DBM130 - Designing Intelligence in Interaction - Q2 2016

Assignment resources

Schedule

Lecture and workshop materials and instructions

Lecture 1: Intelligence for optimization - Loe Feijs

Intelligence for optimization. There exists algorithms for solving a class of problems called "combinatorial optimization". Typical examples are finding the shortest path in a network (packet routing), finding a tour along cities (delivery service) or finding an optimal path along graph edges (3D printing).

Loe Feijs will illustrate the famous traveling salesman problem showing an interactive and intelligent G-code generator project for shoe soles (joint work with Troy Nachtigall). Another illustration is the optimal computerized embroidery path for the Pied de Pulse project (joint work with Marina Toeters).

References:

Feijs, L.M.G. & Toeters, M.J. (2016). Pied de pulse: packing embroidered circles and coil actuators in pied de poule (houndstooth). In C. Sequin, D. McKenna, E. Torrence, R. Sarhangi, B. Torrence & K. Fenyvesi (Eds.), Proceedings of Bridges 2016: Mathematics, Music, Art, Architecture, Education, Culture, Jyväskylä, August 9-13, 2016 (pp. 415-418), http://archive.bridgesmathart.org/2016/bridges2016-415.html

Feijs, L.M.G., Nachtigall, T., Tomico, O. Sole Maker: Towards Ultra-personalised Shoe Design Using Voronoi Diagrams and 3D Printing. Proceedings of SMI'2016 Fabrication and Sculpting Event (FASE), Reprinted in: Hyperseeing, Tthe Publication of the international Society of Arts, Mathematics and Architecture; SUMMER 2016, http://www.geometrysummit.org/proceedings/fase2016/papers/1.pdf

Lecture 2: Information processing for learning systems - Matthias Rauterberg

Lecture 3: Adaptivity in games - Erik van der Spek

Games and playful interactions have the unique ability to entice users to start interacting and keep them fully engaged with products for long periods of time. A wellknown theory in this regard is that of Flow, which describes that people should be optimally challenged according to their skill level. But what is cognitively interesting for one might be boring for another, necessitating intelligent adaptation to the player's cognition and abilities. This lecture will show a number of these adaptive approaches, and also, why they might be wrong.

Lecture 4: Design of social scenarios - Emilia Barakova

This lecture will introduce four classes of computational problems that most often need to be resolved in design applications. Especially Pattern recognition will be featured as potentially most frequently encountered problem.

Lecture 5: Supporting the design of intelligent systems with OOCSI - Mathias Funk

Designing intelligent products nowadays often means designing intelligent systems, or simply designing systems that act in smart ways for the purpose of improving the quality of life for stakeholders. System design is neither easy, nor well-understood as it happens in a multi-disciplinary design space that borders interaction design, complex adaptive systems, and engineering distributed systems. In this lecture/week we will look at the challenges, important values and tools for system design. We will use OOCSI and its system-level building blocks to design systems in a hands-on way; this will be done in a short group project.

Lecture 6: Control systems, dynamic systems, and Blocks World - Frank Delbressine

Topics to be dealt with are:

Regarding Dynamical Systems: All systems, especially mechanical systems, show oscillations. How do these oscillations look like. What are factors influencing these oscillations? How can these oscillations be minimized?

Regarding Feedback Control: One method to control the behavior of systems, not only mechanical, is feedback control. How does this work? Why does it work? What kind of controllers are there and what is their performance?

Regarding Robot Programming: If I have time left. I want to talk about Blocks World. A famous example from the book of Terry Winograd, "Understanding Natural Language”.

Lecture 7: Pattern recognition and neural network - Jun Hu

Lecture 8: Simulating behavior - Peter Peters

Simulating complex systems and behaviors using NetLogo

Deliverables

Rubrics

  1. Follow the instructions on the NetLogo website to download and install NetLogo

  2. Check out the description here and the videos