Exam Questions

Question 1

What is the difference between Creative Computing and Computational Creativity?

Answer 1

Creative computing is as it sounds, to compute in a creative way, using computing as a tool to be creative. Computational creativity is using computing to create creativity itself. It aims to generate machine creativity through imitating human creativity. For example AARON, a project which a robot creates art and the user can learn, observe, and be inspired.

Question 2

What is Generative Art?

Answer 2

”Generative art refers to any practice where the artists uses a system, such as of natural language rules, a computer program, a machine, or other procedural invention, which is set into motion with some degree of autonomy contributing to or resulting in a completed work art.”
- Generative art is more focused on how, than on what

Question 3

What are some key ideas of New Media Art?

Answer 3

Materiality: artist and audience engaged with the physical world
Embodiment: being in the work in which the mind and the body are connected.
The Cyborg: The artistic piece should exhibit a certain degree of autonomy
Hybridity: Hybridity of disciplines and of the real world with the virtual world
Narrative: There should be an evolution in time (includes interactivity)
Interactivity: The viewer has the power to be an active participant in the infolding of a works flow of events, influencing and modifying its forms and narratives.

Question 4

What is Evolutionary Art and Interactive Evolutionary Art?

Answer 4

In Evolutionary art the artist dos not do the work of constructing the artwork, but rather lets a system do the construction.

In Interactive Evolutionary Art the selection is made by the artist itself.

They are both typically based on evolutive systems like Genetic Algorithms.

Question 5

What are the different types of “knowledge combination”

Answer 5

• Multidisciplinary: the knowledge in multidisciplinary
  approaches may collaborate, but they maintain a
  separation of their disciplines
• Interdisciplinary: It could fuse different knowledge deeply and finally end up with the extension knowledge
  of existing disciplines – boundary-crossing cooperation
• Trans-disciplinary: focuses on all the disciplines in the
  world, which makes it a more holistic approach than
  interdisciplinary combination. It tends to merge the
  disciplines (es. biotechnology)

Question 6

What are the two divisions of code in Processing?

Answer 6

The setup section which is run at the beginning and defines initial environment properties. (Screen Size, background color, images, fonts)

The draw section which is continuously looped/updated if not told to stop. This is to execute draw commands (Line, Point, Eclipse, Image). The speed of the draw loop is controlled by frameRate().

Question 7

What are some Audio Controls of Processing?

Answer 7

AudioIn
Sampling
Effects: LP, HP, BP, Delay, Reverb
Noise: White, Pink, Brown
Oscillators: Sin, Saw, Sir, Try, Pulse
Analysis: Amplitude, FFT

Question 8

What is Wekinator, it’s inputs and outputs, and uses?

Answer 8

Wekinator is an open source software created for interactive real time machine learning.

It’s inputs and outputs are OSC messages.

It can create a mapping between gestures, sounds animations, etc. Examples are gestures from webcam, Kinect, Arduino, to visual environments such as Processing, OpenFrameworks, or Quartz Composer.

It can create real-time music information retrieval and audio analysis such as: Instrument detection, genre, pitch, rhythm.

Inputs could be sent from any real-time input, such as game controllers, webcam input, motion tracking, audio input, sensors connected to Arduino, etc.

Outputs could be sent to any real-time process, such as music (Max/MSP, ChucK, PD, SuperCollider, Ableton/Max4Live…), animation (Processing, OpenFrameworks, …), games
(Unity3D), robots or physical actuators (e.g. using Arduino), etc.

Question 9

What is Dynamic Time Warping?

Answer 9

DTW is a measure of the similarity between two temporal sequences, which may vary in speed. Gesture Recognition for example, used for training and using a data set.

Question 10

What is an Agent?

Answer 10

An agent is anything that can be viewed as perceiving its ENVIRONMENT through sensors and acting upon that environment through ACTUATORS.

Agents behavior is described by the AGENT FUNCTION that maps any given percept sequence to an action.

Agents can sometimes have feedback (Robot interacting with environment) or be completely dependent on feedback. Interactive/Evolutionary agents).

Agents have perception of their environment (sensors, sounds, images..) then an agent function, and then an action takes place.

Question 11

What kinds of knowledge can Agents have?

Answer 11

ENCODED: the knowledge is encoded in the system by the creators (rule-based systems). In this case no knowledge is needed, it’s an action, reaction system.
INPUT KNOWLEDGE: the agent is exposed to some input even if part of the knowledge is already in the system. AARON for example, or a Jazz continuator)
KNOWLEDGE LEARNED: extracted by data (machine learning).

Question 12

What kind of behaviors can creative Agents have?

Answer 12

Fixed
Periodic - Similar to fixed but usually a simple input -> output
Complex - Chaotic but with some kind of structure/rules.
Chaotic - You can’t imagine the output, it’s different every time, there’s no rules.

Question 13

Provide an exhaustive description of the three agent paradigms studied during the course.

Answer 13

The three overall agent architectures that we studies were
COGNITIVE - Cognitive agents maintain a internal symbolic representation of the world. The example we used was a robot which has a camera and keeps a map (maybe even in 3D) of the environment.
REACTIVE - Reactive agents have no explicit representation of the world and focus on behavioral rules. Reactive agents can be with or without memory. With no memory Reactive agents are just REFLEXIVE agents, condition-action rule based with no representation of the world, actions based only on the current perception. With memory, Reactive agents can store information (Position, History of actions, …).
HYBRID - Hybrid agents are a mix of reactive and cognitive, to balance reactiveness and deliberativeness.

Question 14

What is Interactive art and the New Musical Instrument paradigm ?

Answer 14

Interactive Art invites the audience to take part and is focused on the user experience. The engagement curve should be quick and then dwindle.
New Musical Instrument Design is creating new instruments that are innovative and expand artistic ability. The learning curve should be not for the use of the system but what to use it for. Users should go back to it without being annoyed.

Question 15

Define the three different models of interaction (interaction design).

Answer 15

Didactic Models: The system is running constantly and the user doesn’t regulate the output but learns from it. An example of this is generative art.
Semi-Conversational Model: This type of interaction means the user communicates something to a system and then the system responds with something back. Full-Fledged Conversation Model: User and System fully and constantly interact. An example of this is a user and their cellphone.

Question 16

What is Computer Vision?

Answer 16

Computer Vision is defined as a field of study that seeks to develop techniques to help computers “see” and understand the content of digital images such as photographs and videos

Question 17

What are the types of interaction when it comes to Muti-Agents?

Answer 17

• No Interaction: agents do not interact, they do not have the knowledge of the existence of the others. An example of this was the energy of wind used to move particles. In this case they have no interaction, they only move due to an external force.
• Person goal: Each agent has a goal, it has the knowledge of the existence of the others and interacts with other ignorer to accomplish it’s goal. Examples of this can be to Avoid collision, Attraction, Connection.
  Collective goal: The complex systems of agents has a goal and single agents give their personal contribution to reach the goal. (In this scenario agents may communicate).

Question 18

What is the steering force model to design a particle system? Particles attract to one another but not allowed to collide with one another.

Answer 18

The seek problem: An agent looks for a way to reach a target. In this case each particle is considered an agent and they are attracted together like a swarm. Each particle can be characterized by its Location, Velocity, and Acceleration, and Life Span. Each should look how it needs to move to reach its target. Steering force = desired velocity - current velocity. To attract the particles should attract to one another, but to avoid collision once they reach a certain parameter of one another they should repulse.

Question 19

What is Boids Theory?

Answer 19

Boids (Bird-oid) is a theory developed to prove that flocks and swarms are self-organizing.

The collective behavior of the group is emergent because the rules concerning the parts of the swarm do not contain any notion of the whole.

Swarms assume that individual particles have a finite range of perception in which a given individual feels the influence of neighbors. (Repel at close range, attract at medium range, and are oblivious to each other at long range.)

The basic rules governing the interactions between neighboring particles in bonds are Separation, Alignment, Cohesion.

Boids is an example of a multi-agent because each particle (boid) is considered an agent.

Question 20

What is Stigmergy

Answer 20

Stigmergy (/ˈstɪɡmərdʒi/ STIG-mər-jee) is a mechanism of indirect coordination, through the environment, between agents or actions.

Question 21

What is Cellular Automata?

Answer 21

A cellular automaton consists of a regular grid of cells, each in one of a finite number of STATES such as on and off.

The grid can be in any finite number of dimensions. The grid is called a UNIVERSE.

For each cell, a set of cells called its NEIGHBORHOOD is defined relative to the specified cell.

An initial state (time t = 0) is selected by assigning a state for each cell.

TRANSITION RULE:
A new generation is created (advancing t by 1), according to some fixed rule (generally, a mathematical function that determines the new state of each cell in terms of the current state of the cell and the states of the cells in its neighborhood.)

Typically, the rule for updating the state of cells is the same for each cell and does not change over time, and is applied to the whole grid simultaneously, though exceptions are known, such as the stochastic cellular automaton and asynchronous cellular automaton.

“Game of Life” is a set of rules according to the status of the neighborhood. Using the Moore behavior (also diagonal neighbors) If a cell is alive it will die if it becomes overpopulated (The cell has three or more neighbors), or if it becomes lonely (The cell has only one alive neighbor). A cell is birthed when it has exactly 3 alive neighbors. Otherwise the state of the cell does not change. Alive + 2 Alive neighbors -> Stays Alive.

Question 22

Describe the techniques of modeling the creativity through the random process.

Answer 22

The most common way to model creativity is through the random process. (Introducing uncertainty).

We select something which consists of parameters which have a certain range of values, (music note, color, movement,…) and we randomly compute the value.

Given a variable X defined over a discrete values space (the domain D), a random draw is giving equal probability to all the option from D.

There can be different probability distributions of X. P(X).

This is considered a “straight-forward probability based walk”.

Question 23

What is the Levy flight Method and what is an application in music composition?

Answer 23

The Lévy flight method or “Monte Carlo” method is a specific random walk in which you pick two random numbers.

We pick a number R1 and then compute a probability P that R1 should qualify. Then we pick another random number R2 and if R2 is less than P(R1) then we take R1, otherwise if R2 is not less than P we start over the procedure.

In the application of music composition, the probabilities of each note can be learnt from an already existing composition, and then used in our Levy Flight random walk to compose music. One note will be randomly chosen, the probability of it appearing will be taken, then another note and it’s probability and if the probability of the second note is lower than the first then the first note will be used, otherwise we start over.

In computer graphics the random walk can be applied in direction but also distance.

Question 24

Provide a description of Grammars and of Markov Models, in particular focus of the how the Markov Models can be view as an evolution of grammars.

Answer 24

A grammar is a set of rules that define the structure of a language or a set of strings. The rules specify the allowed sequences of symbols, such as letters, words, or phrases, and can be used to generate or analyze strings in the language.

?

Question 25

What is Impro-Visor?

Answer 25

Impro-Visor is a grammar-based software for automatic Jazz solo melody creation.

The grammar was built manually based on jazz music theory.

The terminal alphabet is {h, q} (half notes/quart notes), then types of tones.
The non-terminal alphabet is {S, M, H} (sequence/measure/half measure).

Question 26

What are L-Systems? Stochastic L-Systems?

Answer 26

L-systems are a grammar-based system to model the growth patterns of plants. In L-systems there is no difference between terminal and non-terminal symbols.

L-System is defined as L = (v,w,P) where v = alphabet of infinite symbols, w = seed/indicator, P = finite set of rules.

L-Systems are great for generating fractals.

w is the initial shape (could be a line, could be a shape), the set of rules P tells you how to use that shape.

L-Systems can be used to create a binary tree, or even a fractal plant.

Stochastic L-Systems introduce a probability to each rule. This creates a like of “uniform” randomness. (patterns but with an introduced randomness).

Question 27

What are tranisitional Networks?

Question 28

What are Linked Automata? Applied to music?

Question 29

Markov Model? Markov Chain?

Answer 29

Markov models, are a class of models used to generate sequences of symbols or states. They are based on the idea that the probability of a symbol or state at a given time step depends only on the previous symbol or state, and not on any earlier history. This is known as the Markov property.

A Markov chain is a discrete-time random process with the Markov property * A discrete-‐time random process means a system which is in a certain state at each step,
with the state changing randomly between steps * The changes of the state are transition with a given transition probability

In general, Markov models can be seen as an evolution of grammars, as they use probability to describe how words or symbols follow each other instead of using a set of fixed rules. A Markov chain can be represented by a transitional network. Markov models can be trained on large amounts of data, such as texts written in a specific language, and then used to generate new sentences or texts that follow the same transition probability.

The simplest Markov chain uses the probability of one state give the previous one. But even with music for example, it’s often that the next note is not based off the previous one but a string of previous notes. Including a larger string into our probability creates a higher order Markov chain, and with reason a more accurate model.

However, a problem still remains -> capture the inner structure of a piece that, generally, has
not a fixed length * Variable Length Markov Chains attempt to overcome the issue