IBM Watson
@RomeoKienzler
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1502091/Twitter_logo_white.png)
the goal
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494449/brain2colors.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494451/BULB02.png)
the goal
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494449/brain2colors.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494451/BULB02.png)
you: 20 Watts
the goal
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494449/brain2colors.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494451/BULB02.png)
you: 20 Watts
IBM Watson: 200.000 Watts
the human brain
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494466/neurone.png)
the human brain
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494463/58294main-The.Brain.in.Space-32-science-diagram-of-synapse.png)
Supervised Machine Learning
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494479/Linear_Regression.png)
Supervised Machine Learning
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494500/GI5ax.png)
y = \alpha + \beta_{1} x_1 +\beta_{2} x_2 +\beta_{3} x_2 +\beta_{4} x_4+\beta_{5} x_5
y=α+β1x1+β2x2+β3x2+β4x4+β5x5
Source: http://i.stack.imgur.com/GI5ax.png
functions beyond linearity
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494606/index.png)
Neural Networks
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494591/500px-Colored_neural_network.svg.png)
Watson (distance view)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494618/Space-Sketched.png)
Feature Extraction
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494624/Question-Guy.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494628/1384943818.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494630/1286980063.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494632/matt-icons_text-x-doc-cleaned.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494632/matt-icons_text-x-doc-cleaned.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494632/matt-icons_text-x-doc-cleaned.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494632/matt-icons_text-x-doc-cleaned.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494632/matt-icons_text-x-doc-cleaned.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494635/mystica-Arrow-set-Bulb-3.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494641/Untitled.png)
Training
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494628/1384943818.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494630/1286980063.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494635/mystica-Arrow-set-Bulb-3.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494641/Untitled.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494673/1382196361.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494689/isoleucine.png)
GO !!!
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494624/Question-Guy.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494630/1286980063.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494635/mystica-Arrow-set-Bulb-3.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494689/isoleucine.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494690/Exclamation-Mark.png)
Watson Architecture
Watson Architecture
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495011/Untitled.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495016/Untitled.png)
Query Decomposition
- Focus: this Michigan City
- LAT: Michigan
- Keywords: 1894, C.W. Post, created, warm, cereal, drink, Postum, Michigan, City
In 1894 C.W. Post created his warm cereal drink Postum in this Michigan city
Parsing
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1494996/Untitled.png)
Parsing
Watson Architecture
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495011/Untitled.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495029/Untitled.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495029/Untitled.png)
Primary Search
- Lucene and Indri search engine
- Preprocessing generated keyword->text based documents
- Keyword associated with found document added to candidate answer list
- Primary Search extended with all synonyms
Watson Architecture
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495011/Untitled.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495016/Untitled.png)
Hypothesis Generation
Unlike most sea animals, in the sea horse this pair of sense organs can move independently of one another
- A Sea Horse can move its eyes independently.
- A Sea Horse can move its ears independently.
- A Sea Horse can move its skin independently.
- A Sea Horse can move its nose independently.
- A Sea Horse can move its tung independently.
Watson Architecture
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495011/Untitled.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495029/Untitled.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495029/Untitled.png)
Evidence Scoring
- Generated Candidate Answer is
- ESG'd
- PAS'd
- searched against corpus
- LATs used to determine whether a candidate answer is an instance of the answer types
Evidence Scoring
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495312/Untitled.png)
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495322/Untitled.png)
Oncology Expert Advisor
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495377/putting-ibm-watson-to-work-saxena-14-728.jpg)
Oncology Expert Advisor
- Dynamic patient summary
- Evidence-based treatment options
- Care pathway advisory
![](https://s3.amazonaws.com/media-p.slid.es/uploads/167657/images/1495394/Untitled.png)
- N=400 historical patient cases
- false-positive rate 2.9%
- false-negative rate 0.4%
- accuracy 82.6%
Watson in the Cloud
Do it yourself!
http://ibm.biz/joinIBMCloud
Personality Insights
Bibliography
- Wie lernen Kinder? Aktuelles aus der Gehirnforschung, Manfred Spitzer
- IEEE IBM Journal of Research and Development, Issue 3.4, This is Watson
- IBM Watson for Oncology
-
Superintelligence: Paths, Dangers, Strategies, Bostrom, Nick
- IBM Cloud
IBM Watson
By Romeo Kienzler
IBM Watson
A short overview how IBM Watson works
- 2,997