Parity function Neural Networks
I have read an interesting paper on limitations of machine learning models: Scaling Learning Algorithms towards AI. It mentions limitation of two-layer neural networks and other two-layer models (SVMs). These shallow models are unable to learn some functions without an exponential number of components. For example, to learn the parity function over N input bits, they would need 2N hidden neurons.
On the other hand, a deep model with N layers could compute the parity with just N components.
Source: Lessoned Code
You might also like:
Foreign Exchange trading up 21%-- Adios $2005-01-12 07:32:11 by puro
Did the Romans curse Caligula, or only in historical retrospect?
"Forex trade volumes hit record levels
By Jennifer Hughes in New York
Published: January 11 2005 17:26 | Last updated: January 11 2005 17:26
Foreign exchange trading volumes leapt to record levels in the first week of 2005, according to EBS, the largest interbank trading platform, and the Chicago Mercantile Exchange.
Average daily volume for the week on EBS reached $162bn, up 21 per cent on the same week last year and considerably higher than the average of about $100bn. Total volumes for the week were $811bn.
Trading in FX products over the CMEs Globex electronic platform last week was up 182 per cent over the same period in 2004, and pit-traded options volumes were...
Profit & Loss Names CitiFX Pro Best Retail Platform for Margin FX Trading for … — PR Web
CitiFX Pro is Citi's online forex trading platform for active individual and small institutional clients including commodity trading advisors, broker-dealers, money managers, and hedge funds. Additional information may be found at …
Financial Analysts Journal - Vol. 49, No 4 - July/August 1993: Is Beta Dead Again?; Using Artificial Neural Networks to Pick Stocks; The Value of Control
Book (Association for Investment Management and Research)