OUCH - Outing Unfortunate Characteristics of Hidden Markov Models

Explore the limitations and problematic aspects of Hidden Markov Models in this 45-minute seminar presented by Jordan Cohen from Spelamode at the Center for Language and Speech Processing at Johns Hopkins University. Examine the unfortunate characteristics and inherent weaknesses of HMMs that researchers and practitioners should be aware of when applying these statistical models to language and speech processing tasks. Gain critical insights into the theoretical and practical shortcomings of Hidden Markov Models through detailed analysis and discussion of their structural limitations, computational challenges, and performance constraints in real-world applications.