What is the primary goal of a language model?

What is the primary goal of a language model?

• A. Classification of images into categories
• B. Storing data for retrieval
• C. Translating speech to text
• D. Probability distribution over a sequence of words

Answer: (D)

The essential aim of a language model is to assign probabilities to sequences of words, i.e., model the probability distribution over word strings. This lets the model judge how plausible a given sentence is, predict the next word given preceding words, and generate new text by sampling from that distribution. During training, it learns by maximizing the likelihood of observed text, so fluent and common sequences get higher probabilities. This probabilistic view—capturing P(w1, w2, ..., wn) or P(next | history)—is what enables both generation and scoring in language tasks. The other options describe different problems (image classification, data storage, or speech-to-text) that aren’t the core aim of a language model, even though language models can be components in broader systems.

The essential aim of a language model is to assign probabilities to sequences of words, i.e., model the probability distribution over word strings. This lets the model judge how plausible a given sentence is, predict the next word given preceding words, and generate new text by sampling from that distribution. During training, it learns by maximizing the likelihood of observed text, so fluent and common sequences get higher probabilities. This probabilistic view—capturing P(w1, w2, ..., wn) or P(next | history)—is what enables both generation and scoring in language tasks. The other options describe different problems (image classification, data storage, or speech-to-text) that aren’t the core aim of a language model, even though language models can be components in broader systems.