Learn by Source
DeepLearning.AI · on Coursera

Natural Language Processing with Probabilistic Models

Intermediate English SpecializationFREE
SkillsNatural language processingN-gram modelsHidden Markov modelsDynamic programmingPythonProbability

Is this course right for you?

Our take
The second course of DeepLearning.AI's NLP Specialization, building the probabilistic groundwork the field was built on.

Good for: The probabilistic foundations of NLP, hands-on in Python.

Skip if: You want a quick, library-based NLP tutorial or you lack Python/maths.

It works hands-on in Python through dynamic programming, N-gram language models, hidden Markov models and more — the classical mechanics that make the later deep-learning courses make sense. As with course one, that rigour sets the bar: you'll want Python and some maths, so it's not the place for a quick library-based tutorial or a gentle first look.

Worth knowing where it sits: this is deliberately the classical, pre-transformer half of the specialisation, with attention-based methods arriving in later courses. Free to audit on Coursera; a certificate needs a subscription (about $49/month) or Coursera Plus, with financial aid available. These probabilistic foundations remain worth knowing (as of 2026).

Comparison · LBS

Compare alternatives for Natural Language Processing with Probabilistic Models

Same topic, different options. We surface the trade-offs others hide so you can pick the course that actually fits your time, budget, and goals.
Coursera
Natural Language Processing with Probabilistic Models
Price
Free
Audit free · Certificate available
Duration
—
Level
Intermediate
Certificate
Specialization
Coursera
Natural Language Processing with Classification and Vector Spaces
Price
Free
Audit free · Certificate available
Duration
—
Level
Intermediate
Certificate
Specialization
View course →
Coursera4.9(39,000)
Machine Learning Specialization
Price
Free
Audit free · Certificate on subscription
Duration
86 hrs
Level
Intermediate
Certificate
Specialization
View course →
Coursera
Fitting Statistical Models to Data with Python
Price
Free
Audit free · Certificate available
Duration
—
Level
Intermediate
Certificate
Specialization
View course →
Prices & availability can change — confirm on the provider's site. We're not affiliated with any single provider.

About this course

Natural Language Processing with Probabilistic Models is Course 2 of DeepLearning.AI's NLP Specialization. It builds the probabilistic foundations of NLP: dynamic programming, N-gram language models, hidden Markov models, and Word2Vec — used to build autocorrect, autocomplete, and part-of-speech taggers in Python.

Browse all AI & ML courses Browse all Coursera courses

Instructor

YB
Younes Bensouda Mourri
Coursera instructor

Taught by Younes Bensouda Mourri (Stanford, DeepLearning.AI) and Łukasz Kaiser, a Google Brain research scientist and co-author of the Transformer paper and TensorFlow. The pairing brings both clear teaching and frontier research depth.

Frequently asked questions

It is strongly recommended. This is course two of DeepLearning.AI's NLP specialization, building on the classification-and-vector-spaces material from course one. You could take it alone with enough background, but it assumes familiarity with the earlier concepts, so most people take them in order. Arriving without that grounding makes the probabilistic-models material harder than it needs to be.

No — that comes later in the specialization. This course covers classical, probabilistic NLP techniques: things like autocorrect, part-of-speech tagging, and word models using probability and older sequence methods. Transformers and attention are taught in the specialization's later courses. So this gives you the foundational, pre-transformer techniques rather than the modern architectures behind today's large language models.

Yes, as foundation. Large language models dominate now, but understanding the classical probabilistic methods this course teaches builds genuine intuition for how NLP works underneath, which helps you grasp and use modern models more deeply. It is not the cutting edge, and you should pair it with transformer-based learning, but the fundamentals remain a worthwhile grounding rather than obsolete.

Yes, plus some machine-learning background. It involves Python programming assignments and assumes you already know the language and understand basic ML concepts. It is not a place to learn Python or ML from scratch. Arriving with comfortable Python and the earlier NLP course behind you is the right preparation; without those, the material moves too fast.

Yes, you can audit it on Coursera and watch the lectures at no cost. The graded programming assignments and the certificate require a subscription, with financial aid available. Since the hands-on coding is where NLP concepts genuinely stick, the paid track is worth considering if you want that practice rather than only the lectures.
Free
to audit
Enroll now