Abstract
Machine learning and Statistical techniques are powerful analysis tools yet to be incorporated in the new multidisciplinary field diversely termed as natural language processing (NLP) or computational linguistic. The linguistic knowledge may be ambiguous or contains ambiguity; therefore, various NLP tasks are carried out in order to resolve the ambiguity in speech and language processing. The current prevailing techniques for addressing various NLP tasks as a supervised learning are hidden Markov models (HMM), conditional random field (CRF), maximum entropy models (MaxEnt), support vector machines (SVM), Naive Bays, and deep learning (DL). The goal of this survey paper is to highlight ambiguity in speech and language processing, to provide brief overview of basic categories of linguistic knowledge, to discuss different existing machine learning models and their classification into different categories and finally to provide a comprehensive review of different state of the art machine learning models with the goal that new researchers look into these techniques and depending on these, develops advance techniques. In this survey we reviewed how avant-grademachine learning models can help in this dilemma.