Barga DIDC'14 Invited Talk

50 %
50 %
Information about Barga DIDC'14 Invited Talk

Published on April 24, 2016

Author: rsbarga

Source: slideshare.net

1. A Look into the Future by Learning from the Past Roger S. Barga Cloud Machine Learning, Cloud and Enterprise Microsoft Corporation

2. This isn’t an academic talk…

3. This isn’t an applied research talk…

4. 1 1 5 4 3 7 5 3 5 3 5 5 9 0 6 3 5 2 0 0

5. 1. Learn it when you can’t code it 2. Learn it when you can’t scale it 3. Learn it when you have to adapt/personalize 4. Learn it when you can’t track it

6. • Distributed computing and storage • Deep Neural Networks • Learning = Scalable, Adaptive Computation for Various Big Data 2011 (“Big Data, DNN”) • Wide application in products • Statistical Modeling of Data • Learning = Parameter Estimation or Inference 2005 (“Graphical Models”) • Statistical Learning Theory • Scoring Systems • Learning = Optimization of Convex Functions 2000 (“Kernel Machines”) • Expert Systems • Decision-Tree Learning (C4.5) • Learning = Methods to automatically build Expert Systems 1990 (“Symbolic”) • Neural Networks • Artificial Intelligence • Learning = Adaptation of Neurons based on External Stimuli 1980 (“Neuro”)

7. • Distributed computing and storage • Deep Neural Networks • Learning = Scalable, Adaptive Computation for Various Big Data 2011 (“Big Data, DNN”) • Wide application in products • Statistical Modeling of Data • Learning = Parameter Estimation or Inference 2005 (“Graphical Models”) • Statistical Learning Theory • Scoring Systems • Learning = Optimization of Convex Functions 2000 (“Kernel Machines”) • Expert Systems • Decision-Tree Learning (C4.5) • Learning = Methods to automatically build Expert Systems 1990 (“Symbolic”) • Neural Networks • Artificial Intelligence • Learning = Adaptation of Neurons based on External Stimuli 1980 (“Neuro”)

8. • Distributed computing and storage • Deep Neural Networks • Learning = Scalable, Adaptive Computation for Various Big Data 2011 (“Big Data, DNN”) • Wide application in products • Statistical Modeling of Data • Learning = Parameter Estimation or Inference 2005 (“Graphical Models”) • Statistical Learning Theory • Scoring Systems • Learning = Optimization of Convex Functions 2000 (“Kernel Machines”) • Expert Systems • Decision-Tree Learning (C4.5) • Learning = Methods to automatically build Expert Systems 1990 (“Symbolic”) • Neural Networks • Artificial Intelligence • Learning = Adaptation of Neurons based on External Stimuli 1980 (“Neuro”)

9. • Distributed computing and storage • Deep Neural Networks • Learning = Scalable, Adaptive Computation for Various Big Data 2011 (“Big Data, DNN”) • Wide application in products • Statistical Modeling of Data • Learning = Parameter Estimation or Inference 2005 (“Graphical Models”) • Statistical Learning Theory • Scoring Systems • Learning = Optimization of Convex Functions 2000 (“Kernel Machines”) • Expert Systems • Decision-Tree Learning (C4.5) • Learning = Methods to automatically build Expert Systems 1990 (“Symbolic”) • Neural Networks • Artificial Intelligence • Learning = Adaptation of Neurons based on External Stimuli 1980 (“Neuro”)

10. • Distributed computing and storage • Deep Neural Networks • Learning = Scalable, Adaptive Computation for Various Big Data 2011 (“Big Data, DNN”) • Wide application in products • Statistical Modeling of Data • Learning = Parameter Estimation or Inference 2005 (“Graphical Models”) • Statistical Learning Theory • Scoring Systems • Learning = Optimization of Convex Functions 2000 (“Kernel Machines”) • Expert Systems • Decision-Tree Learning (C4.5) • Learning = Methods to automatically build Expert Systems 1990 (“Symbolic”) • Neural Networks • Artificial Intelligence • Learning = Adaptation of Neurons based on External Stimuli 1980 (“Neuro”)

11. The future will belong to those who can turn their historical data into predictive models…

12. Vision Analytics Recommenda- tion engines Advertising analysis Weather forecasting for business planning Social network analysis Legal discovery and document archiving Pricing analysis Fraud detection Churn analysis Equipment monitoring Location-based tracking and services Personalized Insurance Machine learning and predictive models are core new capabilities that will touch everything in the new enterprise

13. training data (expensive) synthetic training data (cheaper)

14. solve hard problems value from Big Data data analytics

15. Machine learning enables nearly every value proposition of web search.

16. Hundreds of thousands of machines… Hundreds of metrics and signals per machine… Which signals correlate with the real cause of a problem? How can we extract effective repair actions?

17. solve hard problems value from Big Data data analytics human intelligence

18. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 WER %

19. Training English training data English words, with some errors English speech input More with fewer errors How much data? – About the same as a human needs… Runtime

20. Training English training data English words, with some errors English speech input with fewer errors Runtime Can we learn the internal representation of human speech? French training data or French words or French Chinese training data or Chinese words or Chinese

21. Shetland Sheepdog (0.72) Shoe Store (0.56) Attack Aircraft Carrier (0.81) Steel Arch Bridge (0.74) Ballplayer, Baseball Player (0.86) Catamaran (0.51) Wood Rabbit, Cottontail, Cottontail Rabbit (0.18)

22. The first image returned is Rajiv Gandhi (her husband) in the Answer. An image of Lindsay Lohan appears in the Images Answer not really X X

23. solve hard problems value from Big Data data analytics human intelligence engineering practices

24. intelligence will become ambient intelligence from machine learning

25. 55 57 59 61 63 65 67 69 71 Overall NDCG Bing NDCG Google NDCG   

26. The razor-toothed piranhas of the genera Serrasalmus and Pygocentrus are the most ferocious freshwater fish in the world. In reality they seldom attack a human. Template matching

27. The razor-toothed piranhas of the genera Serrasalmus and Pygocentrus are the most ferocious freshwater fish in the world. In reality they seldom attack a human. py pyg ygo goc Pygocentrus

28. The razor-toothed piranhas of the genera Serrasalmus and Pygocentrus are the most ferocious freshwater fish in the world. In reality they seldom attack a human. Sentence-level decoding The razor-toothed piranhas of the genera Serrasalmus and Pygocentrus are the most ferocious freshwater fish in the world. In reality they seldom attack a human.

29. Massive The Intelligent Cloud Machine Learning & Analytics Crowd Sourcing Massive & Diverse Data The Cloud - Where Everything Comes Together

Add a comment

Related pages

Invited Talks 2015 2014 - ETH Z

Invited Talks Ruth Signorell Page ... (Barga), Italy, 27 April - 02 May 2014. ... Chair talk, Niigata, Japan, ...
Read more

Schoenemann: Conferences

International Conferences ... Gordon Research Conference (invited): Lucca (Barga), Italy ... invited lecture Talk: ...
Read more

Talks - Javier A. Bargas-Avila - Google Sites

Javier A. Bargas-Avila. Search this site. About. CV. ... " User Experience @ Google", talk at the lecture "Applied Computer Sciences" at Department of ...
Read more

IC2E 2016 - IEEE Computer Society

IC2E 2016 Keynotes and Invited Talk: IC2E Keynotes: ... Roger Barga (General Manager, Amazon Web Services) Title: Processing Big Data in Motion.
Read more

Fachvorträge+Poster (ab 1997) | Vorträge | Clusterphysik ...

... "Clusters as new materials" (invited talk). 2002. ... Barga, Italy, G. Ganteför, "Hydrogen Chemisorption to Metal Clusters" (invited talk)
Read more

DIDC 2014 - Fatih Üniversitesi

... (DIDC'14) Vancouver, Canada, ... DIDC 2012; Keynote Address. ... In this talk I will present a brief overview of machine learning, ...
Read more