Introduction:What Landmarks Are, and Why They Are Important - Landmarks: GIScience for Intelligent Services - page 21

Fig. 1.7 Interpreting common language involves geospatial intelligence

computers are increasingly “capable of understanding, even if that understanding
is not isomorphic to our own”, and points to growing capabilities of computers in
“[sensing,] representing and contextualizing patterns, making links to other patterns
and analyzing these relationships” ( ibid. l) that will shape our interactions with them.
Thinking ahead along these lines, we are reaching the age of calm computing [ 48 ]
where 'the machine' is no longer a disembodied isolated apparatus (although IBM's
Watson still was disconnected from the Internet when it won its game of Jeopardy!
in 2011). Instead, with the Internet of Things, where every object is addressable,
has sensors, can interact with the environment and communicate with other objects,
'the machine' gets embedded in our environment. Objects in the environment can
sense our presence and our intentions, and hence determine, whether, for example,
they can serve as landmarks in the current circumstances. They can communicate
their readiness to our smart (shall we say: intelligent?) service that helps us
navigate in this environment, or maintaining our orientation. The smart service can
communicate back to the objects that they have been selected to support a particular
task, and could be triggered to stand out even more as landmarks (e.g., turning on
their lights) for a particular encounter or period of time.
In this context the original question of what is an intelligent machine poses itself
now different from an imitation game. Nevertheless, a whole range of challenges
has to be solved in order to build an intelligent system capable to communicate via
landmarks. Among these challenges are understanding references to landmarks in
verbal or graphical human place and route descriptions, understanding context of the
conversation and relevance within this context, understanding personal preferences
or knowledge of an environment, understanding the human embodied experience of
environments, especially salience, and understanding to use landmarks effectively
and efficiently in conveying information. Given the rest of the topic will talk about
approaches tackling them, let us go into depth here with the illustration of some of
these challenges.
Consider a request a traveller might pose to an intelligent system: “Can you tell
me the way to the airport?” (Fig. 1.7 ) . Sounds simple? Surely, people would not have
difficulties to answer this question in a manner that is easily understood and applied
by the traveller (let's ignore here that some people are really bad in direction giving).
They most likely would be able to resolve the meaning of the destination, which in
this case is also a landmark, from context, and in their response they most likely
would refer to some landmarks. They might say: “Sure. Tullamarine, you mean?
Follow this street [points] to the hospital, then turn right. From there, just follow the
signs”, referring to a hospital as a landmark at a critical point of the route, namely
the point where to turn right.