Interested in data mining and computational creativity? We hire researchers and students and supervise thesis work. On this page, we list some possible topics for student projects/theses in our research group.

Interactive poetry writing

We are currently working on the use of constraint programming for generation of poetry. In this project, the goal is to develop interactive tools where the user and computer write poetry collaboratively. First, the user provides a topic and the computer provides a first draft of a poem on that topic. After that, the user and the computer take turns in editing the poem until the user (and the computer?) is satisfied. A poetry writing computer program exists, and the main task of the student is to turn it to an interactive tool. An ideal candidate would have good programming skills and some combination of studies or interests in linguistics, literature, and constraint programming.

Kinetic typography for poetry

Poetry is typically printed on paper. Isn't that a bit old fashioned these days? Think of the potential of graphical, animated and even interactive visualizations for making poems richer experiences. This project applies (automated) kinetic typography, i.e., animation of letters and words, on poetry. In combination with the topic above, an ambitious goal would be an interactive, potentially game-like setting where the user and computer collaborate to produce poetry by moving words around and modifying/deleting/adding them.

Interactive story teller

Once upon a time, in a faraway kingdom university lived a beautiful princess computer science student and her widowed father windowed computer (or linuxed, just as well). During a hot Finnish summer (in stories, anything is possible), she programmed the computer to understand the structure of stories and to support kids in composing their own fairy tales. And they lived happily everafter. She earned first an MSc degree and then a PhD, the computer was retired and replaced by a young, slim tablet, and the kids made the world a bit better place when they grew up.

Conversational trolling agent

The Turing test is difficult to pass, but what about what we call the trolling test: how well can a computer program participate in on-line discussions and pass as a real human user? Some statistical natural language text analysis is needed to first figure out what is being discussed, and then the program should produce and post comments that do not stand out as computer generated or as being out of topic. Can the computer program even evoke constructive discussion?

Creative WordBench

There is an increasing interest on developing tools supporting humans in linguistic creativity. The goal of this project is to integrate available computational resources (textual corpora, computational linguistic tools) in order to develop an authoring tool for building short creative texts such as advertising headlines or funny titles. The tool is conceived as a web application allowing users to assign tasks (e.g. "find a list of movie titles containing a word in the input topic and rhyming a given word") performed as parallel threads that are visualized in real time. The GUI will show the best outputs available at the current time, emphasized in a visual way with colours and kinetic typography. As a prerequisite, the candidate is required to have good programming skills in Python and Javascript (or HTML5 and WebSockets).

Musical sonificator

Sonification and musicalization are two ways to perceptualize data. Sonification includes a wide range of auditory display, not necessarily of musical type (e.g. Geiger counter is a way to achieve sonification for counting radiation levels). Musicalization, in turn, is employed to represent information as a combination of melody, harmony, and timbre (see sleepmusicalization.net). In this project we want to focus on the timbre and explore the boundary between noise and sounds. To what extent could we rearrange the noise obtained through the low-level sonification of input data, and manipulate its tone quality in order to use it as a musical effect? For this project good we require good programming skills and, possibly, some experience in data analysis and computer music.

Mobile musicalization

This project is a variant of the musical sonificator above and explores the use of a mobile phone as an intelligent musical instrument: read data from acceleration sensors, camera, and microphone, and immediately produce live music. Again, programming skills are needed, also on mobile platforms, and knowledge of (computer) music is a definite plus.

Graph browser

Graphs and networks are abundant, e.g., in biology, social networks, citation networks, traffic -- and in computational creativity as simple representations of  world knowledge. However, they often are much too large to be visualized. This project develops methods and tools for interactive visualization and browsing of large networks and their contents. The work combines graphics with graph mining and user interaction, and a successful candidate should have interest in this combination. Prof. Giulio Jacucci will collaborate with us and will supervise the user interaction part.