Are events in this experiment simultaneous if observed in platform's frame? In some contexts e.g. on Wikipedia it is defined as a matter of happening . In others(e.g. as defined by Einstein in his book "Relativity the special and general theory") it is defined as a matter of observation.
Both definitions are distinct from each other. In case of Wikipedia's thought experiment two events $A$ and B(reaching of light at the back and front of train) are declared not simultaneous without observing them.
acc. to the Einstein's definition events can be empirically dissimultaneous only if they are 
observed to be dissimultaneous. So A and B are not dissimultaneous as par the criteria imposed by Einstein. 

Are A and B (as explained on Wikipedia's experiment) are dissimultaneous in platform's frame, if yes then explain how these will be observed dissimultaneously.  

 A: First, to be clear, events are points in spacetime that exist independent of any coordinate system as is the interval associated with two events.
So, we can say, without introducing a coordinate system, that the interval associated with two events is timelike, lightlike, or spacelike
In the timelike case, we can say that one event is later than the other, i.e., there is a separation in time between the events with one event after the other.  It is possible to assign identical spatial coordinates to these events, i.e., they can be co-located in a coordinate system.
For the spacelike case, we can say that the two events are spatially separated and it is possible to assign identical time coordinates to these events, i.e., the events can by co-located in time (simultaneous) in a coordinate system.
Thus, one cannot say that two events are simultaneous in an absolute way.
Only by introducing a coordinate system (reference frame) do we add the necessary structure to discuss if two events are simultaneous according to some (but not all) coordinate systems.
Simultaneity is relative; it is in "the eye of the beholder".
A: Simultaneity depends on the frame of reference. The statement that two events are simultaneous in a reference frame is just the statement that in the time coordinate of the two events in that frame are the same. This definition in general depends on the reference frame.
However, there are two exceptions where whether or not the events are simultaneous is the same in all reference frame. The first exception is when the events occur at the same place and the same time in some reference frame. Then it occurs at the same place and time in all reference frames. So they events are simultaneous in all reference frames.
The other exception is when, in some reference frame, you can get from the first event's space-time coordinates to the second event's space-time coordinates moving slower than the speed of light. In this case the events will never be simultaneous in any reference frame. To see this, imagine a person moving at constant speed from one event to the other. Now imagine going to a frame where the events are simultaneous. The person would need to be in two places at once, but this is impossible.
The example on the wiki of the light bulb turning on in the train is an example where simultaneity does depend on reference frame. In the reference frame of the train the two events are simultaneous, but in other reference frames the events will not be simultaneous.
A: Simultaneity is a convention in special relativity, not an observation. The Einstein clock synchronization procedure is one way of defining a plane of simultaneity in space-time; however, special relativity can be (and has been by, e.g., John Winnie in 1970) reformulated with a wide range of allowable clock synchronization schemes with no effect on the things you can actually measure. Einstein himself emphasized in the 1905 paper that this was a matter of definition, not empirical fact.
A: 
[...] as defined by Einstein in his book "Relativity the special and general theory") it [simultaneity] is defined as a matter of observation.

Einstein's literal prescription can be read for instance here:
http://en.wikisource.org/wiki/Relativity:_The_Special_and_General_Theory/Part_I#Section_8_-_On_the_Idea_of_Time_in_Physics
To summarize in my own (hopefully carefully enough chosen) words:
Some particular indication (e.g. "being struck by lightning, $A_s$") of one particular element of a railway embankment ($A$) and some particular indication (e.g. "being hit by a thunderbolt, $B_h$") of some other particular element of a railway embankment ($B$, which was and remained at rest wrt. $A$) are called 


*

*"having been simultaneous to each other" if there exist an particiant ($M$) who was and remained "middle between" $A$ and $B$ and who observed $A$s indication $A_s$ and $B$s indication $B_h$ in coincidence;

*"having been dissimultaneous to each other" if there exist an particiant ($M$) who was and remained "middle between" $A$ and $B$ and who observed $A$s indication $A_s$ and $B$s indication $B_h$ not in coincidence, but in sequence ("first one, and later the other"; regardless of which "first");
or else, if $A$ had seen $B$s indication $B_h$ before, or at latest in coincidence with, stating its own indication $A_s$;
or else, if $B$ had seen $A$s indication $A_s$ before, or at latest in coincidence with, stating its own indication $B_h$; 

*or otherwise there is at best only circumstantial evidence about simultaneity or dissimultaneity of $A$s indication $A_s$ and  $B$s indication $B_h$.

In case of wiki's thought experiment [ http://en.wikipedia.org/wiki/Relativity_of_simultaneity#The_train-and-platform_thought_experiment ] two events A and B (reaching of light at the back and front of train) are declared not simultaneous without observing them.

In the thought experiment as described in the Wikipedia article there are relevant elements of the train being explicitly considered and named ("the front and back of the traincar"); and the "one observer midway inside a speeding traincar" is obviously meant as the "middle between the front and back of the traincar", and thus capable of contributing observations for determining whether and which indications of these two "ends of the traincar" were simultaneous or dissimultaneous to each other.
But there are no corresponding particular elements of the "platform" being named; and there is apparently no consideration given to the "other observer standing on a platform as the train moves past" having been "middle between" particular elements of the "platform", and thus capable of contributing observations for determining whether and which indications of any such elements of the "platform" might have been simultaneous or dissimultaneous to each other.
(In fairness: there is a similar shortcoming in Einstein's (related, though not quite equal) discussion of http://en.wikisource.org/wiki/Relativity:_The_Special_and_General_Theory/Part_I#Section_9_-_The_Relativity_of_Simultaneity -- where the two "places of the embankment, $A$ and $B$" are explicitly named and $M$ is the "middle between $A$ and $B$"; but "ends" or corresponding "elements of the train" are not explicitly and distinctly named, making it difficult to identify and address "$M'$" explicitly as a "middle between" particular participants.)

Are events in this experiment simultaneous if observed?

It has been noted (correctly) in other answers already that "one cannot say that two events are simultaneous in an absolute way".
Simultaneity or dissimultaneity are not attributes of pairs of entire events; because each event may always involve several distinct participants (such as "the front of the traincar" and some particular element of the embankment, passing each other) who were not at rest to each other.
(That's "what simultaneity is not", or "what dissimultaneity is not", respectively.)     
Instead, as indicated above, and as apparent in Einstein's definition, but unfortunately not explicitly verbalized by Einstein:
Simultaneity or dissimultaneity are attributes of pairs of indications of individual distinct participants who were at rest to each other (such as two "places of the railway embankment"; or two "ends of a train").
A: Sure, flash lights sent within the train cannot be observed by the observer on the platform. Nobody (or no device) can see light that is not sent directly to its eyes (receiver). Observer on the platform can see light sent within the train only if it is resent to him after it travelled locally (i.e. unseen directly from the platform).
Therefore, there will never be any problems with simultaineity. Both flashes reach the ends of the train simultaneously as registered by local (travelling) devices, and nobody on the platform can see them directly. There is no way to see the flashes reach the ends of the train dissimultaneousy.
