CSCI 101 - Exam 2

Artificial Intelligence (AI)

The capability of a machine to imitate intelligent human behavior. NOT all artificial intelligence is machine learning.

Found in Lecture ML Part 1 - Machine Learning

Machine Learning

Giving computers the ability to learn without being explicitly programmed. All machine learning is artificial intelligence.

Found in Lecture ML Part 1 - Machine Learning

Deep Learning

Creating machine learning models using deep neural networks.

Found in Lecture ML Part 1 - Machine Learning

Supervised Learning

Teach the machine by providing examples (with labels). Provide examples to the machine and then let the machine learn a model that represents the examples. Then use the model to predict values for data that are unknown. Found in Lecture ML Part 1 - Machine Learning

Unsupervised Learning

Try to recognize patterns in the data without any examples (no labels). Provide features to the machine and then let the machine determine the patterns that exist in the data. Found in Lecture ML Part 1 - Machine Learning

Reinforcement Learning

Perform an action then learn based on a reward or punishment.

Reinforcement Learning considers definition of states, a set of possible actions that can be taken, and a reward score. For a given state AND the action the machine takes, it receives a positive or negative reward.

Found in Lecture ML Part 2 - Machine Learning

Exploration vs. Exploitation in Reinforcement Learning

The core idea in reinforcement learning is trying to balance between trying new things and using what you know. In training, a model should have a high exploration factor so it can learn. When deployed, a model should focus on applying the knowledge it learned.

Found in Lecture ML Part 2 - Machine Learning

Exploration factor

Parameter that determines whether you should try something new.

Found in Lecture ML Part 2 - Machine Learning

Exploitation factor

Parameter that determines whether you should use the knowledge you already have.

Found in Lecture ML Part 2 - Machine Learning

Features

The values that describe the data (e.g., x values).

Found in Lecture ML Part 1 - Machine Learning

Label

The value that you are trying to determine (e.g., y-values). Also called the target class.

Found in Lecture ML Part 1 - Machine Learning

Classification

A type of supervised learning. Target class has discrete values (e.g. dog, cat, mouse).

Found in Lecture ML Part 1 - Machine Learning

Regression

A type of supervised learning. Target class has continuous values. Found in Lecture ML Part 1 - Machine Learning

K-Means Clustering

One example of how to do unsupervised learning.

Cluster the dataset into k groups.

Start with k points (centroids)

Determine which cluster a point belongs (closest centroid)

Move centroid to the average of all points in that cluster

Repeat 2&3 until convergence

Found in Lecture ML Part 1 - Machine Learning

Centroid

Center of mass ("average" of the points).

Found in Lecture ML Part 1 - Machine Learning

Neural Networks

A neural network is a network of neurons (analogous to biological neurons in our brains). Neural Network forms the core of every Deep Learning algorithm. These networks are represented as systems of interconnected "neurons", which send messages to each other.

Found in Lecture ML Part 3 - Machine Learning

Neurons

Each neuron has the following: Multiple inputs with a weight for each input, an activation function, and an output

Found in Lecture ML Part 3 - Machine Learning

Value Iteration

Finding the expected eventual reward of being in each state.

Q-Learning

Starts with an unknown set of rewards and learns each value by doing.

Neuron Step-Activation "Algorithm"

Multiply each input by its weight

Sum up all the productsIf the sum is >= the threshold value, output 1

Else, output 0

Training a Neural Network

Start with random weights (or educated guess) and produce an incorrect result

Determine the error relative to correct results

Update the weights to account for the error

Repeat Steps 2 & 3 until the output converges to the correct result

Computer Network

A set of independent computer systems interconnected by communication links for the purpose of sharing information and resources.

Found in lecture Chapter 7 Part 1 - Networks

Node

Individual computer on a network. Range in size from smartphone, tablet, laptop, desktop, to massive supercomputers.

Found in lecture Chapter 7 Part 1 - Networks

Gigabit ethernet

The fastest networking technology for data communication today.

Found in lecture Chapter 7 Part 1 - Networks

Bandwidth

The rate at which information can be sent and received across a network

Found in lecture Chapter 7 Part 2 - Networks and Crash Course Video #28

Dial-up

The slowest networking technology for data communication.

Found in lecture Chapter 7 Part 1 - Networks

PAN

Personal Area Network (typically 30-50 feet for Bluetooth).

Collection of privately owned devices in close proximity to person. e.g. Bluetooth connection from phone to computer

Found in lecture Chapter 7 Part 1 - Networks

LAN

Local Area Network.

Connects devices in close geographic proximity

Found in lecture Chapter 7 Part 1 - Networks

WLAN

Wireless Local Area Network (typically 150-300 feet for WiFi)

Found in lecture Chapter 7 Part 1 - Networks

MAN

Metropolitan area network (typically a few blocks to a city).

Computer network that connects a geographic area or region larger than LAN but smaller than WAN

Found in lecture Chapter 7 Part 1 - Networks

WAN

Wide Area Network.

Computer network that spans large geographic distance.

Found in lecture Chapter 7 Part 1 - Networks

Bus

single shared communication line

There is a theoretical MAX on a bus length due to collision detection scheme.

Found in lecture Chapter 7 Part 2 - Networks

Repeater

a device that simply amplifies and forwards a signal

Found in lecture Chapter 7 Part 2 - Networks

Bridge

a "smarter" device that has knowledge about the nodes on each bus or LAN

Found in lecture Chapter 7 Part 2 - Networks

Switch

Multi-port device that uses packet switching to forward data to appropriate destination

Found in lecture Chapter 7 Part 2 - Networks

Packet

An information block with a fixed maximum size that is transmitted through the Internet as a single unit A packet has a MAX size a long message is broken into multiple packets to transmit

Found in lecture Chapter 7 Part 4 - Networks and Crash Course Video #29

Router

Like a bridge, but can transmit packets between two different types of networks. E.g. Router can send info from a WLAN to a wired Ethernet LAN

Found in lecture Chapter 7 Part 2 - Networks

ISP

Internet Service Provider Both a user's individual computer and a company's private network are connected to the rest of the world through an ISP.

Found in lecture Chapter 7 Part 2 - Networks

Internet

The Internet is not a single computer network; instead, it is a huge interconnected "network of networks" that includes nodes, LANs, MANs, WANs, bridges, routers, and multiple levels of ISPs.

Found in lecture Chapter 7 Part 2 - Networks

WWAN

Wireless Wide Area Network Found in lecture Chapter 7 Part 1 - Networks

LAN Topologies

Bus, Ring, Star Found in lecture Chapter 7 Part 2 - Networks

Ring

connects nodes in a circular fashion (not used much)

Found in lecture Chapter 7 Part 2 - Networks

Star

central node routes info directly to any other node

Found in lecture Chapter 7 Part 2 - Networks

Collision

Occurs when two messages sent at same time

Messages are lost

Found in lecture Chapter 7 Parts 2 and 3 - Networks

Tier 1 Networks

International ISPs (= Internet backbone)

These international ISPs connect with other tier-1 networks

Example: Level 3 Communications provides a Tier-1 network

Found in lecture Chapter 7 Part 2 - Networks

Tier 3 Networks

The last mile providers

Needed to deliver Internet to residential homes and businesses

Found in lecture Chapter 7 Part 2 - Networks

Tier 2 Networks

Connect tier 1 and tier 3

Sometimes tier 2 term used interchangeably with tier 3

Found in lecture Chapter 7 Part 2 - Networks

Internet Society

A nonprofit, nongovernmental, professional society composed of more than 100 worldwide organizations (e.g., foundations, governmental agencies, educational institutions, companies) in 180 countries united by the common goal of maintaining the viability and health of the Internet. For smooth functioning, it's imperative to have a protocol set that everyone agrees to, the Internet Society is responsible for this.

Found in lecture Chapter 7 Part 3 - Networks

TCP/IP

The "common language" of the Internet. Found in lecture Chapter 7 Part 3 - Networks

Internet Protocol Stack layers

TOP

5\. Application (Ex: HTTP, SMTP, FTP)

4\. Transport (Ex: TCP, UDP)

3\. Network (Ex: IP)

2b. Logical Link Control (Ex: PPP, Ethernet)

2a. Medium Access Control (Ex: Ethernet)

1\. Physical (Ex: Modem, DSL, Cable modem, WiFi- 4G)

BOTTOM

Found in lecture Chapter 7 Part 3 - Networks

Physical Layer

Goal: Create a "bit pipe" between two computers.

The physical layer does NOT provide an error free communication medium. Need error detection AND correction, which is why we need layer #2!

Found in lecture Chapter 7 Part 3 - Networks

Two main types of communication channels

Cable and broadcast

Found in lecture Chapter 7 Part 3 - Networks

Cable communication line examples

fiber-optic, twisted pair/copper

Found in lecture Chapter 7 Part 3 - Networks

Broadcast communication line examples

satellite, microwave, infrared, radio

Found in lecture Chapter 7 Part 3 - Networks

Data Link Layers (2 sub layers)

Goal: Create an error free "bit pipe"

2a. Medium Access Control Protocol

2b. Logical Link Control Protocol

Found in lecture Chapter 7 Part 3 - Networks

Medium Access Control Protocol

Goal: Provides access to the physical channel.

Determines who sends on a shared communication line when multiple nodes want to send messages at the same time.

Found in lecture Chapter 7 Part 3 - Networks

Logical Link Control Protocol

Goal: Detects and corrects errors.

Its job is to solve the error detection and correction problem (to ensure the packet traveling across the bit pipe arrives correctly).

Found in lecture Chapter 7 Part 3 - Networks

ARQ

Automatic Repeat reQuest

This algorithm is the basis for all Data Link Control protocols that are currently used.

How it works: Sender sends packet

Sender maintains copy of packet

Receiver sends ACK (if packet received correctly)

Sender discards saved packet (if ACK received)

Sender re-sends packet (if no ACK received)

Found in lecture Chapter 7 Part 3 and 4 - Networks

ACK

Acknowledgement

Found in lecture Chapter 7 Part 3 - Networks

ARQ: What happens if packet 2 is lost?

Sender will time out since no ACK has been received ... then re-send packet Found in lecture Chapter 7 Part 4 - Networks

ARQ: What happens if ACK 2 is lost?

Sender will time out, re-send packet (as, again, no ACK received)

ARQ: What happens if receiver gets packet 4 but never saw packet 3?

Receiver sends ACK-4 if packet #4 is received correctly

Eventually sender will see ACK for 3 was NOT received and will resend.

Found in lecture Chapter 7 Part 4 - Networks

Network Layer

Goal: Deliver a message from source to destination.

The network layer in the internet is called the Internet Protocol (IP).

IP is a BEST EFFORT protocol

NO GUARANTEE a given packet will arrive

NO GUARANTEE what order several packets from one msg will arrive.

Found in lecture Chapter 7 Part 4 and 5 - Networks

IP-address

On the internet, nodes identify each other using a 32-bit IP-address. IPv4 has 32 bits (four 8-bit numeric quantities, each in range 0-255 such as http://74.125.224.72/)

Found in lecture Chapter 7 Part 4 - Networks and zyBooks Section 7.2 and Internet History reading

Host names

Host names are WAY easier than IP addresses to use for us humans Ex: google.com

Analogy: we prefer Assembly instead of machine language

Found in lecture Chapter 7 Part 4 - Networks

Port Numbers

Connected to different applications

Transport Layer gets packet to port

Well-known port numbers exist on the Internet e.g. DNS always uses port 53 for its communications

Found in lecture Chapter 7 Part 5 - Networks

DNS

Domain Name System Converts from a symbolic host name to its equivalent IP Address.

Found in lecture Chapter 7 Part 4 - Networks and zyBooks Section 7.2 and 7.4

TCP

Transmission Control Protocol A reliable transport protocol. Creates an error-free delivery service. TCP will also order the packets correctly for the message. TCP uses the same ARQ algorithm as DLL.

Found in lecture Chapter 7 Part 5 - Networks

UDP

User Datagram Protocol

An unreliable transport protocol

Used for quick/fast communication, e.g. streaming audio or video.

Found in lecture Chapter 7 Part 5 - Networks

Why do we need ARQ at both the data link layer AND the transport layer?

Need ARQ at data link layer as physical medium has issues. No sense passing packet to network layer if there are issues in it Also, network layer is BEST effort protocol - Packet could be lost in this layer TCP needs to ensure all packets all arrive

Found in lecture Chapter 7 Part 5 - Networks

DLL fixes \____ issues

physical layer issues

TCP fixes \____ issues

network layer

Application Layer

Goal: Provide rules to offer end-user services

TCP will get messages to the application.

The application then needs to have rules on what to do with those messages.

Found in lecture Chapter 7 Part 5 - Networks

Client/Server Model

In most network protocols, source acts as a client and destination acts as a server HTTP client requests a web page, server responds! IP is connectionless, so packets may NOT follow the same path!

Found in lecture Chapter 7 Part 5 - Networks

URL

Uniform Resource Locator A symbolic string that identifies a webpage.

Found in zyBooks Section 7.8

CSS

Cascading Style Sheets Helps to stylize a webpage with color, sizes, fonts and more

Found in lecture Chapter 7 Part 5 - Networks

Transport Layer

Goal: Create a program-to-program delivery service Get the packet to the program on the computer that is expecting it

Found in lecture Chapter 7 Part 5 - Networks

Top-level Domain

.edu in cs.mines.edu

Found in lecture Chapter 7 Part 3 - Networks and zyBook Section 7.8

Web Crawler

Pre-visits web pages to build an index of words

Found in lecture Chapter 7 Part 3 - Networks and zyBook Section 7.6

Hyperlink

A way to move from one page on the web to another.

Found in lecture Chapter 7 Part 5 - Networks and Crash Course Video #30

Web browser

A program for reading and interpreting websites.

Found in lecture Chapter 7 Part 5 - Networks and Crash Course Video #30

How did early search engines work?

Search results were based on number of times the word appeared on the page.

Found in lecture Chapter 7 Part 5 - Networks and Crash Course Video #30

Augmented Reality

Virtual objects overlaid on a view of the real world

Virtual Reality

The real word is entirely replaced with a virtual environment

Optical See-through

The real-world view is seen directly through a lens or smart glass where virtual objects are rendered

Video See-through

The real world is seen indirectly through a video camera, virtual objects are rendered on the video screen

Image Tags

Calculates 3D distance and orientation based on skewed views of a known image

Inertial Measurement Unit (IMU)

Electronic sensor that can measure 3D movements based on gravity and inertia

Robotics Disciplines

Robotics involves Mechanical Engineering, Electrical Engineering, and Computer Science

Robot

A system that can:

SENSE things about itself or its environment

PLAN what to do based on its goals and sensors

ACT on its goals and affects the environment in some way

Major Robot Subsystems

Power: Making sure all components have enough electrical power

Perception: What a robot can sense and understand about itself and its environment

Decision Making: The step between Perception and Manipulation (the majority of coding goes here!) Deciding what our goal is and how to achieve it

Manipulation: How a robot can move within and interact with objects in the environment

Sensor

perception of the external world and the robot’s internal state

Ex: Cameras, buttons, remotes, distance sensors, accelerometers, GPS

Controller

decision making component, reads sensor’s data and determines actuator’s responses

Ex: microcontrollers, embedded computers

Actuator

Actuators: the manipulation of the robot and its surroundings

Ex: movement, arms, tools, motors, servos, hydraulics

The Three D's

Robots are most commonly for situations or tasks that are:

DULL: Repetitive, predictable, or simple tasks like assembly lines

DIRTY: Sanitation, sewer repair. Used UV robot to sterilize masks for hospital workers!

DANGEROUS: Outer space, underwater, de-mining battlefields, navigating resource mines

Autonomy

The ability of a robot or machine to act without human control or intervention

Autonomy Levels

Level 0 - No Autonomy (Full human control, normal machinery)

Level 1 - Robotic Assistance (Robot assistance, helps with safety limits)

Level 2 - Task Autonomy (Small portions of functionality are automated)

Level 3 - Conditional Autonomy (Decisions and movements are approved by supervisor)

Level 4 - High Autonomy (Mostly automatic, supervisor for emergency interventions)

Level 5 - Full Autonomy (Robot acts fully alone throughout entire task or program)

Digital footprint

information about a person on the Internet as a result of their online activity

Found in Lecture 8 - Privacy

Metadata

a set of data that describes and gives information about other data

Found in zyBooks Section 8.1

strategies to protect user information

choose apps used carefully (How much does the app collect and record information? How does the app use the collected data?)

think before posting (Could the information you are posting harm yourself if made public?)

get the facts (Before signing up for a "free" service, consider how the service makes money)

check and set preferences carefully (If you don't check/set user preferences, default may be to "share")

Found in Lecture 8 - Privacy