On the difference with a regular in-presence university course, this MOOC is based on voluntary participation and does not lead to a diploma. So, we will do our best to make it enjoyable in addition to being instructive.
We are well aware that most of you have a busy life and little extra time. For more efficiency, we will state the key concepts of each lesson straightforwardly (as in the outline below), and apply them concretely. We will help you get these concepts right quickly through activities that we designed to be as straightforward and efficient as we could imagine. We will, nonetheless, suggest extra readings for those who wish to dig deeper into the theory.
For those of you who are ready to invest more time in programming, we will make all the Java demo files and all the necessary explanations available to set you on track.
We too are doing this work aside from our usual professional activities (and job search activities). We do it with as much professionalism as we can, but please be forgiving for possible flaw. Just let us know when we can improve.
Debates and discussions will constitute an important dimension of the MOOC. We created the Developmental AI MOOC Google+ community to host them; it is important that you join it so you can follow and contribute to these discussions. If you do not wish to use Google+, you have the other option of using the forum of our Claroline Connect platform but, sadly, it might be less convenient.
|#||Date||Title||Key concepts||Reading||Extra reading|
|1||10/13||Embodied paradigm||Do not consider the agent's input data as the agent's perception of its environment.||Embodied Cognition (Wikipedia).||
Inverting the interaction cycle (Georgeon & Cordier 2014).
The Embodied Mind (Varela et al., 1991).
|2||10/20||Sensorimotor paradigm||Focus on sensorimotor interactions rather than separating perception from action.||
Cognitive development (Piaget / Wikipedia).
Interactional motivation (Georgeon et al. 2012).
Sensorimotor contingencies (O'Regan 2001)
|3||10/27||Constructivist epistemology||Cognitive agents must discover, learn, and exploit regularities of interaction.||Constructivist cognition (Riegler 2007).||Construction of Reality in the Robot (Ziemke 2001).|
|4||11/3||Self-programming||Self-programming arises from active contextual exploitation of regularities of interaction.||Enactive artificial intelligence (Froese & Ziemke 2009).||
(Thórisson et al. 2013).
Radical interactionism (FAQ) (Georgeon).
|5||11/17||Radical interactionism||Use sensorimotor interactions as primitives of the model.
Self-programming underpins autonomous sense-making.
|Desiderata for Cognitive Architectures (Sun 2004).||Enactive e-puck robot (Georgeon et al. 2013). Sense-making (Georgeon & Marshall 2013).|
|6||11/24||Cognitive architecture||Spatio-temporal regularities of interaction lead to ontological knowledge of the world.||Enactive Cognitive Architecture (Georgeon et al. 2013)||Memory and navigation (Buzsáki 2013).|
|7||12/1||Research agenda||Future roadmap.|
Lessons and activities will be published on Monday mornings, European time, from week 1 to 6 at this address. The activities must be completed by the next Monday. The 7th week will be used for conclusion and discussions about Developmental AI research perspectives. Most activities will use questionnaires with automatic assessment but some assignments will involve peer assessment. The certificate of participation will be delivered via email to participants who completed at least 5 of the 6 activities.
The readings may be done before or after the activities. If you start with the reading, you will probably better understand the significance of the activities. If you start with the activities, you will better understand the concrete implications of the reading. So it is your call.
The language of the course is English, but you can use your prefered language in discussions and assignments. We will try to arange peer assessment amongst participants who speak the same language :-).
We strongly advise the participants who wish to do the optional programming activities to do them regularly every week, otherwise it might be hard to catch up. Weeks 1 and 2 will be easy but crucial to get the concepts right. The next programming activities will require more work. On demand, we can include a personalized mention in the certificate of participation for participants who share their programming work online at the end of the course.
Back to the IDEAL MOOC home page.
Last updated October 28th 2014.
See public discussions about this page or start a new discussion by clicking on the Google+ Share button: