Multilingual models in NLP - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is a multilingual model in AI?

A multilingual model is an AI system trained to understand and generate text in multiple languages, allowing it to work across different languages without needing separate models for each.

beginner

Why are multilingual models useful in real life?

They help people communicate and access information across languages, like translating websites or understanding customer feedback from different countries, all with one model.

intermediate

How do multilingual models learn multiple languages?

They are trained on large collections of text from many languages, learning patterns and meanings that apply across languages, so they can understand and generate text in each one.

intermediate

What is a challenge when building multilingual models?

One challenge is balancing performance: the model might do well in popular languages but less well in rare ones because it sees less data for those languages during training.

beginner

Name a popular multilingual model used in natural language processing.

One popular multilingual model is mBERT (multilingual BERT), which supports over 100 languages and is widely used for tasks like translation and text classification.

What does a multilingual model do?

A Processes multiple languages with one model

B Uses one model per language

C Only translates text

D Works only with English text

Multilingual models handle many languages using a single model, unlike separate models for each language.

Which is a common training method for multilingual models?

A Training on one language at a time separately

B Training only on English text

C Training on images only

D Training on text from many languages together

Multilingual models learn by training on mixed text data from many languages simultaneously.

What is a limitation of multilingual models?

A They only work with numbers

B They may perform worse on less common languages

C They cannot translate text

D They require separate models per language

Multilingual models often have less data for rare languages, which can reduce their accuracy for those languages.

Which model is an example of a multilingual model?

A VGG16

B ResNet

C mBERT

D GPT-1 English only

mBERT is a well-known multilingual model supporting many languages.

Why might companies use multilingual models?

A To handle customer support in many languages with one system

B To only work with English customers

C To avoid using AI

D To translate images

Multilingual models help companies serve customers speaking different languages efficiently.

Explain what a multilingual model is and why it is important.

Think about how one model can work with many languages.

Describe one challenge faced when training multilingual models and how it affects performance.

Consider how the amount of data per language matters.