We are a (still) young and growing institute with innovative projects and motivated young researchers. The focus of our research and teaching is on Artificial Intelligence and Machine Learning.
We develop and study computational models and algorithms that permit computers to perceive and 'understand' aspects of the external world, where we interpret 'perception' in the widest sense of the word, as the extraction of useful high-level information and knowledge from complex, possibly low-level data (audio, video, images, sensor data, texts, or even the Internet).
This requires fundamental and applied research on AI, machine learning, pattern recognition, text / data / social media mining, signal processing, statistical data modeling and classification.
Our current research has a particular focus on intelligent audio and music processing. Further research topics include biometrics, cryptography, personalization, and recommender systems.
Our goal is to offer state-of-the-art research and teaching in this area, and to provide a teaching environment that permits students to get involved in real research projects as early as possible.
Explainable Models and Their Application in Music Emotion Recognition
ARI guest talk by Verena Haunschmid and Shreyan Chowdhury at the Acoustics Research Institute of the Austrian Academy of
Current machine learning models used for music processing, while generally working quite well, do not give meaningful or intuitive explanations for their predictions. In this talk we will give an overview of existing tools for interpretable machine learning and lay out how those tools (mostly developed for the image domain) can or cannot be applied in the music domain. In particular, we will talk about emotion recognition in music and a few approaches to make emotion predictions more useful in terms of explaining them in a perceptually meaningful way.
This research is supported by the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme, grant agreement No 670035 ("Con Espressione").
Gerhard Widmer to give a Keynote Presentation at the Augmenting Performance Conference
Much of current research in Artificial Intelligence and Music, and particularly in the field of Music Information Retrieval (MIR), focuses on algorithms that interpret musical signals and recognise musically relevant objects and patterns at various levels – from notes to beats and rhythm, to melodic and harmonic patterns and higher-level structure -, with the goal of supporting novel applications in the digital music world. This presentation will give the audience a glimpse of what computational music perception systems can currently do with music, and what this is good for. However, we will also find that while some of these capabilities are quite impressive, they are still far from showing (or requiring) a deeper „understanding” of music. An ongoing project will be presented that aims to take AI & music research a step further, going beyond surface features and focusing on the *expressive* aspects, and how these are communicated in music. We will look at recent work on computational models of expressive music performance and some examples of the state of the art, and will discuss possible applications of this research. In the process, the audience will be subjected to a little experiment which may, or may not, yield a surprising result.
CP’s researcher Christine Bauer on the Austrian public radio Ö1 on music and technology, and their impact on the music industry
Next week’s Ö1 Radiokolleg, by the Austrian Public Broadcaster ORF, is devoted to the topic of music and technology, and their impact on the music industry.
Among other experts in the field, the Radiokolleg will feature CP’s researcher Christine Bauer in part 2 and 4 of the program (1st and 3rd October 2019, respectively).
MediaEval Benchmarking Initiative for Multimedia Evaluation
MediaEval is a benchmarking that offers challenges in multimedia retrieval, access and exploration to allow researchers working in computer science and other multimedia related field an opportunity to work on tasks that are related to human and social aspects of multimedia. MediaEval aims is to promote reproducible research that makes multimedia a positive force for society.
Emotion and theme recognition is a popular task in music information retrieval that is relevant for music search and recommendation systems. The goal of the task is to automatically recognize the emotions and themes conveyed in a music recording using machine learning algorithms.
The team of the Institute of Computational Perception, namely Khaled Koutini, Shreyan Chowdhury, Verena Haunschmid, Hamid Eghbal-zadeh, Gerhard Widmer, won the challenge by developing the highest performing models to detect emotion and themes in music.