hidden Markov models

Latent state machines

Hidden Markov models are statistical systems where an underlying process moves through unobservable states, each producing observable outputs with certain probabilities.

The "hidden" refers to these invisible states—you see only the evidence they generate, not the states themselves.

Used extensively in speech recognition, genomics, and natural language processing: The algorithm used a hidden Markov model to predict which words the speaker intended.

speech

speech recognition

Machines understanding human talk

decoder · word recognition · voice · speech synthesis · natural language processing · dictation · transcription

probability

stochastic

Randomly determined, statistically describable

Bayesian · likelihood · Monte Carlo · statistical · estimation · parameters · latent · conditional · prior · posterior

bioinformatics

sequence alignment

Arranging molecular strings for comparison

transcription · annotation · homology · CpG islands · DNA sequence · amino acid · evolutionary

Categories
Opposites
stochastic vs deterministic modeling

hidden Markov modelsexpert systems

AI using encoded rules

regular expressions · decision trees

model complexity/expressiveness

hidden Markov modelstransformers

Voltage-altering electrical devices

Family

model, hidden, hiding, models, modeling

Familiarity
Top 1,000,000
Profile
Used worldwide · Timeless · Abstract · Rated G