Context based Text-generation using LSTM networks

Understand

Long short-term memory(LSTM) units on sequence-based models are being used in translation, question-answering systems, classification tasks due to their capability of learning long-term dependencies.

• In Natural language generation, LSTM networks are providing impressive results on text generation models by learning language models with grammatically stable syntaxes.
• But the downside is that the network does not learn about the context.
• The network only learns the input-output function and generates text given a set of input words irrespective of pragmatics.