Emerging Technologies – Glossary of Important Terms by Topic for

CEC04 (VIVA)
Difference between AI, ML, Deep Learning & Neural Network, NLP and Text Analytics. Also how are they related to
each other?
1. Artificial Intelligence (AI):
 Definition: AI is a broad field of computer science focused on creating systems capable of
performing tasks that require human-like intelligence, such as reasoning, problem-solving, decision-
making, and perception.
 Scope: Encompasses various subfields, including ML, NLP, robotics, and expert systems.
 Example: A self-driving car or a recommendation system.
2. Machine Learning (ML):
 Definition: ML is a subset of AI that involves algorithms that allow systems to learn and improve
from data without explicit programming.
 Types: Supervised learning, unsupervised learning, and reinforcement learning.
 Example: Predicting house prices based on historical data.
 Relationship to AI: ML is a technique used to achieve AI.
3. Deep Learning (DL):
 Definition: DL is a subset of ML that uses artificial neural networks (ANN) to process large amounts
of data and automatically extract features.
 Strengths: Particularly powerful for tasks involving unstructured data like images, audio, and text.
 Example: Image recognition or language translation using deep neural networks.
 Relationship to ML: DL is a specialized form of ML.
4. Neural Networks (NN):
 Definition: NN are algorithms inspired by the structure of the human brain, consisting of layers of
interconnected nodes (neurons). These form the foundation of DL.
 Types: Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), etc.
 Example: Detecting handwritten digits using a neural network.
 Relationship to DL: NNs are the building blocks of DL.
5. Natural Language Processing (NLP):
 Definition: NLP is a subfield of AI focused on enabling machines to understand, interpret, and
generate human language.
 Applications: Sentiment analysis, machine translation, chatbots, etc.
 Example: Virtual assistants like Siri or Google Assistant.
 Relationship to AI and ML: NLP uses AI principles and often employs ML/DL techniques.
6. Text Analytics:
 Definition: Text Analytics refers to extracting meaningful insights and patterns from textual data
using techniques like statistical analysis, ML, and NLP.
 Focus: Summarization, sentiment analysis, keyword extraction, etc.
 Example: Analyzing customer reviews to determine satisfaction levels.
 Relationship to NLP: Text Analytics often overlaps with NLP as it uses many NLP methods for
analysis.
How They Are Related:
 Hierarchy:
o AI is the umbrella field.
o ML is a subfield of AI.
o DL is a subfield of ML.
o NN are a key technique within DL.
 NLP and Text Analytics:
o Both are specific domains within AI that often leverage ML and DL.
 o NLP focuses on understanding and generating language, while Text Analytics emphasizes
deriving insights from text.
 Overlap: Techniques like DL are used in NLP for advanced language modeling, which in turn
supports tasks like Text Analytics.

BIG DATA
Quantitative Data/Qualitative Data
STRUCTURED DATA/ UNSTRUCTURED DATA
5V’s
Applications

CLOUD COMPUTING
Cloud Service Models
• Infrastructure as a Service (IaaS)
• Platform as a Service (PaaS)
• Software as a Service (SaaS)
Cloud Deployment Models
• Public Cloud
• Private Cloud
• Hybrid Cloud
Cloud Computing Advantages
• Cost: “Moving to cloud"
• Elasticity:
• Collaboration & Sharing:
• Technology:
• Infrastructures Management
Cloud Computing Challenges
• Control is relinquished to third party services.
• Unauthorized access.
• Data corruption.
• Data Security & Privacy
• De-perimeterization
Examples of cloud computing

AI
What Artificial Intelligence
AI Applications
1. Virtual Personal Assistants (Siri, Google Now, Alexa, Cortana)
2. Video Gaming
3. Smart Cars
4. Purchase Prediction
5. Fraud Detection
6. Healthcare
7. Finance
8. Marketing
9. Logistics
10. Manufacturing
11. Agriculture
 12. Energy

Machine Learning
Definition of Machine Learning
Machine Learning Project
1. Data Collection
2. Data Preparation
3. Choosing Learning Algorithm
4. Training the Model
5. Evaluating the Model
6. Predictive Analytics

Types of Machine Learning

• Supervised Learning - training dataset with predefined answers by use of labelled (data with
meaningful tags) datasets to train algorithms to classify data or predict outcomes accurately.
• Episodic (current action will not affect a future action).
• Regression and Classification (Logistic).
• Unsupervised Learning - model itself find hidden patterns and insights from the given data from
historical data.
• Clustering, Association rules (Mkt Basket Anay), and Dimensionality Reduction (Factor
Analysis).
• Reinforcement Learning - maximize cumulative rewards in a given situation; learning through
experience.
 Agent
 Environment:
 State
 Action
 Reward

Applications

Deep Learning
Definition of Deep Learning
The ‘Learning’ in ‘Deep Learning’.
Simple Neural Network components
 Input Layer
 Hidden Layer
 Output Layer

Types of Artificial Neural Networks

• Feedforward Neural Network (FNN) - information moves in only one direction
• Convolutional Neural Network (CNN) - patterns recognition enabled by hidden ‘convolutional
layers’
• Recurrent Neural Network (RNN) – for Sequential data is data such as words, sentences, time-
series data
• Long Short-Term Memory (LSTM) Network
• Generative Adversarial Network (GAN) - ‘adversarial’ deep-learning architecture with Generator
Network and Discriminator Network
Natural Language Processing (NLP) & Text Analytics
NLP - broad field encompassing the automated interpretation of human language.
Text analysis - process of extracting meaning from digital text
Text Analytics project.
1. Data Collection
2. Data Preparation
 Tokenization - transforms the text into a series of word tokens.
 Part of speech (POS) tagging - data is assigned a grammatical category like noun, verb, adjective,
and adverb.
 Out of Vocabulary – OOV removed (proper nouns, professional jargon, compound words,
abbreviations, and slang).
 Parsing - understanding the syntactical structure of the text.
 Lemmatization and stemming - reducing words down to their base form, known as the lemma.
 Removing Stop-words - ‘and’, ‘the’ and ‘a’, etc.
3. Text analytics - techniques and algorithms can now be used to derive insights. Prominent among
them are Text Classification, Text Extraction, Adjacent Words and Word frequency.
Applications

Generative AI
• Working - Large Language Models" (LLM not AI), based on the "Generative Pre-trained
Transformer" architecture (or GPT—as invented by Google).
• Responses are based on the information that the model was trained on, which includes a broad
range of publicly available data up until cut-off date of Oct 2023
Prompt Engg – CTO; Context (Who), Task (What), Output (How).
Examples
Limitation & Bias
 Confirmation Bias (3rd month Jan)
 Hallucinations (LLM; 6-3 = 3.14)
 Algorithmic Bias (Nurse –> Female)
CRAAP Test – Current, Relevance, Authority, Accurate, Purpose

AR, VR. MR
Extended Reality (XR)
 Augmented Reality (AR)
 Mixed Reality (MR)
 Virtual Reality (VR) – Devices, VRML (Virtual Reality Modeling Language)
Telepresence
Examples

IoT
What is?
IoT Working – Sensors collect & transmit, Cloud receives information, Communication between networks
or devices, Actuator trigger action based on IF-THEN-ELSE, Analysis
Characteristics of IoT
IoT Technologies
• Sensors and actuators (Temperature, Image, Gyro, Obstacle, RF/IR, Ultrasonic, Distance)
• Connectivity technologies: Wi-Fi, Bluetooth, cellular, etc
 • Cloud computing & Big data analytics
• Security and privacy technologies

Applications of IoT
• Consumers
• Organizational
• Infrastructural
• Military
• Home automation
• Elder care
• Internet of Medical Things (IoMT)
• Transportation
• Industrial & Manufacturing (IIoT)
• Agriculture
• Smart Cities
• Energy management

IoT Challenges
• Big Data Explosion
• Interoperability issues
• Standardization & universal protocol to combine several heterogeneous devices.
• Managing and fostering rapid innovation
• Cost and complexity
• Privacy & Security.
• Absence of legislation

Blockchain & Cryptocurrencies

• Characteristics of Blockchains
• Distributed database
• Peer-to-peer (P2P) Transmission
• Spread over nodes
• Consensus Mechanisms
• Transparency yet pseudonymity
• Authenticity, integrity and security with Signatures
• Immutable (Append-only, no edit/modify/delete)
• Real world use-cases of Blockchains
• Payment processing
• Digital IDs for Financial Services
• Digital Voting
• Real estate, land, and auto title transfers
• Medical recordkeeping
• Equity trading
• Smart Contracts & their Use Cases
• Limitations and vulnerabilities (depends on active users, scalability, security)
• What is Cryptocurrency?
• How Cryptocurrencies and Blockchain Work Together - Proof of Work (PoW), Proof of Stake (PoS)
• Challenges of Cryptocurrencies (Volatility and Speculation, Regulatory Uncertainty, Security
Concerns, Illicit Activities, Interoperability)
Application Programming Interface (API) & Web Services
• What is an API?
• Characteristics of APIs (middleware)
• Types of API by Use Cases (Open APIs, Internal APIs, Partner APIs)
• Types of API by Application
• Local API: provides services from Front-end/GUI to back-end DB
• Program API: allow software applications to communicate and interact with each other
• Web API (Web Services): is application / device communication via World Wide Web.
• Examples of APIs
• API Economy

Cyber Security
• Security Vulnerabilities
• Data Security
• Network Security
• Application Security
• Cloud Security
• Emerging Technologies and Security
• Mitigating Security Vulnerabilities
• Backup and Recovery (Cold Site, Warm Site, Hot Site)
• Identity and Access Management - IAM (Authentication, Authorization, Access Logging and
Auditing)
• Cryptography (Digital Signatures, Digital Certificates)
• Encryption and Decryption
• HTTPS and TLS/SSL: Secure Web Communication
• Cyber Forensics