Cloud and Emerging Technologies

By admin123
From Sensors to Intelligence
CLOUD FOUNDATIONS / AI
Study Deck · 1.16–1.18

From Sensors to Intelligence

How Artificial Intelligence, Machine Learning, and the Internet of Things depend on the cloud to collect, move, and make sense of data — and how each connects to the next.

Use → / ← or click the arrows below. Cards with a + can be opened.

Artificial Intelligence

Simulating human intelligence, at scale

AI works with structured, semi-structured, and unstructured data to solve complex problems, using a set of rules to manage its analysis. It's the broader science — Machine Learning is one component inside it.

Artificial Intelligence · Types

Six ways AI shows up in industry

Click a card to see what it does.

Artificial Intelligence · Generative AI

Generates content from a prompt

Text, code, or images — often powered by deep learning, and common across code generation, content research, content generation, and chatbots.

Training

Relies on large datasets of examples from related sources — books, databases, images — as source material for inferencing.

Inferencing

The operational, production stage — an already-trained model generates new content.

Artificial Intelligence · Large Language Models

Pre-trained knowledge, focused on language

LLMs are built on deep learning and specialize in text analysis, chatbots, and other human-like interaction.

  • OpenAI Generative Pre-trained Transformer (GPT) — foundation for ChatGPT
  • Microsoft Turing Natural Language Generation (NLG)
  • Google Pathways Language Model (PaLM)
Artificial Intelligence · Cloud Vendors

AI is a first-class cloud service now

Microsoft Azure

  • CycleCloudProvision and manage HPC solutions
  • Azure Machine LearningAI-based machine learning tool
  • Azure AI StudioGenerative AI development environment

AWS

  • Amazon QAI-powered business assistant
  • Amazon PollyText-to-speech service
  • Amazon RekognitionInformation from images/video

Google Cloud

  • Vertex AI StudioGenerative AI development environment
  • Vertex AI NotebooksToolset for data science workflows
  • DialogflowConversational AI environment
Machine Learning

A subset of AI that learns instead of being told

AI
ML
DL

ML is not explicitly programmed to find one outcome — it learns from data and makes decisions with minimal human interaction. Deep Learning is a more complex subset that emulates the brain's learning process for greater accuracy on unstructured data.

Machine Learning · Where it shows up
Machine Learning examples
  • Traffic prediction
  • Product recommendations
  • Medical diagnosis
  • Sports strategy
Deep Learning is found in
  • Gaming
  • Language processing
  • Agriculture
Machine Learning · Tools

Common frameworks in cloud environments

Internet of Things

Network connectivity meets smart devices

IoT combines connectivity and smart devices — software, sensors, robotics — that exchange data and instructions over the internet or internal networks. It's enabled by near-global connectivity, low-cost sensors, and cloud management platforms.

Smart homes
Medical monitoring
Agriculture management
Energy management
Manufacturing / Industrial
Internet of Things · The Cloud's Role

IoT generates. The cloud uses.

IoT devices generate large volumes of information; cloud services store and process it, consolidating data from geographically dispersed networks. Scaling and resource pooling give flexibility when working with that data.

Generates data
IoT Devices
→
Uses data
Cloud Services
→
Acts on data
Analysis & ML/AI
Internet of Things · Components

Three layers of communication

LOCAL
Communication between IoT devices, sensors, or other information-gathering components.
GATEWAY
Communication between a population of IoT devices and the internet.
CENTRAL SERVERS
Accept inbound communications to centralize and analyze data.

Sensors detect changes — position, temperature, chemicals — but communication is limited by device capability, power, range, media, and security requirements.

Internet of Things · Sensors

Collecting data across locations

Air quality
Accelerometer
Temperature
Biomedical
Water / humidity

Example: a smart irrigation system reads soil water levels and automatically adjusts settings — the optimal amount of water, without waste or dry crops.

Internet of Things · Transmission Stack

Organized on the OSI model, in three layers

APPLICATION
Connects applications that consume IoT data.
NETWORK
Provides connectivity and security.
PERCEPTION / SENSOR
Gathers information on an IoT device.
Internet of Things · Protocols

Trade-offs by protocol

ProtocolProfile
Wi-FiCommon wireless protocol; high power requirements make it tricky for many IoT deployments.
BluetoothShort point-to-point range, low power, limited security.
HTTPSynchronous — high processing and connectivity needs limit its fit for many IoT cases.
NB-IoTNewer wireless protocol; low power, limited transmission capability.
AMQPAsynchronous and reliable — queues messages to enable high availability.
Internet of Things · Gateways & Edge

Translating protocols, computing closer to the source

Gateways give IoT connectivity to the cloud, adding a security layer and translating protocols — not every device speaks IP. Gateways can also connect to edge devices, which preprocess data near the source rather than deep in the cloud.

01
Reduced bandwidth utilization
02
Reduced latency
03
Improved real-time analysis and decision-making
Summary

One data pipeline, three disciplines

AI

The broad science of simulating human intelligence across structured, semi-structured, and unstructured data.

ML → DL

ML is a subset of AI that learns from data; Deep Learning is a subset of ML for harder, unstructured problems.

IoT

Generates the real-world data — sensors, gateways, edge — that the cloud stores, and AI/ML ultimately learn from.

01 / 17
↑ arrow keys or click to navigate