Smart Objects: Challenges for Semantic Web Research

Paper Title: 
Smart Objects: Challenges for Semantic Web Research
Marta Sabou
The increased availability and robustness of sensors, the wide-spread use of the internet as a communication environment and the intensified adoption of semantic technologies foster the vision of embedding intelligence in physical objects. The race of realizing this vision is pervasive to a variety of research fields, most notably ambient intelligence and semantic web, and lead to the proliferation of several overlapping definitions and terminologies: smart products, semantic devices, semantic gadgets - to which we collectively refer to as smart objects. What exactly are smart objects? And what are the research challenges in realizing them? We hereby explore the answers to these questions.
Full PDF Version: 
Submission type: 
Responsible editor: 
Krzysztof Janowicz

Review 1 by Boyan Brodaric:
This paper characterizes Smart Objects and describes associated research challenges. Smart Objects are physical things such as products, devices and gadgets that are characteristically aware of their contexts, can adapt their behaviour to these contexts and form organized structures within them. Implicit in the realization of these characteristics is a foundational reliance on Semantic technologies, which are currently challenged by the limited hardware resources, complex reasoning, poor data, and broad and emergent knowledge requirements of Smart Objects.

The paper is well-written and, given the proliferation of sensors and mobile devices, is clearly an important topic. Apart from some minor presentation issues listed below, I would like to see the paper better differentiate Smart Objects from existing AI endeavours to build intelligent machines, such as robotics. How does the incorporation of Semantic Technologies make this different and beneficial? The general characterization of semantic devices, for example, would seem to also broadly apply to some of these established efforts. While some of the differences might be implicit in the research challenges, it would be worthwhile to explicitly draw them out in the description of Smart Objects.

In terms of presentation:
-"and lead to the proliferation" should be "and leads to the proliferation"
-"This adaptivity is enabled by three main technologies"—the 3 technologies should be more clearly identified
-"the resources characteristic to the computer" should be "characteristic of the computer"
-"From our analysis of smart objects and their characteristics, we can conclude that their representation needs are much richer and more diverse"----can you say a little more about why this is the case?
-First sentence of section 4. Summary is too complex

Review 2 by Lora Aroyo:
(abstract) lead => leads
Add references when referring to terms or projects. When explaining the central nature, clarify where the centralization applies to: the management of the devices or the definition of the semantics. The last part of section 1's second paragraph needs to be reconsidered: be more to the point. It is important to explain what the steps forward, and therefore the innovations and challenges are. Why does a look at several fields give good challenges? This is only relevant when these fields necessarily complement each other.

Add references when mentioning concrete products or projects. When coming to a joint definition in the diverse fields (towards the end of section 2), it is essential to explain why an integration makes sense. For example, there are many fields that use the word "context" and while there seems to be a common idea, the different notions of context and what to do with them differ significantly. After all, the objective is not to find a concept that is common in different fields, but to have a concept that helps solve a concrete and relevant research problem. Refer to publications when mentioning them.
There appears to be a discrepancy between "smart objects" and "smart products". In section 2 the story appears to concentrate on smart products, but the paper's title refers to smart objects. Rephrase the last two paragraphs of section 2 to arrive at a concrete lesson learned from the state of the art.

At the beginning of section 3 you are caught between two motivations: one is to create the smartness that products need and the other is to embed semantic web technologies also in objects with limited resources. The one is the ambition to create the smartness that autonomous products want. The other is the ambition to extend the semantic web to devices as well. These are separate ambitions.
The above consideration makes some of the goals discussed in section 3 not trivial to understand: they depend on the ambition one is having. The paragraphs on the reasoning algorithms do not give a clear picture of the general challenges.
The paragraph on the suboptimal quality is too informal and without evidence.
The paragraphs on the representation of a variety of information are rather abstract and need to be reconsidered.
The paragraph on further challenges is without evidence.

The identification of challenges is performed rather bottom up, i.e. from the problems of current partial solutions. This makes it hard to see which are the grand challenges that help to solve a concrete problem. Without doubt, the partial solutions can evolve in their respective fields, but for a coherent approach to smart objects more cohesion is needed in (the identification of) the challenges. It is best to keep the summary short in a text that is already short. There are some errors in the last sentence.