Given a bigram language model, in what scenarios do we encounter zero probabilities? How should we handle these situations ?

Recall the Bi-gram model can be expressed as :     Following scenarios can lead to zero probability in the above expression : Out of vocabulary(OOV) words – such words may not be present during training and hence any probability term involving OOV words will be 0.0┬áleading entire term to be zero. This is solved…