How would you best define a language model?

How would you best define a language model?

• A. A model that translates speech into text.
• B. A model that classifies documents by topic.
• C. A model that stores large corpora for retrieval.
• D. A model that assigns probabilities to word sequences in a language.

Answer: (D)

A language model is a model that assigns probabilities to word sequences in a language. It captures how likely a given sequence of words is, or, equivalently, the likelihood of the next word given the preceding words. This probabilistic view lets the model score sentences, choose plausible continuations, and generate text by selecting the most probable next word at each step. This isn’t about translating speech, classifying documents by topic, or simply storing large collections. Those are separate tasks—speech recognition turns audio into text, document classification assigns topics to texts, and storing corpora relates to retrieval or storage. A language model’s defining role is to learn and use the distribution over language itself, typically expressed as probabilities for word sequences.

A language model is a model that assigns probabilities to word sequences in a language. It captures how likely a given sequence of words is, or, equivalently, the likelihood of the next word given the preceding words. This probabilistic view lets the model score sentences, choose plausible continuations, and generate text by selecting the most probable next word at each step.

This isn’t about translating speech, classifying documents by topic, or simply storing large collections. Those are separate tasks—speech recognition turns audio into text, document classification assigns topics to texts, and storing corpora relates to retrieval or storage. A language model’s defining role is to learn and use the distribution over language itself, typically expressed as probabilities for word sequences.