Exam Notes

Skills measured as of April 24, 2024

Describe Artificial Intelligence workloads and considerations (15–20%)
Describe fundamental principles of machine learning on Azure (20–25%)
Describe features of computer vision workloads on Azure (15–20%)
Describe features of Natural Language Processing (NLP) workloads on Azure (15–20%)
Describe features of generative AI workloads on Azure (15–20%)

Aritificial Intelligence (AI) : The theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.

Narrow AI : Computer systems which use human intelligence but have very strong limitations in what they can do. Ex . Siri, Cortana & Google Assistant are examples of Narrow AI

Machine Learning

The study of computer algorithms that improve automatically through experience.

→ Unsupervised Learning : ability to find patterns in data without human help → Supervised Learning : humans label the data and gives general guidance

Natural Language Processing

Allows a machine to read and understand human language, Machine translation, question answering, sentiment analysis, etc

Perception

The ability to use input from sensors - images, audio, lidar, sonar, radar, touch etc. Covers things like facial recognition, speech recognition and object recognition.

ML Model

ML allows computers to use data to forecast the future without specifically being programmed.

In ML, a model is a program that can be used to recognize a pattern in data

A Model can be used to recognize people, objects and landmarks using unseen images
A Model can be used to understand the context of natural human text or speech
We train a model using "training data"
We evaluate a model using "test data" to measure how accurate is it
Once model has been deployed, it can recognize patterns in data it has never seen before

Common AI Workload :

Prediction and Demand Forecasting
Anomoly Detection
Computer Vission
Natural Language Processing
Knowledge Mining - Cognitive Search : Ingest content from Blob storage, Tables, SQL Database, and Cosmos DB, PDFs, Images, Word Documents, Powerpoints, Audio uses AI tools such as images classification, face recognition, language detection, key phrase extraction to create a searchable index.
Content Moderation : Find profanity, adult, hateful, offensive, violent or phishing links in text, video and images, we can set the severity levels to allow some free/open discussion.
Generative AI

Guding Principals :

Unintended Consequences :

Decisions that are wrong,
Decisions that are illegal,
Decisions that cannot be explained by anybody
Decisions that are harmful to society at large

Six Principles Should Guide AI Devlopment

Fairness
Reliability and Safety
Privacy and Security
Inclusiveness
Transparency
Accountability

Machine Learning Types

Supervised Learning

Regression : The ability to predict the outcome variable given 1 or more predictor variables.
Classification : Assign a score to the odds of it belonging to a cluster.

binary classification,
Multi-class classifications.

Unsupervised Learning

Clustering : Find groups of related things among data

Deep Learning

→ Deep Learning is a subset of machine learning where neural networks with many layers from a large amount of data. → This is how a ML algorithm can perform complex tasks sucha as image recognition, speech processing or language translation. → It does so using a hierarchy of neural networks, each iteration getting more complex and understanding the problem at a deeper level.

→ Feature is an input variable. → Label is the thing we're predicting

Feature and Labels in a Dataset Given a pile of data, data scientist need to determine which bits are relevant to make decisions on

Experiment
Domain Knowledge
Keep in mind the principles of AI

Training the Model

The more input(historical) data you have, the more accurate the results.
Don't use ALL of your data to train the model
Divide your available labelled data into training and validation/test datasets ie. 1,00,000 rows of data - 500K to train the model, 500K to test the model

Evaluate the Results - Regression

Use the validation dataset to test the model, and measure how close or far the actual results are from predicted results

Mean Square Error Large differences are much worse than small differences

Evaluate the Results - Classification

The result is to given a prediction score that the subject is part of the group. 70% confident this is an apple, 30% confident this is a pear. So if an apple is mis-identifed as a pear, that's ok as long as it only happens 30% of the time

False Positives vs False Negatives

Compare true positives with false positives and true negatives with false negatives when evaluating the model. How important is it to you that it never has a false positive? Accuracy vs Precision

True Positive Rate(TPR) : also called hit rate, sensitivity, or recall False Positive Rate(FPR) : also called fallout rate.

TPR and FPR together can be combined onto a curve. = Receiver Operation Characteristic (ROC) Curve

AUC : Area Under the Curve

Azure Machine Learning

Azure Machine Learning is a cloud-based service that helps simplify some of the tasks and reduce the time it takes to prepare data, train a model, and deploy a predictive service.

Azure ML Designer : Another no code approach to model design.

Computer Vision workloads

Image Classification
Object Detection
Semantic Segmentation
Optical Character Recognition
Facial Detection and Recognition

In Azure, we can use Computer Vision service or we can use Cognitive Services

1. Computer Vision Service

Pre-trained ML model
Can recognize over 10,000 objects
Can generate automatic captions for images and tags
Content moderation for adult, racy or gory content
Detect faces
Text recognition

2. Custom Vision Service

A model that you can build and train
Classification or object detection
Upload an existing data set of images and classed
Publish the model so that you and others can use it

What can I do with Azure AI Video Indexer ?

Deep Search
Content creation
Accessibility
Monetization
Content Moderation
Recommendations

NLP Workloads

Key Phrase Extraction
Entity Recognition
Sentiment Analysis
Language Modeling
Speech Recognition and Synthesis
Translation

NLP Services in Azure

1. Text Analytics Service

All of the Azure services that include key phrase extraction, entity detection, and sentiment analysis.
It can detect the language of text.
Can detect when text has multiple language and identify the predominant language.
Returns NaN when it can't determine.

2. Language Understanding Service ( LUIS)

Understanding Natural Language

Three core concepts :

Utterances : Something user will say
Entities : An item to which utterance refers.
Intents : The purpose or goal expressed by the user.

3. Speech Service

4. Translator Text Service

Generative AI

We’ve now firmly entered the era where computer ML algorithms can create text, images, videos and audio.

Image Generation : Art Creation, Face Generation
Text Generation : Chatbots and conversational agents, Content creation, Code generation.
Music or Audio Generation : Music composition, Voice synthesis, Sound effects
Video Generation : Deep fakes, Animation, Improving quality of video footage

Responsible Generative AI

There are clearly dangers of generative AI:

The ability of crete “fake news”, videos and images that are fake but trick the viewer into believing they are real.
Impersonating someone’s likeness or voice
Generating content that encourages harm or assists with illegal activities
Perpetuating stereotypes or discriminatory behavior

Responsible Generative AI Process

Identify potential harms that are relevant to your planned solution.
Measure the presence of these harms in the outputs generated by your solution.
Mitigate the harms at multiple layers in your solution to minimize their presence and impact, and ensure transparent communication about potential risks to users.
Operate the solution responsible by defining and following a deployment and operational readiness plan.

Azure OpenAI Service

Natural Language Generation (NLG) : NLG is a subfield of AI focused on creating text that mimics human language. It enables AI models to generate coherent and contextually appropritate responses based on input data or prompts.

Azure OpenAI Service provides access to advanced AI models like GPT(Generative Pre-trained Transformer), optimized for various tasks, including NLG.

These models can perform tasks such as drafting text, summarizing content, translating languages, or generating creative outputs like stories or poems.

Capabilites of Azure OpenAI for NLG

Contentd Generation
Text Summarization
Conversational Agents
Code Generation
Translation
Creative Writing

Microsoft AI-900