Machine Learning & AI

https://azure.microsoft.com/en-us/overview/what-is-machine-learning-platform/

 

What is machine learning and how does it work?

Machine learning (ML) is the process of using mathematical models of data to help a computer learn without direct instruction. It’s considered a subset of artificial intelligence (AI). Machine learning uses algorithms to identify patterns within data, and those patterns are then used to create a data model that can make predictions. With increased data and experience, the results of machine learning are more accurate—much like how humans improve with more practice.
The adaptability of machine learning makes it a great choice in scenarios where the data is always changing, the nature of the request or task are always shifting, or coding a solution would be effectively impossible.

How machine learning relates to AI:

Machine learning is considered a subset of AI. An “intelligent” computer thinks like a human and carries out tasks on its own. One way to train a computer to mimic human reasoning is to use a neural network, which is a series of algorithms that are modeled after the human brain.

How machine learning relates to deep learning:

Deep learning is a specialized form of machine learning, using [multiple hierarchical levels of] neural networks (NN) to deliver answers [with higher levels of accuracy]. Able to determine accuracy on its own, deep learning classifies information like a human brain—and powers some of the most human-like AI.

The benefits of machine learning

• Identify patterns  within both structured and unstructured data, helping gain insight. 
• Data mining to help improve data integrity.
• Provide adaptive interfaces, targeted & personalized content, chatbots, and voice-enabled virtual assistants, anticipate customer behaviour, helping  enhance & optimize user experience
•  Automate processes, freeing up resources and improving productivity 

How does ML solve problems:

• Collect & compile data
• Clean data (provide structure, ensure data integrity, resolve inconsistencies)
• Train models with large datasets
• Evaluate / test models and select best performing model
• Apply model and predict outcomes

What does ML do:

• Predict outcomes based on regression models and studying cause-effect relationships, used for sales forecasts, expense estimates, anticipate customer requirements
• Detect anomalies, outliers & abnormal behaviour, used in fraud detection and identifying equipment malfunction
• Cluster and classify data to provide structure to data to aid in analysis

Machine learning algorithms identify patterns within data, helping data scientists solve problems. Machine learning algorithms can predict values, identify unusual occurrences, determine structure, and create categories. Depending upon the type of data you have and the outcome you’re looking for, you’ll use different algorithms. Algorithms are typically grouped by technique (supervised learning, unsupervised learning, or reinforced) or by family of algorithm (including classification, regression, and clustering).

Applications of ML in various industries:

Banking/Finance:  risk management & fraud prevention

Healthcare: diagnostic tools, patient health monitoring, outbreak prevention

Transportation: Delivery route optimization, traffic anomaly detection, self driving cars

Customer service: virtual assistants, chatbots, speech recognition, targeted / personalized ads, anticipating customer needs

Retail: Analyzing buying trends/patterns, improving customer experience

Agriculture: Harvest forecasting based on seasonal patterns

________________________________________________________

https://azure.microsoft.com/en-us/overview/cloud-computing-dictionary/

https://docs.microsoft.com/en-us/azure/machine-learning/service/concept-deep-learning-vs-machine-learning

Deep learning is a subset of machine learning that's based on artificial neural networks. The learning process is deep because the structure of artificial neural networks consists of multiple input, output, and hidden layers. Each layer contains units that transform the input data into information that the next layer can use for a certain predictive task.

 Machine learning is a subset of artificial intelligence that uses techniques (such as deep learning) that enable machines to use experience to improve at tasks.

By using machine learning and deep learning techniques, you can build computer systems and applications that do tasks that are commonly associated with human intelligence. These tasks include image recognition, speech recognition, and machine translation.

In machine learning, the algorithm needs to be told how to make an accurate prediction by consuming more information (for example, by performing feature extraction). In deep learning, the algorithm can learn how to make an accurate prediction through its own data processing, thanks to the artificial neural network structure.

	All machine learning Only deep learning Number of data points Can use small amounts of data to make predictions. Needs to use large amounts of training data to make predictions. Hardware dependencies Can work on low-end machines. It doesn't need a large amount of computational power. Depends on high-end machines. It inherently does a large number of matrix multiplication operations. A GPU can efficiently optimize these operations. Featurization process Requires features to be accurately identified and created by users. Learns high-level features from data and creates new features by itself. (Self-trained through several levels of hierarchical learning) Learning approach Divides the learning process into smaller steps. It then combines the results from each step into one output. Moves through the learning process by resolving the problem on an end-to-end basis. (end-to-end complex problem solving) Execution time Takes comparatively little time to train, ranging from a few seconds to a few hours. Usually takes a long time to train because a deep learning algorithm involves many layers. Output The output is usually a numerical value, like a score or a classification. The output can have multiple formats, like a text, a score or a sound.
	Because of the artificial neural network structure, deep learning excels at identifying patterns in unstructured data such as images, sound, video, and text. For this reason, deep learning is rapidly transforming many industries, including healthcare, energy, finance, and transportation.
	Some of the most common applications for deep learning include named-entity recognition, object detection, image recognition / caption generation, machine translation (including speech to text, or between languages, speech/sound/music recognition), textual data analytics.

Neural networks

The feedforward neural network is the most basic type of artificial neural network. In a feedforward network, information moves in only one direction from input layer to output layer. Feedforward neural networks transform an input by putting it through a series of hidden layers.

Recurrent neural networks save the output of a layer and feed it back to the input layer to help predict the layer's outcome. They're widely used for complex tasks such as time series forecasting, learning handwriting and recognizing language.

Convolutional neural networks are more complex 3D structures and used for video recognition, image recognition and recommender systems.

Big Data and Analytics

Big data is a large Volume of high-Velocity information from highly Varied information resources, which require innovative and cost-effective forms of automated information processing that allows for better understanding, decision-making, and communication. 

Sources of data: business transactions & records, addresses/contacts/emails, scientific and industrial sensors, medical devices, social media, satellite imagery, video surveillance, digital content, software tools, web repositories.

There are several Big Data technologies, tools and utilities used for data acquisition/preparation, processing, analysis &  reporting that are different from traditional database systems. These can be broadly classified into the categories of Storage, Analytics, Mining & Visualization.

Hadoop: Popular storage technology supporting distributed data batch processing
Spark: Real-time processing & streaming
Other examples: MongoDB, Hunk, Cassandra, Presto, ElasticSearch, Kafka, Splunk, R-language, Tableau & Plotly.

Hadoop is data processing engine that processes large volumes of data in batches in a distributed parallel environment across clusters & computers. Although more secure and cheaper than Spark, this is slower and more complex to use.

Spark is data analytics engine that processes real-time data designed for high-speed / low latency computing. This is less secure and more expensive to use (due to in-memory computational needs), but is faster and easier to use than Hadoop.

Analysis of Big Data is used by organizations to develop various forms of business intelligence / analytics to aid in decision making, help build customer relationships, provide targeted advertising and personalized services, assess marketing trends, and for business forecasting. Nature of this analysis may be Descriptive, Predictive and Prescriptive.

Data mining is a machine learning technique that can analyze big data and uncover interesting  correlations in large data sets. Basic data mining methods use clustering, classification, regression trees and neural networks. Data mining can be applied on databases, data warehouses / lakes and repositories.

Business analytics tools are types of application software, ranging from spreadsheets with statistical functions through sophisticated data mining tools that retrieve data from one or more business systems and combine it in a repository, such as a data warehouse, to be reviewed and analyzed.  Business analytics provide key insights and understanding of the business so smarter decisions may be made regarding business operations, customers and more.

While business intelligence tools also collect and display aggregate data, business analytics tools go a step further to not only report the results of the data, but explain why the results occurred to help identify weaknesses, fix potential problem areas, alert decision makers to unforeseen events, and even forecast future results based on decisions the company might make.

Business intelligence (BI) tools are types of application software that collect and process large amounts of unstructured data from internal and external systems, including books, journals, documents, health records, images, files, email, video, and other business sources. While not as flexible as business analytics tools, BI tools provide a way of amassing data to find information primarily through queries. These tools also help prepare data for analysis so that you can create reports, dashboards, and data visualizations. The results give both employees and managers the power to accelerate and improve decision making, increase operational efficiency, pinpoint new revenue potentials, identify market trends, report genuine KPIs, and identify new business opportunities.

MACHINE LEARNING / DEEP LEARNING / NEURAL NETWORKS with HPC / SUPERCOMPUTING capabilities is supported by chips such as Nvidia's Tesla/Volta and Google TPU, and requires HBM memories and advanced packaging technologies such as 2.5D/TSV for package integration.