Chapter 10: AI

AI customer behavior created by

feeding large quantities of data to machine learning algorithm

How supervised learning works?

-Develop predictive model based on both input and output data -Trains a model to generate reasonable predictions for the response to new data. -Supervised learning uses classification and regression methods

AI Content Personalization

(Artificial) Intelligent Personalization: Algorithms that can personalize websites to individual site users Based on: Geographical Location Demographics Types of Device being Used On-Site Interaction Companies like WSJ, Pandora, and TopFan use AI to improve conversion rates

What is Deep Learning?

-A subset of machine learning based on algorithms inspired by the human brain -Utilizes neural networks- a computer system modeled after the human brain and nervous system -Uses a series of algorithms to attempt to identify relationships in data by copying the process that the human brain uses -Allows machines to adapt to changes in order to maximize results

AI MARKETING: Data Flow

-AI detects when unexpected traffic is visiting a site -Analytics is checked autonomously by AI platforms -Hunch: An AI platform that analyzes data from Google Analytics Hunch tracks: -Website Performance -Customer Experience -Advertising on return-on-investment -Summarizes performance insights based on data

SWARM AI: Marketing Applications

-Analysis allows marketers to make decisions that are lower in risk and with less data -Potential to reduce marketing costs → especially for newer companies -More accurately predict future movements in the market -Could replace expensive customer research and more efficiently create personas and characterizations

2 Categories: Artificial Intelligence

-Artificial Applied Intelligence -Artificial General Intelligence

ARTIFICIAL INTELLIGENCE TOPICS

-Artificial Intelligence -Machine Learning -Deep learning -Marketing Applications -SWARM Artificial

Things machine learning helps with?

-Computational finance, for credit scoring and algorithmic trading -Image processing and computer vision, for face recognition, motion detection, and object detection -Computational biology, for tumor detection, drug discovery, and DNA sequencing -Energy production, for price and load forecasting -Automotive, aerospace, and manufacturing, for predictive maintenance -Natural language processing, for voice recognition applications

The Development of Deep Learning A.I.

-Conceptualized in the 1950's but hasn't been a possibility until as early as 2010 -Requires massive amounts of processing power that haven't been available until recently -R&D has been extremely expensive until companies like Facebook began open sourcing their work, allowing smaller developers to build on their research

Important Elements of Machine Learning

-Data Preparation Capabilities -Algorithms -Automations -Iterative Processes -Scalability -Ensemble Modeling

Neural Network

-Different connections leads to different network structure -Each nuerons can have different values of weights and biases -Weights and biases are network parameters (use theta symbol)

AI MARKETING Downsides

-Difficult to use depending on company or industry being marketed -Can create marketing tactics that are predictable = too boring -Has the potential to stifle the creative process in marketing

Unsupervised Machine Learning

-Does NOT need algorithm input (does NOT need to be trained) -No Human Guidance -No reference data set and outcomes are unknown

Unsupervised Machine Learning

-Does not need inputs or need to be trained -Has led to "Deep Learning"

Industry Leaders in Deep Learning

-Google -DeepMind -Facebook

How Unsupervised Learning works?

-Group & interpret data based only on input data -Finds hidden patterns or intrinsic structures in data. -Applies cluster analysis, association rules and dimension reduction

Propensity model consumer metrics:

-Likelihood of conversion -Price Point of conversion -Which consumers to target to become repeat customers -Predict lead effectivity

2 Areas of AI

-Machine Learning -Deep Learning

Machine Learning vs. Deep Learning

-Machine learning: In the current state of machine learning, algorithms need to process millions of photos of an object before learning to recognize it. In other words, machine learning looks at one general pattern and tries to make predictions based off them. -Deep Learning uses neural networks which can recognize relationships between various parts of an object so that they can more easily and quickly recognize an object in a photo. In other words, machine learning looks at one general pattern and tries to make predictions based off them.

Benefits of Machine Learning for Marketing

-Massive data input from unlimited sources -Rapid processing, analysis, and predictions -Action Systems -Learning from past behaviors

Why you should care about AI?

-One of the largest projects moving forward is making AI's that can make other AI -This would let computers learn from their own mistakes, and improve themselves based on what they have found important -Very human thing to do "oh i have trouble walking to and from school, so I'll work out more" -The rapid advancement in the last 20 years puts AI on the cusp of the same type of explosive growth as personal computers and cell phones.

AI MARKETING: Data Collection

-Organisms amplifying group intelligence by forming flocks, schools, shoals, colonies, and swarms (eg. bees, birds, fish, ants) -Research shows → groups outperform most individuals in decision making & prediction regardless of expertise -Most surveys, focus groups, and polls fail to recognize this social aspect of decision making

Frank

-Pay-Per-Click tool -Uses Machine learning to find best paid advertising channels depending on specific audience

Applied (More common)

-Perform single task extremely well -Run automated and repetitive tasks -Does NOT involve decision making Ex) Bots, weather update, self driving cars

Why Machine Learning Matters?

-Speed to support faster compute calculations and decision making -Power to process and analyze large volumes of data. -Efficiency to generate more models -Intelligence through the ability to learn autonomously and uncover latent insights.

Supervised Machine Learning

-Supervised → Human guidance to train the algorithm -Data used already labeled with correct answers -Just needs to learn process to get from input to output

2 Categories: Machine Learning

-Supervised: Human guidance -Unsupervised: does not need input

General (Less common)

-System that can handle infinite number of tasks -Involved in decision making + analysis -Think and function similar to human brain -Still in development of artificial neural network

Albert

-Used for autonomous media buying -Buying digital media on behalf of clients -Analyzes, manages, and optimizes paid advertising campaigns

Deep Learning:

-based on algorithms to identify relations in data inspired by the process of the human brain -Sub-area of machine learning -Utilizes artificial neural networks → computer system modeled after human brain & nervous system -Allows ability to → adapt to changes to maximize results

Propensity Model:

A statistical algorithm used to predict customer behavior

AI Decisions

AI is an area of computer science that "emphasizes the creation of intelligent machines that work and react like humans"

AI Marketing: Tool Platforms

Albert and Frank

Sub-area of machine learning

Deep learning

Cat Example of machine learning

Deep learning is able to analyze several different patterns at once, (shape, color, etc.) and therefore will be able to quickly recognize and identify a specific cat after only a few different pictures

Google Deep Learning

Google's self-driving cars will utilize deep learning in order to help them react quickly to unexpected things on the road

Cat Example of machine learning

Machine learning analyzes one part of the cat- shape, color, facial features, or size, and takes millions of different photos to be able to accurately distinguish a specific cat from pictures of other cats

What are Capsule Networks?

Neural networks that recognize relationships between various parts of an object so that they can more easily and quickly recognize that object in different images

Market example of deep learning

One of the main applications of this can be seen from music streaming services such as Spotify when it curates recommended playlists and suggests songs based on previous listens

DeepMind Deep Learning

Revered as the foremost experts in A.I. deep learning, DeepMind has created self-learning A.I. capable of beating globally ranked players at chess, Go, and Shogi

UNU

SWARM's insight tool

Machine Learning

Systems that automatically learn and improve without being programmed to do so using large amounts of data and algorithms

Facebook deep learning

Through the Facebook A.I. Research initiative, (FAIR) Facebook has open sourced many of their A.I. breakthroughs, helping smaller companies begin their own work as well

Artificial Design Intelligence:

Use of AI for content creation including: -Design of websites -Written content and graphics -Writing Reports -Creating website dashboards -Email communications

Machine Learning

is a method of data analysis that automates analytical model building. It is a branch of artificial intelligence based on the idea that machines should be able to learn and adapt through experience.

If a company has __________ and ___________ the company could benefit

numerically based content, data-driven info

Customer behavior predictions are only as good as

the data you put into them

Neuron equation

z= a1 w1 + ... ak wk+ ... aK wK + b