https://arshadhs.github.io/tags/pos-tagging/Recent content in POS Tagging on Arshad SiddiquiHugoen-ushttps://arshadhs.github.io/docs/ai/050-nlp/070-statistical-ml-and-neural-models-of-pos-tagging/Mon, 01 Jan 0001 00:00:00 +0000https://arshadhs.github.io/docs/ai/050-nlp/070-statistical-ml-and-neural-models-of-pos-tagging/<h1 id="statistical-ml-and-neural-models-of-pos-tagging"> Statistical, ML and Neural Models of POS Tagging <a class="anchor" href="#statistical-ml-and-neural-models-of-pos-tagging">#</a> </h1> <p>&lsquo;HMM Inference: Forward and Viterbi Algorithms&rsquo; covers the portion:</p> <ul> <li>Forward Algorithm</li> <li>Viterbi Algorithm</li> <li>HMM inference for POS tagging</li> </ul> <p>The complete Topic also includes:</p> <ul> <li>Maximum Entropy Markov Models</li> <li>Bidirectionality</li> <li>Neural-network models for POS tagging</li> </ul> <h1 id="hmm-inference-forward-and-viterbi-algorithms"> HMM Inference: Forward and Viterbi Algorithms <a class="anchor" href="#hmm-inference-forward-and-viterbi-algorithms">#</a> </h1> <p>Hidden Markov Models create two closely related inference problems:</p> <ul> <li><strong>Likelihood:</strong> How probable is an observed sequence under the model?</li> <li><strong>Decoding:</strong> Which hidden-state sequence most probably generated the observations?</li> </ul> <p>The Forward Algorithm solves the likelihood problem, while the Viterbi Algorithm solves the decoding problem. Both use dynamic programming and a trellis, but they combine paths differently.</p>