Prosjekt 2024

[ Vis hele beskrivelsen ]

To be creative, we need to produce something which is new, meaningful and has some sort of value. Computers are able to support humans in creative processes, but to also themselves be creative or to assess if an idea or a product is creative. A computational creativity project can investigate any creative field matching the interests and backgrounds of the student or students (language, design, music, art, mathematics, computer programming, etc.), and concentrate on one or several aspects of computational creativity, such as the production, understanding or evaluation of creativity, or on computer systems that support human creativity.

In particular, a thesis in computational creativity can investigate the transitions between different creative artforms, e.g., generating music or images based on textual input (as in Stable Diffusion models), generating music based on images or text, or generating text based on music or images.

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

Creativity can be found in nature and in humans, but also in computers, and entails to produce something which is new. However, just “newness” isn’t a sufficient condition for us to consider an idea to be creative, it also has to have some value and meaning: If a 2 year old draws some lines on a paper, we rarely consider it to be art; while if a grown-up does the same, we interpret it as having some deeper meaning – and if the grown-up signs the paper with a well-known artist name, we attribute both an underlying meaning and a monetary value to it. Creativity is thus something which isn’t only a result of the effort of a producer, but also very much the result of how the result is viewed by the consumer.

Computational creativity can involve computer programmes that in themselves are creative, but also systems that are able to recognise and access creativity, as well as programmes that assist humans in creative tasks. There are many creativity-supporting systems (e.g., Adobe PhotoShop), and a few systems that themselves (possibly) are creative, such as “The Painting Fool” and “AARON” (two artificial artists). There are also systems that draws art based on textual or musical input (such as Stable Diffusion), or generates music based on images, or systems for automatic captioning of images and videos that produce short texts matching the visual content (with a specific challenge if parts of the texts may need to be emphasised). A master thesis on the topic could address any of these strands and approaches, depending on the student(s) background and interests.

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

There are three basic machine learning paradigms: supervised learning, unsupervised learning, and reinforcement learning. Supervised learning requires data instances annotated with the correct output labels, but the other paradigms assume no such labelled data. Reinforcement learning (RL) is primarily concerned with what actions an intelligent agent (or group of agents) should take in a specified environment in order to maximise some cumulative reward, while simultaneously exploring the environment and exploiting previously accumulated knowledge.

Within computational creativity applications, reinforcement learning from human feedback has been utilised for language generation in systems such as ChatGPT but can tentatively have a wider usage in art or music, or can be used form part of a generative adversarial network (GAN) that so far mainly have been explored in image generation.

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

Computational linguistic creativity can primarily be aimed at creating systems that either are creative themselves (e.g., generate poetry, generate screen plays, write song lyrics, produce analogies or metaphors) or try to understand creativity (e.g., identify sarcasm, understand humour or interpret rhymes).

A master thesis project in the field could concentrate on one or several of these different aspects of computational linguistic creativity (e.g., generate and evaluate computational poetry, translate on-line jokes between two languages, or generate plot twists in movies).

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

Computers have been used in music both as support for creativity and as creative agents themselves, and both for the composition of the music scores and for writing lyrics. The first algorithmic composition system appeared already in the 1950s (the Illiac suite, Hiller & Isaacson 1958), and since then rule-based systems, stochastic methods, grammar-based methods, neural networks, and evolutionary methods have all been utilised to compose music in a specific genre, style or mood; to identify which key(s) or moods a music piece is written in; for generating lyrics; for music information retrieval; or for automatic music transcription. A master thesis on the topic could address any of these strands and approaches, depending on the student(s) background and interests.

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

It can be vital both for mental health reasons and for security to identify users at risk in social media. That can be users showing signs of mental health issues (depression, suicide, eating disorders, etc.) as well as users being at risk of being radicalised (as extremists, school shooters and other types of terrorists, , etc.), or users being targeted by predators or extremist organisations. 
This thesis work would experiment with applying various machine learners to social media text. To fully utilise deep learning, a substantial amount of data will need to be gathered, either from previous research or specifically for the task. 

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

The way we write texts give a lot of information about the background personalities of the authors: their age, gender, native language (if writing in a foreign language), if they're human or bots, and possibly their actual identity. This type of information can be used to, e.g., give fair indications of user profiles, to deduce if a text (or a part of it) has been plagiarised, or to uncover social media software misuse. The thesis could thus focus on tasks such as author profiling (what can we say about the author, e.g., their gender, age, if they're a human or a bot), author identification (did a specific author write this text?) and/or plagiarism detection (did somebody else than the author claiming the text actually write all or part of the text?)

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

Large language models form the basis of almost all currently topical AI research, making it vital to identify and rectify different types of demographic biases in those models (be that bias based on gender or sexual identity, or on cultural, ethical or social background, etc.). This has triggered intense research on fair representation in language models, aiming both at building and using unbiased training and evaluation datasets, and at changing the actual learning algorithms themselves. There is still a lot of room for improvements though, both in identifying and quantifying bias, in developing dibiasing methods, and in defining bias as such.

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

Properly identifying hatespeech is a pressing issue for social media sites as well as for smaller companies, clubs, and organisations that allow for user-generated content. Many such sites currently use slow, manual moderation, which mean that abusive posts will be left online for too long without appropriate action being taken or that content will be published with delay (which might be unacceptable to the users, e.g., in online chat rooms).

The thesis project would look into previous efforts to identify hate speech and cyber bullying, as well as available flame-annotated datasets from chat rooms, online games, Wikipedia and Twitter, and investigate various machine learning methods to identify such language.

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

Several models can be used to find out how users’ social media networks, behaviour and language are related to their ethical practices and personalities, Such models include Schwartz’ values and ethics model and Goldberg's Big 5 model that defines personality traits such as openness, conscientiousness, extraversion, agreeableness and neuroticism. The thesis project would investigate applying such models to social media text and how the user personalties are reflected by the social networks that they participate in and develop.

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

The project will be carried out in collaboration with Benjamin Cretois at the Norwegian Institute for Nature Research (NINA). NINA is Norway’s leading institution for applied ecological research, with broad-based expertise on the genetic, population, species, ecosystem and landscape level, in terrestrial, freshwater and coastal marine environments. The student(s) will be collaborating with researchers at the Miljødata department who focus on employing advanced technologies to study and protect biodiversity, with a particular emphasis on bioacoustics and bird species.

Problem Description: 

NINA are currently using an off-the-shelf deep learning-based bird classifier (BirdNET) that effectively identifies common bird species. However, its performance significantly drops for the rarer species, limiting our understanding of biodiversity in diverse ecosystems. Properly accounting for rare species is critical to establish ecosystem health and it is essential to develop automatic methods efficient at detecting and classifying them.

The project aims to refine the model using advanced machine learning techniques such as fine-tuning and few-shot learning strategies, potentially exploring other innovative methods to enhance species recognition, particularly for underrepresented species. The expected outcome includes improved model accuracy and recall for rare species, contributing to more comprehensive biodiversity assessments.

Data: 

The student(s) will have access to the Sound of Norway (https://thesoundofnorway.com) dataset. It contains approximatively 7 terabyte of audio files recorded by BUGG recording devices at 41 different sites throughout Norway. All of the detections made by BirdNET have been verified by an ornithologist. Moreover, a few files have completely been reviewed by the ornithologist to compute the model recall. This dataset should provide a robust foundation for evaluation of the methods that will be developed.

Risks / Challenges:
Improving model capabilities to generalise well to rare events (rare bird calls in the case of this subject) is a hot topic and numerous methods are being developed. The primary scientific challenge lies in the method's ability to generalise well to rare species with limited examples. 

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

The project is aimed at the automatic classification of sentiment in texts on Twitter or other social media, tentatively addressing issues involving negation and/or figurative language, that is, language which intentionally conveys secondary or extended meanings (such as sarcasm, irony and metaphor). Figurative language creates a significant challenge for sentiment analysis systems, as direct approaches based on words and their lexical semantics often are inadequate in the face of indirect meanings. A subgoal of the project would then be to find a set of tweets rich in figurative language, but the main goal would be to determine whether the writer of such tweets has expressed a positive or negative sentiment (commonly displaying positive or negative emotions towards a product, person, political party, etc.), and possibly the degree to which this sentiment has been communicated.

[ Skjul beskrivelse ]

[ Vis hele beskrivelsen ]

Languages change rapidly over time and language users adapt to different situations and setting. Studying how language and communicative processes evolves is a highly multi-disciplinary task involving machine and language learning as well as biological and cultural evolution. The aim of this project will be to use methods such as evolutionary algorithms or reinforcement learning to investigate the main dynamics in language evolution.

When trying to understand the origins of languages, we can compensate for the lack of empirical evidence by utilising evolutionary computational methods to create simulations of how language may have evolved over time, e.g., by creating "language games" to simulate communication between agents in a social setting. In general, simulations on language evolution tend to have relatively small and fixed population sizes, something this study could aim to change.

[ Skjul beskrivelse ]