CompSci Objectives

Describe and differentiate between the components of the Von Neumann Computer architecture

In a Von Neumann Computer it can store and process info.  There are 4 parts.  Input, CPU, Memory, and output.  The input device is what inputs the information.  The CPU processes that info based off algorithms (instructions) are stored in memory. The memory houses more data and algorithms as well.  The output device is the result of the computer after processing the input.  This may be pulling up a file or operating a system like a printer.

Describe the function of each major computer hardware component: motherboard, CPU, memory boards, display card, power supply

The motherboard is the physical piece of material that houses the major components of the computer and connects them.  The CPU or central processing unit processes information and carries out commands.  Memory boards such as RAM and hard drives store memory for the computer both long term and short term.  The display card is responsible for the graphics of the computer and displaying the information.  The power supply takes power from a wall outlet and sends it to all areas where it is needed.

Summarize the memory hierarchy and explain the balance between access speed, storage capacity, and cost with regards to computer memory

At the top of the hierarch is the CPU caches and registers (not important for this course).  Then Ram, Hard Drive, and Archival memory like DVDs and CDs.  As access speed increase, storage capacity decreases and cost increases.  So RAM cant store a ton of data, and is expensive but is fast.

Apply Moore's Law to analyze computer performance over time

Moore’s Law is the principle which says that processing power of computers roughly doubles every 2 years due to the number of transistors which can fit onto a CPU.  Also, as this increases the price decreases

Identify and describe the meaning of common file extensions.

Common file extensions are letters that follow a file to tell what operating system the file falls under.

• Text Files: .txt (plain text), .docx (Microsoft Word), .pdf (Portable Document Format)

• Image Files: .jpg/.jpeg (JPEG image), .png (Portable Network Graphics), .gif (Graphics Interchange Format)

• Audio Files: .mp3 (MP3 audio), .wav (Waveform Audio)

• Video Files: .mp4 (MPEG-4 Video), .avi (Audio Video Interleave)

• Executable Files: .exe (Windows executable), .app (Mac application)

• Compressed Files: .zip, .rar (compressed archives)

• Code Files: .py (Python script), .html (HTML document), .java (Java source code)

Navigate the file structure within the computer, including locating, opening, and saving files in different directories.

In the computer there is a root directory which is the topmost hierarchy, this is the operating system (Windows).  From this there are sub directories which house more specific files.  There are paths of saving information is each directory.  An absolute path goes all the way from the root directory to find a file, while a relative path only goes from the current directory the user is in.  Files can be saved by going to each directory and saving them there.

Convert between binary and decimal numbers

So for each place do the number times 2 to the power of the place. To go from decimal to binary divide by 2 and the remainder is the number until you get to zero then read in the reverse order.

Calculate how many values can be represented with b bits

For every number of bits, square that number.  So, if I have 5 bits, I can have 2^5 unique combination.

Calculate how many bits are needed to represent a given number of unique values

Find the power of 2 needed to encompass the number of representations needed.  So if I need 48 combos, I need 6 bits of info because 2^6 = 64 and 2^5= 32 so it must be 6 bits.

Convert between bits, bytes, kilobytes, megabytes, and gigabytes

1 byte is 8 bits.  A kilobyte is 2^10 which is 1024 bytes.  For further units such as mega or giga just multiply or divide by 1024.  So 1 megabyte is 1024^2, and 1 gigabyte is 1024^3.

Explain how characters are encoded in ASCII and UNICODE

ASCII uses 8 bits per character and UNICODE uses 8,16, or 32. Each character has a unique set of binary numbers.

Identify the key characteristics of an algorithm

It must be an ordered set of executable steps that are unambiguous and free of omissions and contradictions that will result in the accomplishment of an objective.

Or

A set of detailed instructions for accomplishing a task.

Explain how algorithmic reasoning is broadly applicable

Algorithmic reasoning is broadly applicable because it helps solve problems, optimize processing, and automate tasks.

How should algorithms be constructed to solve problems of varied complexity

Write down a logic design such though a recipe is being followed to outline completely what needs to be done to achieve an end result, then implement is.

Define Artificial Intelligence (AI)

Artificial Intelligence is the simulation of human intelligence by machines

Summarize how the Turing test has shaped AI

Turing test is the ability for an AI’s output to be indistinguishable from that of a human.  It is the measurement of being able to say if a computer is intelligent.

Differentiate between narrow and general AI

General AI can exhibit human cognitive abilities and can think and accomplish task on its own including logical thinking and performance in uncertain conditions and being able to learn on its own.

Narrow AI can model human intelligent behavior but can only operate within a pre-defined range of functions.  Can’t think for itself, can only learn when provided with data, and can only perform the task it was designed for.

Outline the types of problems AI can (and can potentially) help humans solve

Classification, Clustering, Optimization, Abnormality Detection, Estimation/Prediction, Ranking/Recommendation, and Data Generation.

Describe and differentiate between each phase in the "Sense-Think-Act" cycle

Sense is the detection of information about its surroundings.  Thinking is finding the best move to do in the current state.  Acting is performing that action.

Differentiate between AI decision making techniques such as decision trees, search trees, and finite state machines

Decision trees are a series of if statements that the AI goes through to decide what to do.  Search for trees though.  Finite state machines determine a state it should be on based on the surroundings, like a guard being passive until it sees the player.  Search tress are branches of decisions the AI can choose from and when they choose the best state that branches out.

Define the concept of a game state

A game state is a snapshot of a game at a given point and the configuration of all the pieces of that game at a time.

Define the concept of a heuristic for Artificial Intelligence

A heuristic is a simplified way of approximating how good a certain state is more quickly compared to an optimized algorithm that considers every outcome of a state. 

Differentiate between depth-first and breadth-first search

Analyzing a game state layer by layer for decisions where is looks at all the decisions for a certain level. Breadth-first search depends on how far we choose for the AI to pick.  Depth-first searches chooses one linage and goes as far down it can for one input then does that for all possible choices.  It checks for all possible outcomes.

Use rule-based AI strategies to solve a discrete game

This would be like if car in front, go left, or go right, or for pong, for x position of ball, go to x position of paddle.

Define "model", "instance", "feature", and "label" in the context of machine learning

A model is a representation of what a machine learning system has learned from the training data.  A feature is a value that can be used to make a prediction, a column in a data set.  A label is the answer or the result of an instance.  It represents the concept that the model is being trained to classify.  An instinct is a row in a dataset, a specific example.

Compare and contrast Supervised and Unsupervised learning

Unsupervised learning is good with unlabeled data and its job is to cluster data for a person to determine if it is useful. Supervised learning is with labeled data that it then used to identify relationships between and predict the label by only looking at the features.  The wine and beer example.

Describe and differentiate between the steps required for supervised learning

1.       Gathering data- determining what information needs to be collected

2.        Preparing data- putting the data in an order and such a manner that makes sense

3.       Choosing a model- Choosing what kind of algorithms work the best

4.       Training the model- Giving the data to the model and letting it learn

5.       Evaluating the model- Test the model against new model to see if the model has learned concepts, not memorized data

6.       Tuning- Tweaking the settings to increase the accuracy of the algorithm like learning rate or how much time the algorithm can go through the data

7.       Predicting: using the model to do work