Randomized Dimensionality Reduction in Machine Learning

Channel:
Microsoft Research
Subscribers:
344,000
Published on ● Video Link: https://www.youtube.com/watch?v=WBi3c0IF7qA


Duration: 36:34
1,532 views
9

We show how certain random projections and random sampling methods can be used to design efficient dimensionality reduction techniques for two popular machine learning problems: (i) K-means Clustering, and (ii) Canonical Correlation Analysis. In both cases, we argue that randomized dimensionality reduction is provably efficient.



Other Videos By Microsoft Research


2016-08-11ChronoZoom: Bridging the Gap Between the Humanities and Sciences
2016-08-11Robust Subspace Modeling
2016-08-11New Results in the Theory of Clustering
2016-08-11The Multi-Armed Bandit Problem and Thompson Sampling
2016-08-11Result Aggregation and Risk Estimation in Crowdsourcing
2016-08-11Random Features for Kernel Learning
2016-08-11The Role of Technology in Human Trafficking
2016-08-11Summer Number Theory Day; Session 2
2016-08-11The Changing Landscape of Parallel Computing - Programming Systems
2016-08-11Botnet Detection Using Structured Graph Analysis
2016-08-11Randomized Dimensionality Reduction in Machine Learning
2016-08-11Outlier Purusit: Robust PCA and Coolaborative Filtering
2016-08-11Big Heritage, Big Quilts, and Big Canvases
2016-08-11Summer Number Theory Day; Session 4 - Derivatives of $p$-adic $L$-functions
2016-08-11The Changing Landscape of Parallel Computing ΓÇô I
2016-08-11The Changing Landscape of Parallel Computing - Architecture
2016-08-11The Changing Landscape of Parallel Computing: Can Industry & Academia Collaborations be Effective?
2016-08-11Learning Hierarchical Similarity Metrics
2016-08-11Learning to Identify Author : A ML Approach to Forensic Linguistic
2016-08-11Kernel Learning with a Million Kernels
2016-08-11Active Learning of Actions


Tags:
microsoft research