Edit This Guide Record
Guides Technology Making the most of Predictive Analytics : Exploring Deep Learning

Making the most of Predictive Analytics : Exploring Deep Learning

Published on 11/17/2016 | Technology

246 1

Paul Pallath

Paul Pallath is a Senior Technology Leader, in quest of Excelling in a globally competitive and challenging business oriented environment in the field of Advanced Analytics/Business Intelligence/Internet of Things.

IoT GUIDE

Overview

At SAP, we see predictive analytics as a means for customers to gain insight into their business data in real time and use the information and knowledge embedded in the data for competitive advantage.   Today, we start a six-part blog series that will explore topics that are top of mind for organizations tackling predictive analytics – beginning with the topic of Deep Learning.  In recent years, deep-learning has taken the academic community and business world by storm. Deep Learning has gained increasing popularity with its ability to obtain good representations in machine-learning tasks.

 

Deep learning and data

Deep Learning allows distributed representation of data. Doing so allows the data to be configured into abstract features that are automatically captured and compactly represented across the hidden layers across the network.  Deep Learning Neural Networks have a two-phased approach for training where each layer is pre-trained with an unsupervised learning algorithm for capturing the main variations in the input data, followed by a final supervised training phase that fine tunes the deep learning model.

 

As a result, a system with deep architecture can still show a strong learning capacity while opening the door to a rich form of generalization, even if the problem being solved contains complex behaviors and highly varying mathematical functions. This machine-learning approach is powering the latest generation of commodity computing and deriving significant value from Big Data.

Application of deep learning

Since 2006, deep architectures have been enabling state-of-the-art performance. And with success, this technology has been applied across a wide range of fields such as classification, dimensionality reduction, robotics, image recognition, image retrieval, information retrieval, language modelling, and natural language processing. One whitepaper I recently read, “An Introduction to Deep Learning” by Dr Ying Wu & Dr. Rouzbeh Razavi, provides a high-level overview of deep learning and use cases of deep learning in learning complex machine learning problems like image, audio and language Recognition/ Classification.

Conclusion

Geoffrey Hinton, known as the “Father of Deep Learning,” is quoted as saying

                “Deep Learning is an algorithm which has no theoretical limitations of what it can learn; the more data you give and the more computational time you provide, the better it is”.

To learn how Deep Learning algorithms can be brought into SAP Predictive Analytics and use it for complex modelling, I invite you to read the white paper titled “Embed Deep – Learning Techniques into Predictive Modelling.”

test test