Theme 3: Research and theory

The journal

I’ve chosen to take a look at the Robotics and Autonomous Systems journal with an impact factor of 1.158. I chose the article “A survey of robot learning from demonstration” published in 2009.

The aim of the article is to examine how robots learn the best and how to implement it on a high level. This is one of the main factors that separate humans from robots that humans can interpret situations on a bigger scale than robots and from that withdraw knowledge from that information. That is called experience based learning, and that is not what this article is about This article focuses on how robots learn from demonstrations (kind of monkey see, monkey do) and they map and structure the different choices given during an implementation of designing this type of robots. It is not only focused on gathering the different input information types it is also discussing the different implementation methods. It also discusses its limitations and future research areas connected to the article.

The first part is about gathering the examples and the second part is about deriving them.

My first thought was that the authors is not inventing something new, just testing different methods from other studies and to map them all out, creating a structure of different theories in LfD design and forming a base for future researchers to stand on improving and inventing new methods.

Under the facts about why this (Learning from demonstration) is a important and relevant study, the authors do not refer to actual facts and provides the reader little of trustworthiness in their statements. It does neither clearly explain the positive expects using these methods, only referring to the fact that it opens the robot creating business up for more people than just the robot experts, which I’m not sure that I can agree upon. The positive comments I came to think about reading this article is the analysis that the authors share with the reader. It is a deep and furrow analyze with many factors included.

The Questions

1. Briefly explain to a first year student what theory is, and what theory is not.

Theory could be many things. As stated in the article “THE NATURE OF THEORY IN INFORMATION SYSTEMS”, even the word’s meaning in a dictionary is not precise and lacks a persistent explanation. Thus to answer what theory is, isn’t possible with just one word or one sentence. More over the term “theory” is changing depending on what area of studies it belongs to are within and who are referring to the word.

Within theory there is five parts, dependent on each other; (1) theory for analyzing, (2) theory for explaining, (3) theory for predicting, (4) theory for explaining and predicting, and (5) theory for design and action.

Theory is not references, data, variables, diagrams and hypotheses as many believe. Using these five incorrectly as theory might get the readers to think that the text does not contain enough theory.

2. Describe the major theory or theories that are used in your selected paper.

Which theory type (see Table 2 in Gregor) can the theory or theories be characterized as?

In the paper I have selected they use the “Design and action” theory. This I decide based on the fact that the paper not predict anything, only discuss and compare the information withdrawn from other studies made by others. It is an implementation discussion on how to design robots, thus it is the aim of the paper.

3. Which are the benefits and limitations of using the selected theory or theories?

One fact that is negative in choosing this type of theory is that it is hard to easy define if it has “contributed with knowledge” (as stated in Gregor) that is one main purpose of the paper and without adding new information to the world it is no longer a research paper. Another factor that is important to seize choosing this theory type is to make the area of investigation it narrow, facilitating the overview of the paper for the reader. The fact that the paper is easily read, understandable and have a good lay out and structure is an important thing for papers in the “Design and action” category. Since it is then easier to clearly see that it is contributing with knowledge and summing different resources up, enabling future researchers to base their research on the article.