In this section we will compare our work with that of others that have applied machine learning techniques to the same data sets. First we will discuss the two chunking tasks and then the tasks that required output of embedded structures. Many systems have been applied to the five tasks. Rather that giving a detailed description of each of them, we will list the best performing systems for each task and mention some differences between these systems and ours. This comparison of our memory-based shallow parsers with other work shows that they produce state-of-the-art results for the chunking tasks but not for the tasks which require identification of embedded structures.