Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling –...
-
Upload
francis-carpenter -
Category
Documents
-
view
216 -
download
0
description
Transcript of Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling –...
![Page 1: Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling – Optimization Modeling – Generative: we assume a probabilistic.](https://reader036.fdocuments.in/reader036/viewer/2022082600/5a4d1b207f8b9ab059994e64/html5/thumbnails/1.jpg)
Machine learning optimization
Usman Roshan
![Page 2: Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling – Optimization Modeling – Generative: we assume a probabilistic.](https://reader036.fdocuments.in/reader036/viewer/2022082600/5a4d1b207f8b9ab059994e64/html5/thumbnails/2.jpg)
Machine learning
• Two components:– Modeling– Optimization
• Modeling– Generative: we assume a probabilistic model and
optimize model parameters on data with maximum likelihood
– Discriminative: we select a model guided by the data
![Page 3: Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling – Optimization Modeling – Generative: we assume a probabilistic.](https://reader036.fdocuments.in/reader036/viewer/2022082600/5a4d1b207f8b9ab059994e64/html5/thumbnails/3.jpg)
Optimization
• Most machine learning problems are actually NP-hard
• Unsupervised learning:– Cluster data into two or more sets– NP-hard– K-means local search
• Supervised learning– Separating plane with minimum error
![Page 4: Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling – Optimization Modeling – Generative: we assume a probabilistic.](https://reader036.fdocuments.in/reader036/viewer/2022082600/5a4d1b207f8b9ab059994e64/html5/thumbnails/4.jpg)
Clustering• Suppose we want to cluster n vectors in Rd
into two groups. Define C1 and C2 as the two groups.
• Our objective is to find C1 and C2 that minimize
where mi is the mean of class Ci
![Page 5: Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling – Optimization Modeling – Generative: we assume a probabilistic.](https://reader036.fdocuments.in/reader036/viewer/2022082600/5a4d1b207f8b9ab059994e64/html5/thumbnails/5.jpg)
Clustering• NP hard even for 2-means
• NP hard even on plane
• K-means heuristic– Popular and hard to beat– Introduced in 1950s and 1960s
![Page 6: Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling – Optimization Modeling – Generative: we assume a probabilistic.](https://reader036.fdocuments.in/reader036/viewer/2022082600/5a4d1b207f8b9ab059994e64/html5/thumbnails/6.jpg)
Unsupervised learning• Also NP-complete (see paper by Ben-David, Eiron, and
Long)• Even NP-complete to polynomially approximate
(Learning with kernels, Scholkopf and Smola)
![Page 7: Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling – Optimization Modeling – Generative: we assume a probabilistic.](https://reader036.fdocuments.in/reader036/viewer/2022082600/5a4d1b207f8b9ab059994e64/html5/thumbnails/7.jpg)
Approximate approach
• To overcome NP-hardness we modify the problems to achieve convexity
• Convexity guarantees a unique optimum• Convexity also allow gives us a gradient
descent solution
![Page 8: Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling – Optimization Modeling – Generative: we assume a probabilistic.](https://reader036.fdocuments.in/reader036/viewer/2022082600/5a4d1b207f8b9ab059994e64/html5/thumbnails/8.jpg)
Gradient descent
•
![Page 9: Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling – Optimization Modeling – Generative: we assume a probabilistic.](https://reader036.fdocuments.in/reader036/viewer/2022082600/5a4d1b207f8b9ab059994e64/html5/thumbnails/9.jpg)
Local search
• Standard approach in computer science to solve hard problems
• At a high level it is a simple method:– Start with a random solution– Find neighborhood– Select point in neighborhood that optimizes
objective– Continue until local minima
![Page 10: Machine learning optimization Usman Roshan. Machine learning Two components: – Modeling – Optimization Modeling – Generative: we assume a probabilistic.](https://reader036.fdocuments.in/reader036/viewer/2022082600/5a4d1b207f8b9ab059994e64/html5/thumbnails/10.jpg)
Recent new approaches
• Direct optimization– Recently published work– Our own work: iterated local search
• Stochastic gradient descent