IIT Madras ADS Lectures: Algorithms for Data Science
BSDA500350 lectures10 weeks
Weeks
- Week 19 lecturesIntroduction to algorithms for data science · Concentration inequalities - markov's inequality · Concentration inequalities - chebyshev's inequality · Concentration inequalities - hoeffding's inequality · Useful inequalities in probability · Johnson-lindenstrauss lemma - introduction · Jl lemma, part-1, no isometry exists · Jl lemma, part-2, approximate isometry · Jl lemma, part-3, goodness of estimate
- Week 23 lecturesJl lemma, part-4, towards the final bound · Jl lemma, part-5, final bound · Jl lemma - linear algebra interpretation
- Week 38 lecturesApproximate nearest neighbours · Locality sensitive hashing - introduction · LSH - hash family · LSH - collision analysis · LSH - true positive rate and false positive rate · LSH - reducing fpr with "and" · LSH - increasing TPR with "OR" · LSH - summary and analysis
- Week 45 lecturesSVD - recap and applications · From naive svd to fast svd · Randomized SVD - part-1 · Randomized SVD - part-2 · Randomized SVD - part-3
- Week 53 lecturesClustering - recap · Spectral clustering - graph cuts · Spectral clustering - mincut
- Week 64 lecturesSpectral clustering - algorithm · Spectral clustering at scale - spectral sparsification, part-1 · Spectral clustering at scale - spectral sparsification, part-2 · Spectral clustering at scale - summary
- Week 73 lecturesStatistical learning theory - underlying distribution · Statistical learning theory - test error and bayes classifier · Statistical learning theory - error decomposition
- Week 82 lecturesStatistical learning theory - generalization error, sample complexity, part-1 · Statistical learning theory - generalization error, sample complexity, part-2
- Week 95 lecturesInfinite hypothesis classes and labelings · VC dimension - introduction · VC dimension - axis parallel rectangles hypothesis class · Finite VC dimension and uniform convergence · VC dimension - linear hypothesis class
- Week 104 lecturesGeneralization error - recap · Estimation error and erm · Approximation error and srm · Sample compression