another. It is the automatic translation of text from one natural language (source) to another (target). Types of translation 1. Rough translation 2. Restricted source translation 3. Pre-edited translation 4. Literary translation Machine translation system If translation is to be done fluently and perfectly then the translator (human or machine) must read the original text, understood the situation to which it is referred and finds a good corresponding text in the target language describing same or similar situation. These systems vary in the level to which they analyze the text. Statistical machine translation It is a new approach proposed during last decade. Here the whole translation process is based on finding the most probable translation of a sentence, using data gathered from s bilingual process. The model for p (F/ E) has four set of parameters. 1. Language model 2. Translation model 3. Fertility model 4. Word choice model 5. Offset model Learning probabilities for machine translation Segment into sentence Estimate the French language model Align sentences Estimate the initial fertility model Estimate the initial choice model Estimate the initial offset model Improve estimates