Manolis Kellis
@manoliskellis1·24.9K subscribers·255 videos
MIT CompBio Lecture 05 - HMMs Hidden Markov Models II (Fall'19)
Posted
September 19, 2019
Views
3,524
Likes
37
Comments
2
Engagement
1.11%
September 19, 2019
3,524
37
2
1.11%
Every word spoken in this episode is indexed. Type any phrase to jump straight to the moment it was said.
Type any word or phrase that may have been spoken. Click a result to seek the player to that exact moment.
Try a name, a topic, or a quoted line
See what was published immediately before and after this episode.
1:15:45Now PlayingMIT Computational Biology: Genomes, Networks, Evolution, Health
Prof. Manolis Kellis
Full playlist with all videos in order is here
All slides from Fall 2019 are here
Outline for this lecture:
1. Review: Basics and algorithms from last time
- P(x,π). Viterbi: π*=argmaxπ P(x,π). Forward: P(x)
- Posterior Decoding: Find most likely state πi, all paths.
2. Increasing the ‘state’ space / adding memory
- Finding GC-rich regions vs. finding CpG islands
- Gene structures GENSCAN, chromatin ChromHMM
3. Learning (ML training, Baum-Welch, Viterbi training)
- Supervised: Find ei(.) and aij given labeled sequence
- Unsupervised: given only x annotation + params
4. Conditional Random Fields (CRFs) & dependencies
- HMM limitations for diverse inputs, model dependencies
- CRF definitions. Modeling dependences with CRFs.
- Expressing HMMs as a special case of CRFs
Sentinel Indexing in Progress
Metadata and chapters are available. Claim extraction for this episode is pending.
All video content is delivered via YouTube embedded players in accordance with the YouTube Terms of Service. Sentinel provides research tools that promote discovery and accountability across political media.
