An important concept in the field of Artificial Intelligence is that of Search Spaces or State Space Search.
This concept can be helpful in gaining a deep insight into this field of research by being able to interpret the various AI tools as different ways of how to go about exploring and searching a state space. And any real-world problem can be mapped as a state space search problem.
While teaching an undergrad class in AI, I observed that students can experience difficulties in grasping this concept. They find it difficult to make the jump from an every-day real-world problem to being able to represent it on a higher abstract level, the search space.
But once (and if) students are able to gain a deep understanding of this concept, then it opens new doors for their understanding of AI. And then things start to fall into place - diverse tools and topics suddenly appear to make sense and their inter-connectedness becomes obvious. I observed many a student who passed through this Eureka! moment. Sadly, others didn’t make it. In the field of pedagogy, this is called a Threshold Concept.
To help students better understand the concept of search spaces, I started working on a number of tools, visualisations and videos. This blog entry is about one such experimental video. This work also forms part of my ongoing studies for a Post-Graduate certificate in Higher Education at Middlesex University.
As an initial experiment, I opted to choose a scenario that is grounded in the real world (representing a person’s daily activity) and which in my opinion is a non-trivial example.
Traditionally AI under-grad courses tend to tackle state space search using toy-like simplified examples. Examples include the 8-Queens Puzzle and various flavours from the class of River Crossing puzzles. While these are valid examples, in my opinion, they fail to show how to cross the gap from the problem grounded in the physical world to an abstract representation of the problem.
In this introductory video I am trying to convey the following messages:
The video is essentially a recording of a powerpoint presentation, with voice narration done via the FromTextToSpeech website. Editing of the video was performed in Adobe Premiere and then uploaded to YouTube.
I plan to extend this work further. Apart from video material, interactive visualisations can help a student to ‘experiment’ with this concept. I might use R for the visualisation part.
Look out for more on this in future blogs!