Summary
In this chapter, first we described two tasks which are tackled in the thesis. We define these tasks formally, then look at the general supervised approach employed to solve them, and finally look at the different evaluation schemes used to assess the systems.
Next, we briefly introduced neural networks which are used for effective representation in general also in our experiments.
Further, we detailed about representation learning and its related fields. We presented different approached used for learning representations for words. Also, different general approaches to compose word representations to get phrase, sentence and document representations.
The general composition approaches do not effectively capture task-specific information. Therefore, particular approaches are used to get task-specific linguistic representations. We detail them in the next chapter.