using a screen here. You should find other interesting paper in their related work. I’ve seen (I can’t remember where) studies comparing smileys versus a robotic face mimic emotions.

Two other related very related european projects:

(I’ll complete this post when other references come back to my mind).

1 Like

You should probably give a look at the EMOTE research project website ( ). The goal of this project is to study the use of robotic tutors. They focus on empathy as a way to improve the learning experience.

On the advantage of robotic vs virtual agent, there is also this paper of C. Bennett that tries to derive a minimal set of features to express emotion. They compared a virtual and an embodied robotic head. Their results show a small (but statistically significant) improvement of the likeability and animacy of the robitic head over the virtual head. ( )

1 Like

Hi @pilllea The Alize project is indeed a very good reference. It has been evaluating in the field and on a large scale the use of humanoid robot as social tools to help diabetes children understand and manage better their illnesses. Field experiments were made in hospitals and field camps in several European countries, and comparisons were made between real robots/simulated robots/tablets. The use of real robots was shown to provide better engagement and motivation than simulated robots or tablets. Now, tablets were also used in this project as an interface between the children and the robot (and between the doctors and families and the robot), and this was absolutely key to provide a successful interaction (it is natural, efficient robust). Using only speech synthesis and recognition and gestures is not mature enough to provide robust and thus acceptable interactions. So I strongly suggest you consider using a tablet as an interface.

1 Like


I am part of the same project as @pilllea.
As part of the project, we need to make videos explaining what we are doing.

I would like to interview a resercher on the topic of empathy and the psychological benefits of robots. I only need a minute or two of footage, so it wouldn’t take long.
Would @Pierre or @oudeyer be interested?

I suggest you take a look at very good existing talk on the topic, e.g. by Tony Belpaeme:

and also;

Thank you, that’s exactly the sort of content I need (though on a shorter scale^^) but I don’t think I can just rip off someone else’s video, which is why I wanted to make my own by intervewing local reserchers.

Hi @sophie,

I’ll be at cap sciences this friday and saturday. It could be an opportunity to make the interview. What do you say?

That sounds perfect, I’ll ask my manager if it’s possible to get the camera and everything for friday.

Hi @Pierre, would friday afternoon be ok?

Another very interesting work: