Case Study Artificial Intelligence,ETL and Datawarehousing Examples part 1

Read : IPTV and Augmented Reality using Artificial Intelligence.
https://sandyclassic.wordpress.com/2012/06/27/future-of-flex-flash-gamification-of-erp-enterprise-software-augmented-reality-on-mobile-apps-iptv/

AI is there in many place like one area of AI Fuzzy Set there is already Fuzzy Transformation in SQL Server Integration Services since year 2010.
What it does Fuzzy logic Transformation achieve?
So when we match two records we do it by checking each alphabet using regular matches.
But when we use fuzzy logic it brings out similar sounding and combination matches although alphabet may not be same also it checks meaning is same. Even it can override spelling mistakes to get right results How?
Example Fuzzy logic in SSIS:
USA,us, united states – For country Any person can enter any of these combination.
Usually its taken up for Data cleansing.
If data is not cleaned using De-dup it may not show many of these records in result for matches.
But Fuzzy logic we use Fuzzy set from all records it creates fuzzy set of record with
Set A { ElementA, membershipOfElementA}
membershipOfElementA define in percentage terms the possibility of it being in the similarly grouped set.
{us,0.97} {united states,0.98} {usa,0.99} {united states of America,1} so we can set tolerance level to 3% then all of these matches are there in result.
code you can see at http://www.codeproject.com/Tips/528243/SSIS-Fuzzy-lookup-for-cleaning-dirty-data
SIRI:  Speech Recognition Search Which was introduced in iPhone long back takes speech.
Speech input to pressure sensor –> generate Waveform –> Then Compare wave form
That’s process but.
AI in SIRI
The Waveform may be amplitude modulated but yet same thing let suppose we say
Apple the Two Waveform compared may have boundary level aberrations which can be defined by membership function Then same result within same Tolerance limit can be deemed to be similar. This membership can be calculated each time person do a search dynamically when it says something in on Mike which repeat same process again.
There can be lots of image processing and AI search algorithm can be built to make better.
Like A* search etc.
Already if the words are linked can be understood by Neural Network. Similar way Neural Network is used to predict The  traffic congestion aggregating data paths from street light sensors in japan Tokyo.
Aggregation of words can be achieved by neural Network in not exact but similar way to some Extent. Thus completing the search.
This aggregation may be used in text, covariance matrix of images or covariance of sound score or speech search.
Using Laplace Transform’s Cross correlation. Read (http://en.wikipedia.org/wiki/Cross-correlation512px-Comparison_convolution_correlation.svg
Now TV is large platform just like difference between watching movie on laptop or TV Vs on 70 mm screen. Each of those has there own market.
Costly Miniaturisation
What effect you can provide on TV may not be provided on mobile until there is technical break through in miniaturisation. I am not saying it cannot be provided but it will require relative less technical  break through compared with miniaturized chips or may be less costly.
Second TV is like we have last mile connectivity in Telecom.
So When you have something to watch in any storage device you can just throw that on TV Ubiquitously . As TV would be there in every house so you need not carry screen to watch. Just like Last mile wireless connectivity using HotSpot.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s