AFArthur Flexer

Arthur Flexer

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Arthur Flexer has a total of 39 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Experimental sense-making, musicking and no control: Performing black box music instruments

    by Thomas Grill, Arthur Flexer, Angélica Castelló, Marco Döttlinger, Patrik Lechner, Christian Grüny

    The text explores ‘Black Box Music,’ an artistic research experiment investigating human–technology relations in musical improvisation. The authors describe performances in which each musician plays an unfamiliar, complex, custom-built instrument designed by another team member. These ‘black box’ devices – ranging from AI-based systems to assemblages of analogue devices – cannot be fully understood or controlled, thus foregrounding questions of agency, sense-making, and aesthetic experience. Drawing on Actor-Network Theory, phenomenology, and philosophies of technology, the authors show how performers, instruments, and audience form a dynamic network of actants whose roles and intentions remain ambiguous. Players initially struggle to ‘read’ the instruments’ behaviour, shifting from reactive analysis toward proactive improvisation. This process blurs embodiment and alterity: the instruments alternately function as transparent extensions of the body and as quasi-autonomous others. Such ambiguity invites anthropomorphisation and even ritualised interaction, echoing historical entanglements of music, magic, and spirituality. Audience members, too, encounter indeterminate agency and must negotiate aesthetic meaning without clear attribution of sound to human or machine. The project demonstrates how complexity and unpredictability destabilise traditional notions of control, revealing improvisation as collective, participatory sense-making and highlighting the emergent, co-creative agency of both humans and technological instruments.

  • Black Box Music

    by Thomas Grill, Arthur Flexer, Patrik Lechner, Angélica Castelló, Marco Döttlinger


  • Open Access

    Latent Topic and Tempo Differences Between East and West Coast Hip-Hop

    by Ábel Boros, Arthur Flexer

    Our work presents a computational approach to exploring differences between Hip-Hop music from the US East and West Coast during the so-called “rap wars". This period in time was characterized by an intense rivalry prevalent especially in the lyrics of the music. We built a representative dataset of music from respective eminent artists and used topic modeling to analyze song lyrics. The resulting topic of “Vulgarity and Violence" is more attributed to West Coast artists and “Street Life and Rhythm" more to the East Coast. In addition we show that tempo of the music is slower for the West Coast and in general slowing down over the observation period. Our quantitative results align with existing theories concerning spatiality of Hip-Hop culture.

  • Open Access

    Are Inherently Interpretable Models More Robust? A Study In Music Emotion Recognition

    by Katharina Hoedt, Arthur Flexer, Gerhard Widmer

    One of the desired key properties of deep learning models is the ability to generalise to unseen samples. When provided with new samples that are (perceptually) similar to one or more training samples, deep learning models are expected to produce correspondingly similar outputs. Models that succeed in predicting similar outputs for similar inputs are often called robust. Deep learning models, on the other hand, have been shown to be highly vulnerable to minor (adversarial) perturbations of the input, which manage to drastically change a model's output and simultaneously expose its reliance on spurious correlations. In this work, we investigate whether inherently interpretable deep models, i.e., deep models that were designed to focus more on meaningful and interpretable features, are more robust to irrelevant perturbations in the data, compared to their black-box counterparts. We test our hypothesis by comparing the robustness of an interpretable and a black-box music emotion recognition (MER) model when challenged with adversarial examples. Furthermore, we include an adversarially trained model, which is optimised to be more robust, in the comparison. Our results indicate that inherently more interpretable models can indeed be more robust than their black-box counterparts, and achieve similar levels of robustness as adversarially trained models, at lower computational cost.

  • Open Access

    Charting the Universe of Metal Music Lyrics and Analyzing Their Relation to Perceived Audio Hardness

    by Isabella Czedik-Eysenberg, Oliver Wieczorek, Arthur Flexer, Christoph Reuter

    We analyze the relationship between the musical and the lyrical content of metal music by combining automated audio feature extraction and quantitative text analysis on a corpus of 124,288 song lyrics from this genre. Based on this text corpus, a topic model is first constructed using Latent Dirichlet Allocation (LDA). For a subsample of 503 songs, scores for predicting perceived musical hardness (heaviness) are extracted using an audio feature model and corroborated via listening tests. By combining both audio feature and text analysis, we (1) offer a comprehensive overview of the lyrical topics present within the metal genre and (2) are able to establish whether levels of hardness are associated with the occurrence of particularly harsh (and other) textual topics. Twenty typical topics are identified and projected into a topic space using multidimensional scaling (MDS). Positive correlations are found between musical hardness and textual topics dealing with “brutal death,” “dystopia,” “archaisms and occultism,” “religion and satanism,” “battle,” and “(psychological) madness,” while there is a negative association with topics like “personal life” and “love and romance.” Further, we show different prevalences of the identified topics within different metal subgenres. In summary, this article presents the first comprehensive computational corpus study of musical and lyrical content in metal music.


  • Open Access

    Validity in Music Information Research Experiments

    by Bob L. T. Sturm, Arthur Flexer

    Validity is the truth of an inference made from evidence, such as data collected in an experiment, and is central to working scientifically. Given the maturity of the domain of music information research (MIR), validity in our opinion should be discussed and considered much more than it has been so far. Considering validity in one's work can improve its scientific and engineering value. Puzzling MIR phenomena like adversarial attacks and performance glass ceilings become less mysterious through the lens of validity. In this article, we review the subject of validity in general, considering the four major types of validity from a key reference: Shadish et al. 2002. We ground our discussion of these types with a prototypical MIR experiment: music classification using machine learning. Through this MIR experimentalists can be guided to make valid inferences from data collected from their experiments.

  • Open Access

    Constructing adversarial examples to investigate the plausibility of explanations in deep audio and image classifiers

    by Katharina Hoedt, Verena Praher, Arthur Flexer, Gerhard Widmer

    Given the rise of deep learning and its inherent black-box nature, the desire to interpret these systems and explain their behaviour became increasingly more prominent. The main idea of so-called explainers is to identify which features of particular samples have the most influence on a classifier’s prediction, and present them as explanations. Evaluating explainers, however, is difficult, due to reasons such as a lack of ground truth. In this work, we construct adversarial examples to check the plausibility of explanations, perturbing input deliberately to change a classifier’s prediction. This allows us to investigate whether explainers are able to detect these perturbed regions as the parts of an input that strongly influence a particular classification. Our results from the audio and image domain suggest that the investigated explainers often fail to identify the input regions most relevant for a prediction; hence, it remains questionable whether explanations are useful or potentially misleading.


  • Open Access

    scikit-hubness: Hubness Reduction and Approximate Neighbor Search

    by Roman Feldbauer, Thomas Rattei, Arthur Flexer