Genre Detection Research Articles

Creating musical features using multi-faceted, multi-task encoders based on transformers

Computational machine intelligence approaches have enabled a variety of music-centric technologies in support of creating, sharing and interacting with music content. A strong performance on specific downstream application tasks, such as music genre detection and music emotion recognition, is paramount to ensuring broad capabilities for computational music understanding and Music Information Retrieval. Traditional approaches have relied on supervised learning to train models to support these music-related tasks. However, such approaches require copious annotated data and still may only provide insight into one view of music—namely, that related to the specific task at hand. We present a new model for generating audio-musical features that support music understanding, leveraging self-supervision and cross-domain learning. After pre-training using masked reconstruction of musical input features using self-attention bidirectional transformers, output representations are fine-tuned using several downstream music understanding tasks. Results show that the features generated by our multi-faceted, multi-task, music transformer model, which we call M3BERT, tend to outperform other audio and music embeddings on several diverse music-related tasks, indicating the potential of self-supervised and semi-supervised learning approaches toward a more generalized and robust computational approach to modeling music. Our work can offer a starting point for many music-related modeling tasks, with potential applications in learning deep representations and enabling robust technology applications.

Ensemble deep learning in speech signal tasks: A review

Machine learning methods are extensively used for processing and analysing speech signals by virtue of their performance gains over multiple domains. Deep learning and ensemble learning are the two most commonly used techniques, which results in benchmark performance across different downstream tasks. Ensemble deep learning is a recent development which combines these two techniques to result in a robust architecture having substantial performance gains, as well as better generalization performance over the individual techniques. In this paper, we extensively review the use of ensemble deep learning methods for different speech signal related tasks, ranging from general objectives such as automatic speech recognition and voice activity detection, to more specific areas such as biomedical applications involving the detection of pathological speech or music genre detection. We provide a discussion on the use of different ensemble strategies such as bagging, boosting and stacking in the context of speech signals, and identify the various salient features and advantages from a broader perspective when coupled with deep learning architectures. The main objective of this study is to comprehensively evaluate existing works in the area of ensemble deep learning, and highlight the future directions that may be explored to further develop it as a tool for several speech related tasks. To the best of our knowledge, this is the first review study which primarily focuses on ensemble deep learning for speech applications. This study aims to serve as a valuable resource for researchers in academia and in industry working with speech signals, supporting advanced novel applications of ensemble deep learning models towards solving challenges in existing speech processing systems.

GenSpecVidOnt: a reference ontology for knowledge based video analytics with multimodal genre detection

GenSpecVidOnt: a reference ontology for knowledge based video analytics with multimodal genre detection

Music genre detection using deep learning models

An automated music classification system is that automates the entire process of music classification, i.e., classifying music without any human intervention. Along with this, it is very important to have a good recommendation system that will help with features such as music classification. This paper proposes an automated music classification system that will be very useful for both informational and entertainment purposes in the field of music. This system, based on a musicrelated Artificial Intelligence (AI) algorithm, automatically categorizes different types of music corresponding to different music genres, e.g., Hip Hop, Jazz, Rock, Blues, etc. Another feature added to this system is to recommend similar songs classified by the system. The architecture of the system and the algorithms used at each stage are described and implemented in this paper. The system also provides a lyrics classification module that generates and provides users with lyrics of the user choice. Compared to realistic musical classification, which has always been a difficult problem as it may lack structure or rationality.

Signal Processing and Machine Learning Approaches and Evaluation for Indian Classical Music

In the field of signal processing, a considerable amount of work has been done in the analysis and implementation of various techniques related to speech signal processing. When it comes to the Music signal processing, the techniques that are originally developed for speech signal processing have been applied to music signal with good results. In comparison with Indian classical music, lot of research work done in Western classical music on music audio analysis like information retrieval, singer/instrument identification, genre detection, etc.,. This paper provides an overview of some signal processing, and machine learning techniques that specifically applied to the Indian classical music for analyzing and identifying features such as tonic identification, genre classification, raga recognition, music transcription, rhythm, and timbre. We will look at how particular features of Indian Classical music signals impact and how these features are extracted.

APPLICATION OF DEEP LEARNING ALGORITHMS IN MUSICOLOGY: AN OVERVIEW

Nowadays, music could be an exceptionally important and perhaps indivisible portion of people's way of life. Music involves various aspects right from generation, its genre detection to its transcription. Composing music may be an exceptionally intrusive challenge that tests the composer’s inventive capacity, whether it is a human or a computer. In spite of the fact that there have been many arguments on the matter, nearly all of the music is some regurgitation or change of a sonic idea created before. Thus, with sufficient information and the proper algorithm, deep learning ought to be able to create music that would sound human. There are numerous genres in musicology which are very different from each other, coming about in individuals having diverse inclinations in musicology. It is a critical and an important issue to classify music and to suggest individuals modern music in applications and stages. Automated music transcription is considered by many researchers to be a key empowering innovation in audio processing. Be that as it may, the execution of translation frameworks is still altogether underneath that of a human master, and correctness detailed in later a long time appears to have come to a constraint, in spite of the fact that the field is dynamic. In the paper, we try to analyse different algorithms enabling music generation, genre classification and music transcription.

Multi-label Movie Genre Detection from a Movie Poster Using Knowledge Transfer Learning

The task of predicting a movie genre from its poster can be very challenging owing to the high variability of movie posters. A novel approach for the generation of a multi-label movie genre prediction from its poster using neural networks that employ knowledge transfer learning has been proposed in this paper. This approach works on two fronts; one is aimed at creating a large, diverse and balanced dataset for movie genre prediction. The second front involves reframing the problem to simpler single-label multi-class classification and generating a multi-label multi-class prediction on a given movie poster as input. The experimental evaluation suggests that our approach generates a remarkable accuracy which is a result of a larger, evenly distributed dataset, simplifying the problem to a single-label multi-class classification problem and because of the use of knowledge transfer learning to extract higher-level feature.

On the influence of low-level visual features in film classification.

BackgroundIn this paper we present a model of parameters to aesthetically characterize films using a multi-disciplinary approach: by combining film theory, visual low-level video descriptors (modeled in order to supply aesthetic information) and classification techniques using machine and deep learning.MethodsFour different tests have been developed, each for a different application, proving the model's usefulness. These applications are: aesthetic style clustering, prediction of production year, genre detection and influence on film popularity.ResultsThe results are compared against high-level information to determine the accuracy of the model to classify films without knowing such information previously. The main difference with other film characterization approaches is that we are able to isolate the influence of high-level descriptors to really understand the relevance of low-level features and, accordingly propose a useful set of low-level visual descriptors for that purpose. This model has been tested with a representative number of films to prove that it can be used for different applications.

Improvised emotion and genre detection for songs through signal processing and genetic algorithm

SummaryMusical tunes are bundle of chords representing emotion which impart diverse of genres. Past history highlighted copious amount of research work emotion and genre classification with still increasingly rapid advancement. Music has various emotional forms as happy, sad, anger and fear. Its various genre forms are Classical, Country, Disco, Hip‐hop, Jazz, and Rock. These emotions and genre can be segregated by identifying the frequency of chords notes (swarams in Tamil music). This paper deals with identifying emotions and genre for classical music both western and south Indian classical music, viz, Carnatic music. The music was clipped and segmented to determining frequency of notes using Shortest Fast Fourier Transformation (STFT). Music features such as mel frequency, pitch beat, zero crossing rate, and spectral centroid were derived from the obtained frequency. Based on the audio features, emotion and genre were identified for the given data set with genetic algorithm as a classification technique. The MIREX‐Mood classification dataset was considered for listing out emotions. The songs from Million song data set and emotion classification repository were considered as ground truth for western classical music and group of Illayaraja Tamil film songs was considered to identify Carnatic music emotions. The classification was done using genetic algorithm. Mel frequency, pitch, and zero crossing rate were considered as individual representations to get best fit ratio and it is found to give accuracy percentage of 99.03%.

Stylometric analysis of classical Arabic texts for genre detection

PurposeIn the context of information retrieval, text genre is as important as its content, and knowledge of the text genre enhances the search engine features by providing customized retrieval. The purpose of this study is to explore and evaluate the use of stylometric analysis, a quantitative analysis for the linguistics features of text, to support the task of automated text genre detection for Classical Arabic text.Design/methodology/approachUnsupervised clustering and supervised classification were applied on the King Saud University Corpus of Classical Arabic texts (KSUCCA) using the most frequent words in the corpus (MFWs) as stylometric features. Four popular distance measures established in stylometric research are evaluated for the genre detection task.FindingsThe results of the experiments show that stylometry-based genre clustering and classification align well with human-defined genre. The evidence suggests that genre style signals exist for Classical Arabic and can be used to support the task of automated genre detection.Originality/valueThis work targets the task of genre detection in Classical Arabic text using stylometric features, an approach that has only been previously applied to Arabic authorship attribution. The study also provides a comparison of four distance measures used in stylomtreic analysis on the KSUCCA, a corpus with over 50 million words of Classical Arabic using clustering and classification.

Open set evaluation of web genre identification

Web genre detection is a task that can enhance information retrieval systems by providing rich descriptions of documents and enabling more specialized queries. Most of previous studies in this field adopt the closed-set scenario where a given palette comprises all available genre labels. However this is not a realistic setup since web genres are constantly enriched with new labels and existing web genres are evolving in time. Open-set classification, where some pages used in the evaluation phase do not belong to any of the known genres, is a more realistic setup for this task. In this case, all pages not belonging to known genres can be seen as noise. This paper focuses on systematic evaluation of open-set web genre identification when the noise is either structured or unstructured. Two open-set methods combined with alternative text representation schemes and similarity measures are tested based on two benchmark corpora. Moreover, we adopt the openness test for web genre identification that enables the observation of effectiveness for a varying number of known/unknown labels.

Tekstgenres analyseren op lexicale complexiteit met T‑Scan

Abstract Using T-Scan to analyse the lexical complexity of text genres T-Scan is a tool for the automatic analysis of Dutch text. This paper presents the first large-scale corpus analysis with T-Scan, focusing on lexical complexity. A collection of nearly 1000 text specimens was assembled, containing ten genres: travel blogs, celebrity news features, novels, textbooks for vocational secondary schools, textbooks for general secondary schools, news reports, opinion pieces, political programs, medical advice texts and research articles. The lexical complexity features in the analysis include morphology, word frequency, various word concreteness indices, personal pronouns, names and verb tense. Systematic genre differences are found, such that a genre detection model comprising 18 T-Scan features correctly identifies 83 percent of the corpus texts. Most lexical features differentiating genres intuitively relate to text topic complexity. A closer analysis is offered of the contrast between the two textbook samples in the corpus, which differ only in the educational levels they cater for. Again, topic variation seems a more important factor than stylistic variation. We demonstrate a new method to examine stylistic variation, which consists of within-genre comparisons using the genre prediction; more specifically, ‘deviant’ texts are compared to ‘typical’ members of their genre.

Learning to recognize webpage genres

Webpages are mainly distinguished by their topic (e.g., politics, sports etc.) and genre (e.g., blogs, homepages, e-shops, etc.). Automatic detection of webpage genre could considerably enhance the ability of modern search engines to focus on the requirements of the user’s information need. In this paper, we present an approach to webpage genre detection based on a fully-automated extraction of the feature set that represents the style of webpages. The features we propose (character n-grams of variable length and HTML tags) are language-independent and easily-extracted while they can be adapted to the properties of the still evolving web genres and the noisy environment of the web. Experiments based on two publicly-available corpora show that the performance of the proposed approach is superior in comparison to previously reported results. It is also shown that character n-grams are better features than words when the dimensionality increases while the binary representation is more effective than the term-frequency representation for both feature types. Moreover, we perform a series of cross-check experiments (e.g., training using a genre palette and testing using a different genre palette as well as using the features extracted from one corpus to discriminate the genres of the other corpus) to illustrate the robustness of our approach and its ability to capture the general stylistic properties of genre categories even when the feature set is not optimized for the given corpus.

Learning Subjective Language

Subjectivity in natural language refers to aspects of language used to express opinions, evaluations, and speculations. There are numerous natural language processing applications for which subjectivity analysis is relevant, including information extraction and text categorization. The goal of this work is learning subjective language from corpora. Clues of subjectivity are generated and tested, including low-frequency words, collocations, and adjectives and verbs identified using distributional similarity. The features are also examined working together in concert. The features, generated from different data sets using different procedures, exhibit consistency in performance in that they all do better and worse on the same data sets. In addition, this article shows that the density of subjectivity clues in the surrounding context strongly affects how likely it is that a word is subjective, and it provides the results of an annotation study assessing the subjectivity of sentences with high-density features. Finally, the clues are used to perform opinion piece recognition (a type of text categorization and genre detection) to demonstrate the utility of the knowledge acquired in this article.

Automatic Text Categorization in Terms of Genre and Author

The two main factors that characterize a text are its content and its style, and both can be used as a means of categorization. In this paper we present an approach to text categorization in terms of genre and author for Modern Greek. In contrast to previous stylometric approaches, we attempt to take full advantage of existing natural language processing (NLP) tools. To this end, we propose a set of style markers including analysis-level measures that represent the way in which the input text has been analyzed and capture useful stylistic information without additional cost. We present a set of small-scale but reasonable experiments in text genre detection, author identification, and author verification tasks and show that the proposed method performs better than the most popular distributional lexical measures, i.e., functions of vocabulary richness and frequencies of occurrence of the most frequent words. All the presented experiments are based on unrestricted text downloaded from the World Wide Web without any manual text preprocessing or text sampling. Various performance issues regarding the training set size and the significance of the proposed style markers are discussed. Our system can be used in any application that requires fast and easily adaptable text categorization in terms of stylistically homogeneous categories. Moreover, the procedure of defining analysis-level markers can be followed in order to extract useful stylistic information using existing text processing tools.