Extreme Learning Machine: Learning Without Iterative Tuning

Subscribers:
351,000
Published on ● Video Link: https://www.youtube.com/watch?v=KiLuYCC33-Y



Duration: 1:33:32
2,603 views
26


Neural networks (NN) and support vector machines (SVM) play key roles in machine learning and data analysis. However, it is known that there exist some challenging issues with them such as intensive human intervene, slow learning speed, poor learning scalability. This talk will introduce a new learning technique referred to as Extreme Learning Machine (ELM). ELM not only learns up to tens of thousands faster than NN and SVMs, but also provides unified implementation for regression, binary and multi-class applications. ELM is efficient to time series, online sequential, incremental applications. ELM is efficient for large datasets.




Other Videos By Microsoft Research


2016-08-16Telling Stories in the Cloud; Communications from the Particle Frontier; Video Analytics
2016-08-16On Users' Mental Models of Security Controls
2016-08-16Why Don't Software Developers Use their Tools?
2016-08-16The Mathematics of Side-Channel Attacks
2016-08-16PyPy's Approach to Implementing Dynamic Languages Using a Tracing JIT Compiler
2016-08-16Fine-Grained Power Modeling for Smartphones Using System Call Tracing
2016-08-16Reputational Bargaining Under Knowledge of Rationality
2016-08-16We Will be Right With You: Managing Customers Expectations with Vague Promises and Cheap Talk
2016-08-16Information That Matters: Investigating Relevance of Entities in Social Media Networks
2016-08-16Efficient Bayesian Algorithmic Mechanism Design
2016-08-16Extreme Learning Machine: Learning Without Iterative Tuning
2016-08-16Extracting Knowledge from Networks: Rumors, Superstars, and Communities
2016-08-16Electrical Flows and Laplacian Systems: A New Tool for Graph Algorithms
2016-08-16Limit Theorems in Pseudorandomness and Learning Theory
2016-08-16Empirical Software Engineering, Version 2.0
2016-08-16A Master Bijection for Planar Maps, and Its Applications
2016-08-16Password-based Authenticated Key Exchange at the Cost of Diffie-Hellman
2016-08-16Generalized Identity-Based Encryption
2016-08-16Cryptography with Weak, Noisy, Leaky and Tempered Keys
2016-08-16Electrical Flows and Laplacian Systems: A New Tool for Graph Algorithms
2016-08-16Broadcast Based Peer Review in Open Source Software Development Projects



Tags:
microsoft research