Application of artificial intelligence in music generation: a systematic review

IAES International Journal of Artificial Intelligence (IJ-AI)
Vol. 13, No. 4, December 2024, pp. 3715~3726
ISSN: 2252-8938, DOI: 10.11591/ijai.v13.i4.pp3715-3726  3715
Journal homepage: http://ijai.iaescore.com
Application of artificial intelligence in music generation: a
systematic review
Frank Manuel Gutiérrez Paitan, María Acuña Meléndez, Christian Ovalle
Departmento de Ingenieria, Universidad Tecnológica del Perú, Lima, Perú
Article Info ABSTRACT
Article history:
Received Jan 3, 2024
Revised Feb 29, 2024
Accepted Mar 21, 2024
Our analysis explores the benefits of artificial intelligence (AI) in music
generation, showcasing progress in electronic music, automatic music
generation, evolution in music, contributions to music-related disciplines,
specific studies, contributions to the renewal of western music, and hardware
development and educational applications. The identified methods
encompass neural networks, automation and simulation, neuroscience
techniques, optimization algorithms, data analysis, and Bayesian models,
computational algorithms, and music processing and audio analysis. These
approaches signify the complexity and versatility of AI in music creation.
The interdisciplinary impact is evident, extending into sound engineering,
music therapy, and cognitive neuroscience. Robust frameworks for
evaluation include Bayesian models, fractal metrics, and the statistical
creator-evaluator. The global reach of this research underscores AI's
transformative role in contemporary music, opening avenues for future
interdisciplinary exploration and algorithmic enhancements.
Keywords:
Artificial intelligence
Automation and simulation
Music generation
Music processing
Optimization algorithms
Systematic review
This is an open access article under the CC BY-SA license.
Corresponding Author:
Christian Ovalle
Departmento de Ingenieria, Universidad Tecnológica del Perú
Lima, Perú
Email: dovalle@utp.edu.pe
1. INTRODUCTION
In recent times, there has been an increase in interest and development in the field of music creation
[1]. Music is a fundamental part of human culture that has undergone a significant evolution over the
centuries, adapting to different cultures, styles, and technologies [2]. In view of this, we must bear in mind
that the complex task of generating musical chords has been tackled by different methods throughout history
[3]. In addition, the combination of music theory with computational algorithms has made possible the
creation of more accurate and adaptable methods of chord generation [4]. During the 1990s, approaches
based on artificial intelligence (AI), such as expert systems and neural networks, have emerged [5]. In
addition, the wide variety of musical styles has generated a demand for methods to generate chords that are
flexible and capable of addressing diverse genres [6], therefore, the initial methods for generating chords
were based on rules derived from musical knowledge [7]. In view of this, the generation of music through
models has undergone a paradigm shift with advances in AI and machine learning (ML) [8]. These
approaches were basically simple and had restrictions in terms of the diversity and expressiveness of the
chords that they could produce [9], bearing in mind that, in the most recent decade, approaches that rely on
ML, such as genetic algorithms and deep neural networks, have emerged [10]. Also, the use of deep learning
strategies to produce a variety of content (sound track, tracks, and bases) is on the increase [11], in view of
this, the incorporation of AI methods has radically transformed the production of chords, making possible the
automatic generation of complicated harmonic sequences [8]. In particular the interactive creation of chords,

 ISSN: 2252-8938
Int J Artif Intell, Vol. 13, No. 4, December 2024: 3715-3726
3716
which has opened up new possibilities in the collaboration between musicians and computer systems,
enabling user participation in the creative process [12]. This analysis therefore investigates the historical
trajectory of the techniques used to create chords, highlighting the fundamental contributions that have taken
this field towards unexplored creative horizons [13]. In this context, as we move into the era of AI applied to
music, chord creation emerges as an exciting and ever-changing area that merges human inventiveness with
computational power [14].
2. METHOD
The methodology of the systematic review was developed with the objective of investigating
information on the evolution of musical chord generation methods. For this purpose, searches were carried
out considering quality criteria related to this topic. To carry out this research, the population, intervention,
comparison, outcome (PICO) approach was applied. This method provides a specialized format for the
formulation of the question, and its elements can be seen in the table corresponding to the PICO
questionnaire [15], [16].
The PICO method is a valuable tool that assists researchers in identifying literature relevant to their
research questions [17], in addition, we have also studied the principles of the preferred reporting items for
systematic reviews and meta-analyses (PRISMA) statement for the proper construction of the systematic
review, which, according to the PRISMA statement, is achieved by using three stages: planning the review,
carrying out the respective review and, finally, reporting on the review [18], it also makes it easier for
researchers to communicate the results of their systematic reviews clearly and concisely. In this specific
context, the application of the PICO approach will help to present the results of the systematic review in a
structured and understandable way [19].
The integration of quality criteria in the searches is essential, as it guarantees the relevance and
reliability of the studies selected in the systematic review [9], effective communication of results is also an
essential element in systematic reviews, and adherence to the PICO approach simplifies the organized and
understandable presentation of findings [20]. The focus on incorporating quality criteria during the searches
not only ensures the choice of relevant studies but also reinforces the methodological basis of the systematic
review, thus promoting the reliability and validity of the results obtained [21], in line with the above, the
effectiveness of the systematic review methodology is supported by the application of the principles of the
PRISMA statement. This ensures systematic planning and execution, as well as the transparent presentation
of the information collected, following the three essential phases of the process [9].
2.1. Research questions
The interest in the development of music generation is evidenced in the literature, where methods
are examined from their initial applications to the most current approaches [22]; recent studies have also
analyzed the influence of technological advances on the evolution of music generation methods, highlighting
the connection between technology and musical creativity [23], because of this, the elaboration of tables in a
systematic review is essential to organize and synthesize the information collected. This process provides a
clear view of the existing literature and facilitates the comparison of results between different studies [24]; on
the other hand, the effectiveness of tables in a systematic review extends to the presentation of information
clearly and concisely. This enables readers to quickly understand the breadth and variability of the included
studies [25]. With this in mind, Table 1 presents the PICO question and its components, and Table 2 details
the keywords used in the PICO tool.
Table 1. PICO question and its components
Question PICO
RQ: How has the generation of musical chords evolved over time and what are the key methods used in this evolution?
Components
RQ1: What were the first methods used for generating musical chords and how have they evolved over time?
RQ2: What technological advances have influenced the evolution of musical chord generation methods?
RQ3: What are the most commonly used contemporary methods for generating musical chords?
RQ4: How do these methods compare in terms of effectiveness and applicability in different musical genres?
2.2. Search strategy
In addition, the keywords used in formulating the search equation were chosen. This was done
because only some words classified in the original search equation were used in the final formulation. This
process aimed to optimize the search for documents that would provide more relevant information for our
systematic review.

Int J Artif Intell ISSN: 2252-8938 
Application of artificial intelligence in music generation: a systematic … (Frank Manuel Gutiérrez Paitan)
3717
Table 3 details the elements of the PICO question, where P: “technology” OR “generation” OR
“technological advances” OR “AI music generation,” I: “evolution” OR “contemporary methods” OR
“applicability,” C: “music” OR “classical music” OR “ pop music,” O: generation, C: “musical.” Initially, the
research was conducted using the search equation derived from these keywords, as shown in the search
equation. After the search, “92” articles were identified in Scopus and “33” in PubMed.
Table 2. Keywords in the PICO tool
Keywords In Pico Tools
P (Problem /
Population)
Evolution of musical chords
How has the generation of musical chords evolved over time
and what are the key methods used in this evolution?
(ALL (history AND of AND chord AND
generation OR early AND chord AND
generation AND methods) AND ALL
(impact AND evolution AND music))
I (Intervention) Chord generation methods
What were the initial methods for generating musical chords
and what has been their impact on the evolution of this
practice?
(ALL (technological AND advances OR
technology AND in AND chord AND
generation) AND ALL (influence AND
evolution AND music))
C (Comparison) Comparison of generation methods
What technological advances have played a crucial role in the
transformation of chord generation methods over time?
(ALL (contemporary AND chord AND
generation AND methods OR current
AND technologies AND in AND music)
AND ALL (difference AND evolution
AND music))
O (Results) Efficiency of generation methods
What are the most relevant and widely used chord generation
methods today, and how do they differ from previous
methods?
(ALL (efficacy AND applicability AND
contemporary AND generation AND
methods OR current AND technologies
AND in AND music) AND ALL (musical
AND genres AND music))
C (Context) Musical genres, application in classical music, pop, and jazz.
Historical context of music. How do contemporary methods
compare in terms of effectiveness and applicability, especially
in different musical genres?
(context OR application) AND (“musical
genres” OR “classical music” OR “pop
music” OR “jazz music”) AND
(“historical context” OR “historical
evolution”) AND music
Table 3. Search on data base
Search Equation TOTAL
SCOPUS (TITLE-ABS-KEY (“technology” OR “generation” OR “technological advances” OR
“generation of music with AI”) AND TITLE-ABS-KEY (“evolution” OR “contemporary
methods” OR “applicability”) AND TITLE-ABS-KEY (“music” OR “classical music” OR
“pop music”) AND TITLE-ABS-KEY (“generation”) AND TITLE-ABS-KEY (“musical”))
92
PUBMED (“Technology” OR “generation” OR “technological advances” OR “AI”) AND (“evolution”
OR “contemporary methods” OR “applicability”) AND (“music” OR “classical music” OR
“musical”) AND (“musical”)
33
TOTAL 125
Starting a systematic review with the accurate formulation of search equations is crucial to ensure
the completeness and specificity of the research [26]. Therefore, the careful choice of keywords in a
systematic search is essential to maximize the relevance and salience of the retrieved papers [27]. Therefore,
to ensure that the articles considered in our research on how AI affects current music development are
relevant and consistent, Table 4 was developed where the criteria used to determine which articles to include
and which to exclude are presented.
Table 4. Inclusion and exclusion criteria
Inclusion criteria Exclusion criteria
CI 1: Studies should specifically address the use of AI in the
creation, production, or analysis of contemporary music.
CE 1: Papers not directly related to the use of AI in
contemporary music are excluded.
CI 2: Papers should provide evidence or empirical analysis on the
impact of AI on contemporary music.
CE 2: Exclude research that does not present relevant results
or significant analysis on the role of AI in the evolution of
contemporary music.
CI 3: Studies may come from a variety of disciplines, such as
computer science, music, AI, or related fields, as long as they are
directly related to the convergence of AI and contemporary music.
A review of titles, summaries, or abstracts was carried out, and based on this review, 45 documents
were excluded from Scopus and 12 documents from PubMed, mainly because they were not studies

 ISSN: 2252-8938
3718
developed within the topic being focused on. According to the inclusion and exclusion criteria, 45 articles
were considered appropriate, after separating 4 duplicate articles present in both databases results. After
completing the selection of studies meeting the inclusion and exclusion criteria in several databases, such as
Scopus and PubMed, a total of 45 papers were collected as part of the systematic review, as can be seen in
Figure 1 in more detail.
Figure 1. PRISMA flow chart
Based on the data provided in the Table 5, it is possible to identify the advantages derived from
progress in the field of electronic music [28], automatic music creation [29]–[36], development in music
[37]–[47], contributions to music and related fields [48]–[57]. Similarly, specific investigations and
explorations have been carried out [58]–[70]. The importance of the contributions to the transformation of
western music is also highlighted [71] and progress in the creation of educational programs and devices [72].
This perspective indicates that developments in electronic music not only have a direct influence on musical
transformation [37]–[47] but also generate significant contributions to related fields [48]–[57]. The
automated creation of music, as demonstrated in the mentions [29]–[36], has created new opportunities and
has acted as a driver for the transformation of western music [71].
Table 5. Benefits obtained
# Benefits Implementation Reference
1 Progress in electronic music Optimization of artistic expression through technology [28]
2 Automatic music generation Facilitating creativity through automatic tools [29]–[36]
3 Evolution in music Adaptation to cultural and technological transformation [37]–[47]
4 Contributions to music and related disciplines Enrichment through connections between different disciplines [48]–[57]
5 Specific studies and explorations In-depth study of specific aspects to broaden knowledge [58]–[70]
6 Contributions to the renewal of western music Modernization of classical music with contemporary
approaches
[71]
7 Hardware development and educational
applications
Improving music education through technological tools [72]
Special contributions to music and related fields [48]–[57] highlight the relevance of these advances
in artistic and academic growth within the musical field. In addition, the detailed research and explorations of
the [58]–[70] point to a constant dedication to research and the constant search for improvements in the field
of electronic music. In addition, it is essential to highlight the importance of progress in the creation of
educational devices and programs [72]. They not only promote technological advances in the field of music
but also collaborate in the training and education of new talents in this field. Table 6 details the methods and
models implemented, which play an important role in the evolution of this field. It is important to highlight
the innovative models that have contributed to the musical transformation.

3719
Table 6. Methods and models implemented
# Method/Models Reference
1 Neural networks [28]–[30], [33], [42], [57], [64], [69], [71]
2 Automation and simulation [34]–[36], [48], [49], [72]
3 Neuroscience techniques [37], [38], [41], [53], [54], [56], [60], [70]
4 Optimization algorithms [62], [67], [68]
5 Data analysis and bayesian models [44], [50]–[52], [59]
6 Computational algorithms [31], [32], [47], [61], [66]
7 Music processing and audio analysis [39], [40], [44]–[46], [55], [58], [63], [65]
3. RESULTS AND DISCUSSION
In this part, a bibliometric review and a thorough analysis of previous research is presented. In the
first section, the connections between the concepts of musical generation and understanding, as well as the
visualization of density, are revealed. In the second part, the scientific gap identified between the papers
suggested in this study is investigated, with the aim of elaborating an architectural model that facilitates the
implementation of an algorithm to optimize and improve efficiency in the musical domain.
3.1. Bibliometric analysis
Bibliometrics is a field that employs statistical and mathematical approaches for the purpose of
examining scientific production [73], in addition, bibliometrics is used to analyze the impact of research,
identify trends in the development of science and technology, and evaluate the scientific productivity of
researchers and institutions [74]. Therefore, bibliometric analysis emerges as a valuable tool not only for
researchers but also for science administrators and policymakers. Its application contributes to enriching the
understanding of scientific research and its influence on society [75].
In Figure 2, you can see the network map provided, which shows three groups or groups of thematic
data that summarize significant relationships between key terms. In the first group, which focused on audio
and music computing, words like “audio acoustics” and “computer music” are interconnected, suggesting a
close link between acoustics and music creation through computer technologies. Furthermore, the presence of
terms such as “fitness functions” and their connection with “evolutionary algorithms” and “genetic
algorithms” indicates a relationship in the field of evolutionary algorithms and fitness functions. Likewise,
connecting “user interfaces” with musical terms suggests the integration of user interfaces into musical
applications.
Figure 2. Linkages between common terms by bibliometric mapping
The second cluster, which encompasses social sciences and humanities, reveals remarkable
connections. The keyword “article” acts as a central node connected to “cultural transmission”, “human”,
“humans”, “iterated learning”, and “music”. This interweaving suggests that the articles address topics
ranging from cultural transmission to music. The connection of “cultural transmission” with “human”,
“humans”, and “iterated learning” reinforces the relationship between cultural transmission and human
interaction. In addition, the connection of “music” con “human” y “humans” highlights a relationship
between music and human studies. At the same time, “iterated learning” connects with “human” and
“music”, highlighting its relationship with cultural evolution and music.
Lastly, the third cluster, focused on AI and evolutionary algorithms, shows that “artificial
intelligence” is connected to “evolutionary algorithms”, “genetic algorithms”, and “fitness functions”. This
points to an intrinsic relationship between AI and evolutionary algorithms. The connection between

 ISSN: 2252-8938
3720
“evolutionary algorithms” and “genetic algorithms” indicates a close relationship in evolutionary computing.
In addition, “genetic algorithms” are connected to “fitness functions” and “user interfaces”, suggesting their
application in the optimization of functions and user interfaces.
Figure 3 visually represents the density from a bibliometric analysis. It provides a quick and
understandable view of the intensity and distribution of connections between bibliometric elements,
contributing to decision-making in monitoring the temporal evolution of bibliometric interconnections.
Artistic expression through music has undergone a captivating evolution over the centuries. From classical
symphonies to contemporary rhythms, musical creativity has been influenced by diverse currents. In the
modern era, it has undergone a unique transformation through the integration of AI. In exploring new musical
possibilities, researchers have employed models based on recurrent neural networks [71], such as the
algorithm Opt-aiNet NS [69], and the neural networks of long short-term memory (LSTM) [33], these
techniques have enabled the generation of distinctive musical pieces and the exploration of previously
unexplored soundscapes.
Figure 3. Visual representation of density from a bibliometric analysis
The incorporation of artificial neural networks and genetic algorithms [30], has elevated music
production to an entirely different sphere. These methodologies, which take inspiration from biological evolution,
make it possible to create music in an evolutionary way, adjusting and perfecting itself over time. Cellular
automata, a dynamic paradigm at its core [72], have found their position in the application of AI to the field of
music [48], its use in the interactive field of evolutionary generation [36] has led to the possibility of creative
collaboration between the machine and the musician, resulting in unique and amazing musical creations.
The application of neuroscience methods, such as magnetoencephalography and
electroencephalography [70], has given researchers the opportunity to explore the minds of musicians and
gain a deeper understanding of creative processes. Neuroimaging techniques, such as functional magnetic
resonance imaging and positron emission tomography, have been used for this purpose [53] and have become
essential tools for investigating the connection between brain activity and music generation. The
incorporation of algorithms such as hierarchical qualitative context [68], the particle swarm optimization
method [67], and the multi-objective evolutionary method NSGA-II [62], has made possible a more effective
optimization of musical production, resulting in the elaboration of compositions that stand out for their
exceptional quality and complexity. With respect to evaluation and comparison, Bayesian models [59], the
statistical creator-evaluator [44], and fractal metrics [50] have established a robust framework for evaluating
the effectiveness and quality of parts generated using AI methods.
The combination between technology and music instruction has led to the development of a music
education system implemented on the android platform [63], additionally, the first-order markov model [55],
has proven to be a valuable resource for understanding and teaching musical organization, as well as the oral
transmission of musical traditions [40], has discovered a new channel through digital content analysis.
Computational approaches and simulations [66], the neuro dynamic code [61], and evolutionary approaches
[47] have enabled musicians and creators to explore a wider range of opportunities, challenging the norms
and taking music creation to new horizons. In short, the inclusion of AI in musical composition represents a
benchmark in the evolution of this artistic manifestation. As we continue to explore new technologies and
approaches, the musical realm will remain a breeding ground for innovation and creative expression.

3721
3.2. Manuscript analysis
Two (2) databases containing research results were used in the initial phase of the search, resulting
in the initial identification of a set of one hundred and twenty-five (125) documents. Subsequently, a
purification process was carried out following the PRISMA methodology. This process included the
elimination of duplicates and the exclusion of those documents that were not relevant to the research topic in
question. As a result of this rigorous procedure, a final selection of forty-five (45) manuscripts was obtained,
as detailed in Table 7. Phase 1 of this process focused on the initial search, while Phase 2 focused on the
elimination of duplicates and the exclusion of manuscripts unrelated to the topic of interest. Phase 3
ultimately revealed the 45 manuscripts deemed relevant to the research.
Table 7. Results obtained from the search
Type of database Name Initial search Final selection
Main database Scopus 92 37
Secondary PubMed 33 8
Total 125 45
For a more detailed understanding of the distribution of these manuscripts, Figure 4 presents the
breakdown of the number of manuscripts by each of the databases used, with Scopus being the main source
with 37 manuscripts, maintaining the highest number after applying the filters. Likewise, PubMed
contributed 8 relevant manuscripts. In addition, Figure 5 presents a line graph illustrating the annual
distribution of manuscript publications, broken down by database and corresponding indexes. To provide a
geographic perspective, Table 8 details the countries that have contributed the largest number of manuscripts
in this study. This breakdown adds an additional layer of analysis, offering perspective on the geographic
distribution of research on the topic under study.
Figure 4. Number of documents in PRISMA phases
Figure 5. Document number by year of publication

 ISSN: 2252-8938
3722
Table 8. Documents by publication location
# Country Quantity Reference
1 Brazil 2 [31], [72]
2 USA 18 [29], [32], [35], [37], [38], [41]–[43], [45], [46], [49], [50],
[52], [53], [56], [57], [59], [70], [71]
3 Australia 2 [48], [69]
4 Portugal 1 [58]
5 Greece 1 [67]
6 Belgium 1 [66]
7 Russia 2 [39], [65]
8 China 2 [63], [64]
9 South Africa 1 [62]
10 Canada 2 [60], [61]
11 Argentina 1 [58]
12 Korea 1 [55]
13 Austria 1 [54]
14 Malasia 1 [51]
15 Japan 2 [44], [47]
16 United Kingdom 2 [28], [40]
17 Sweden 1 [36]
18 Ireland 1 [34]
19 India 1 [33]
20 Kuwait 1 [30]
TOTAL 45
3.3. Work on the benefits of music generation
Over the years, music has undergone significant change, mainly influenced by technological
advances, especially in the field of electronic music, as evidenced in Table 9, these advances provide
significant advantages [28], shaping the music industry and creating new opportunities for artistic expression
[29]–[36]. The automated creation of music, as detailed in [29]–[36] is an innovation that redefines
conventional composition techniques. This method enables the creation of unique works through the use of
algorithms and AI, transforming in a revolutionary way the creative process for musicians and composers.
The development of music, as recorded in [37]–[47], the influence of technological tools has had a significant
impact on the evolution of music. From the introduction of synthesizers to digital audio workstations, these
advances have expanded the variety of musical genres and styles.
Table 9. Studies show the benefits obtained
# Benefits Quantity References
1 Progress in electronic music 1 [28]
2 Automatic music generation 8 [29]–[36]
3 Evolution in music 11 [37]–[47]
4 Contributions to music and related disciplines 10 [48]–[57]
5 Specific studies and explorations 13 [58]–[70]
6 Contributions to the renewal of western music 1 [71]
7 Hardware development and educational applications 1 [72]
TOTAL 45
The contributions of technology to the creation of music, as mentioned in [48]–[57], expand into
other disciplines, such as sound engineering, music therapy, and cognitive neuroscience, generating a
significant impact. Particular research and explorations, described in [58]–[70], have delved into music
creation algorithms and their influence on musical aesthetics, offering significant insights into AI-generated
music and its connection to human creativity. The impact of technology on the generation of music is not
restricted to contemporary styles, as noted in [71]. It has played a pivotal role in revitalizing western musical
traditions, preserving and renewing classical and historical forms. The advancement in hardware and
educational applications, highlighted in [72], has extended the boundaries of technology for music creation,
empowering musicians and facilitating music education. In conclusion, the advantages derived from
technology for music creation, detailed in Table 9, are diverse and encompass different aspects, from the
configuration of electronic music to the transformation of musical styles, thus inaugurating a new stage of
creativity in music.
3.4. Work on the methods used in music generation
The progress of music creation through novel approaches has been significant, and the application of
different techniques is evident in Table 10. Neural networks, cited in [28]–[30], [33], [42], [57], [64], [69],

3723
[71], have emerged as a leading approach. These computational models replicate the structure of the human
brain, enabling music generation through complex and adaptive patterns.
Table 10. Number of methods used with their manuscripts
# Method Quantity References
1 Neural networks 9 [28]–[30], [33], [42], [57], [64], [69], [71]
2 Automation and simulation 6 [34]–[36], [48], [49], [72]
3 Neuroscience techniques 8 [37], [38], [41], [53], [54], [56], [60], [70]
4 Optimization algorithms 3 [62], [67], [68]
5 Data analysis and Bayesian models 5 [44], [50]–[52], [59]
6 Computational algorithms 5 [31], [32], [47], [61], [66]
7 Music processing and audio analysis 9 [39], [40], [44]–[46], [55], [58], [63], [65]
TOTAL 45
Automation and simulation, as highlighted in [34]–[36], [48], [49], [72], is another significant tactic
in music creation. The use of automated algorithms and simulations effectively contributes to the production
of musical compositions, thus exploring new creative opportunities. The neuroscience methodologies,
registered in [37], [38], [41], [53], [54], [56], [60], [70], provide a unique approach to investigating the
connection between music and brain activity. These methodologies offer significant insights into
understanding how music impacts the human mind and is used in music creation.
Optimization algorithms, cited in [62], [67], [68], represent another crucial category. These
algorithms seek to improve music creation by optimizing parameters, raising the quality and diversity of
compositions. Data analysis and Bayesian modeling, explained in [44], [50]–[52], [59], is an approach based
on probability and data analysis. These models provide a deeper understanding of musical structures and
contribute to the generation of coherent and expressive music.
The computational algorithms, cited in [31], [32], [47], [61], [66], are essential tools in music
creation. These algorithms range from traditional approaches to more sophisticated methods, offering
versatility in music production. Music processing and audio analysis, as presented in [39], [40], [44]–[46],
[55], [58], [63], [65], plays a vital role in extracting relevant information from music and audio data. This
field encompasses techniques ranging from the analysis of musical features to the interpretation of audio
signals, thus contributing to the creation of innovative music.
4. CONCLUSION
The meticulous systematic review of 125 papers, following an exhaustive search of key databases such
as Scopus and PubMed, culminated in the final selection of 45 manuscripts. The geographical distribution of
contributions highlights the preeminence of the United States with 18 manuscripts, followed by Brazil (2),
Australia (2), and a diverse representation from countries such as Portugal, Greece, Belgium, Russia, China,
South Africa, Canada, Argentina, Korea, Austria, Malaysia, Japan, United Kingdom, Sweden, Ireland, India, and
Kuwait. This global reach underscores the breadth of research on the impact of AI on contemporary music. The
variety of methods implemented, from neural networks to optimization algorithms and neuroscience techniques,
highlights the complexity and versatility of the approaches adopted by the scientific community. In terms of
benefits, seven (7) categories were identified, ranging from advancement in electronic music to hardware
development and educational applications. These benefits translate into significant contributions to related
disciplines, such as sound engineering, music therapy, and cognitive neuroscience. With an emphasis on the
evaluation and comparison of the results, Bayesian models, fractal metrics, and the statistical creator-evaluator
are highlighted as robust frameworks to evaluate the effectiveness and quality of the pieces generated by AI
methods. In terms of number of methods used, nine (9) studies were found that employed neural networks, six (7)
in automation and simulation, eight (8) in neuroscience techniques, three (3) in optimization algorithms, five (5)
in data analysis and Bayesian models, five (5) in computational algorithms, and nine (9) in music processing and
audio analysis. The research, supported by robust evaluation methods, highlights the integration of disciplines
and persistent challenges. The conclusion underscores the transformation in artistic expression and the opening of
new interdisciplinary avenues, consolidating AI as an innovative force in contemporary music and opening
perspectives for future research and improvements in algorithm implementation.
ACKNOWLEDGMENTS
We would like to express gratitude to the Technological University of Peru, especially for providing
with valuable support at every stage of this research. From the provision of bibliographic resources to access
to specialized laboratories, their support and management have also been essential.

 ISSN: 2252-8938
3724
REFERENCES
[1] T. Carsault, J. Nika, P. Esling, and G. Assayag, “Combining real-time extraction and prediction of musical chord progressions for
creative applications,” Electronics, vol. 10, no. 21, Oct. 2021, doi: 10.3390/electronics10212634.
[2] Y. Zhu, J. Baca, B. Rekabdar, and R. Rawassizadeh, “A survey of AI music generation tools and models,” Cornell University,
vol. 1, pp. 2–39, 2023, doi: 10.48550/arXiv.2308.12982.
[3] D. Bouche, J. Nika, A. Chechile, and J. Bresson, “Computer-aided composition of musical processes,” Journal of New Music
Research, vol. 46, no. 1, pp. 3–14, Jan. 2017, doi: 10.1080/09298215.2016.1230136.
[4] A. A. Al-Omoush, A. A. Alsewari, H. S. Alamri, and K. Z. Zamli, “Comprehensive review of the development of the harmony
search algorithm and its applications,” IEEE Access, vol. 7, pp. 14233–14245, 2019, doi: 10.1109/ACCESS.2019.2893662.
[5] J. Osmalskyj, M. V. Droogenbroeck, and J. Embrechts, “Neural networks for musical chords recognition,” Actes des Journées
d’Informatique Musicale (JIM), pp. 39–46, 2012.
[6] J. Possik, C. Yaacoub, S. Gorecki, G. Zacharewicz, and A. Dambrogio, “A model based systems engineering approach to
automated music arrangement,” in Proceedings of the 2021 Annual Modeling and Simulation Conference, ANNSIM 2021, Jul.
2021, pp. 1–12, doi: 10.23919/ANNSIM52504.2021.9552105.
[7] H. Lim, S. Rhyu, and K. Lee, “Chord generation from symbolic melody using BLSTM networks,” Proceedings of the 18th
International Society for Music Information Retrieval Conference, ISMIR 2017, pp. 621–627, 2017, doi:
10.48550/arXiv.1712.01011.
[8] C. Hernandez-Olivan and J. R. Beltrán, “Music composition with deep learning: a review,” in Signals and Communication
Technology, 2023, pp. 25–50, doi: 10.1007/978-3-031-18444-4_2.
[9] M. Civit, J. C. -Masot, F. Cuadrado, and M. J. Escalona, “A systematic review of artificial intelligence-based music generation:
Scope, applications, and future trends,” Expert Systems with Applications, vol. 209, Dec. 2022, doi: 10.1016/j.eswa.2022.118190.
[10] M. Kikuchi and Y. Osana, “Automatic melody generation considering chord progression by genetic algorithm,” in 2014 6th
World Congress on Nature and Biologically Inspired Computing, NaBIC 2014, Jul. 2014, pp. 190–195, doi:
10.1109/NaBIC.2014.6921876.
[11] S. Ji, X. Yang, and J. Luo, “A survey on deep learning for symbolic music generation: representations, algorithms, evaluations,
and challenges,” ACM Computing Surveys, vol. 56, no. 1, pp. 1–39, Jan. 2023, doi: 10.1145/3597493.
[12] B. Barbot and T. Lubart, “Creative thinking in music: Its nature and assessment through musical exploratory behaviors,”
Psychology of Aesthetics, Creativity, and the Arts, vol. 6, no. 3, pp. 231–242, Aug. 2012, doi: 10.1037/a0027307.
[13] D. Herremans, C. H. Chuan, and E. Chew, “A functional taxonomy of music generation systems,” ACM Computing Surveys, vol.
50, no. 5, pp. 1–30, Sep. 2017, doi: 10.1145/3108242.
[14] Z. Yin, F. Reuben, S. Stepney, and T. Collins, “Deep learning’s shallow gains: a comparative evaluation of algorithms for
automatic music generation,” Machine Learning, vol. 112, no. 5, pp. 1785–1822, May 2023, doi: 10.1007/s10994-023-06309-w.
[15] E. L. -Ramírez and A. J. D. A. -Pantaleón, “Herramienta pico para la formulación y búsqueda de preguntas clínicamente
relevantes en la psicooncología basada en la evidencia,” Psicooncologia, vol. 11, no. 2–3, pp. 259–270, Dec. 2014, doi:
10.5209/rev_PSIC.2014.v11.n2-3.47387.
[16] C. M. D. C. Santos, C. A. D. M. Pimenta, and M. R. C. Nobre, “A estratégia PICO para a construção da pergunta de pesquisa e
busca de evidências,” Revista Latino-Americana de Enfermagem, vol. 15, no. 3, pp. 508–511, Jun. 2007, doi: 10.1590/S0104-
11692007000300023.
[17] I. Korstjens and A. Moser, “Series: Practical guidance to qualitative research. part 2: Context, research questions and designs,”
European Journal of General Practice, vol. 23, no. 1, pp. 274–279, Oct. 2017, doi: 10.1080/13814788.2017.1375090.
[18] R. Sharma, A. Shishodia, A. Gunasekaran, H. Min, and Z. H. Munim, “The role of artificial intelligence in supply chain
management: mapping the territory,” International Journal of Production Research, vol. 60, no. 24, pp. 7527–7550, Dec. 2022,
doi: 10.1080/00207543.2022.2029611.
[19] T. Kelder, B. R. Conklin, C. T. Evelo, and A. R. Pico, “Finding the right questions: exploratory pathway analysis to enhance
biological discovery in large datasets,” PLoS Biology, vol. 8, no. 8, pp. 11–12, Aug. 2010, doi: 10.1371/journal.pbio.1000472.
[20] A. MacFarlane, T. R. -Rose, and F. Shokraneh, “Search strategy formulation for systematic reviews: Issues, challenges and
opportunities,” Intelligent Systems with Applications, vol. 15, Sep. 2022, doi: 10.1016/j.iswa.2022.200091.
[21] J. M. L. Menon, F. Struijs, and P. Whaley, “The methodological rigour of systematic reviews in environmental health,” Critical
Reviews in Toxicology, vol. 52, no. 3, pp. 167–187, Mar. 2022, doi: 10.1080/10408444.2022.2082917.
[22] S. Ji, J. Luo, and X. Yang, “A comprehensive survey on deep music generation: multi-level representations, algorithms,
evaluations, and future directions,” arXiv-Computer Science, pp. 1-96, 2020, doi: 10.48550/arXiv.2011.06801.
[23] S. Shukla and H. Banka, “An automatic chord progression generator based on reinforcement learning,” in 2018 International
Conference on Advances in Computing, Communications and Informatics, ICACCI 2018, Sep. 2018, pp. 55–59, doi:
10.1109/ICACCI.2018.8554901.
[24] A. C. Bethel, M. Rogers, and R. Abbott, “Use of a search summary table to improve systematic review search methods, results,
and efficiency,” Journal of the Medical Library Association, vol. 109, no. 1, pp. 97–106, Jan. 2021, doi: 10.5195/jmla.2021.809.
[25] A. Johnston, S. E. Kelly, S. C. Hsieh, B. Skidmore, and G. A. Wells, “Systematic reviews of clinical practice guidelines: a
methodological guide,” Journal of Clinical Epidemiology, vol. 108, pp. 64–76, Apr. 2019, doi: 10.1016/j.jclinepi.2018.11.030.
[26] G. M. Tawfik et al., “A step by step guide for conducting a systematic review and meta-analysis with simulation data,” Tropical
Medicine and Health, vol. 47, no. 1, Dec. 2019, doi: 10.1186/s41182-019-0165-6.
[27] P. Veginadu, H. Calache, M. Gussy, A. Pandian, and M. Masood, “An overview of methodological approaches in systematic
reviews,” Journal of Evidence-Based Medicine, vol. 15, no. 1, pp. 39–54, Mar. 2022, doi: 10.1111/jebm.12468.
[28] B. Hayes, J. Shier, G. Fazekas, A. McPherson, and C. Saitis, “A review of differentiable digital signal processing for music and
speech synthesis,” Frontiers in Signal Processing, vol. 3, 2024, doi: 10.3389/frsip.2023.1284100.
[29] J. E. Tatem, J. R. Sochinski, and J. W. Roach, “Spatial planning for musical output,” International Journal of Intelligent Systems,
vol. 4, no. 2, pp. 201–218, Jun. 1989, doi: 10.1002/int.4550040206.
[30] I. A. Doush and A. Sawalha, “Automatic music composition using genetic algorithm and artificial neural networks,” Malaysian
Journal of Computer Science, vol. 33, no. 1, pp. 35–51, Jan. 2020, doi: 10.22452/mjcs.vol33no1.3.
[31] H. D. Bezerra, N. Nedjah, and L. D. M. Mourelle, “Reconfigurable hardware architecture for music generation using cellular
automata,” in 2014 IEEE 5th Latin American Symposium on Circuits and Systems, LASCAS 2014 - Conference Proceedings, Feb.
2014, pp. 1–4, doi: 10.1109/LASCAS.2014.6820309.
[32] M. Mauch, R. M. MacCallum, M. Levy, and A. M. Leroi, “The evolution of popular music: USA 1960–2010,” Royal Society
Open Science, vol. 2, no. 5, May 2015, doi: 10.1098/rsos.150081.

3725
[33] F. Shah, T. Naik, and N. Vyas, “LSTM based music generation,” in International Conference on Machine Learning and Data
Engineering, iCMLDE 2019, Dec. 2019, pp. 48–53, doi: 10.1109/iCMLDE49015.2019.00020.
[34] A. Breen, C. O’Riordan, and J. Sheahan, “Evolved aesthetic analogies to improve artistic experience,” in Computational
Intelligence in Music, Sound, Art and Design, vol. 10198, 2017, pp. 65–80, doi: 10.1007/978-3-319-55750-2_5.
[35] R. N. McArthur and C. P. Martin, “An application for evolutionary music composition using autoencoders,” in Artificial
Intelligence in Music, Sound, Art and Design, vol. 12693, 2021, pp. 443–458, doi: 10.1007/978-3-030-72914-1_29.
[36] K. Bäckman and P. Dahlstedt, “A generative representation for the evolution of jazz solos,” in Applications of Evolutionary
Computing, vol. 4974, 2008, pp. 371–380, doi: 10.1007/978-3-540-78761-7_40.
[37] O. L. Krasner, “Birth pangs, growing pains and sibling rivalry: Musical theatre in New York, 1900–1920,” in The Cambridge
Companion to the: Musical, Second Edition, Cambridge University Press, 2008, pp. 54–71, doi:
10.1017/CCOL9780521862387.004.
[38] V. Velardo and M. Vallati, “On the stylistic evolution of a society of virtual melody composers,” in Evolutionary and Biologically
Inspired Music, Sound, Art and Design, vol. 9027, 2015, pp. 249–260, doi: 10.1007/978-3-319-16498-4_22.
[39] B. D. Napreyev, “Five fugues from J. S. Bach’s well-tempered clavier (Book II) ‘In the rendition’ of Mozart,” Music Scholarship,
no. 1, pp. 37–45, Mar. 2019, doi: 10.17674/1997-0854.2019.1.037-045.
[40] M. A. -Tort, P. M. C. Harrison, H. Lee, and N. Jacoby, “Large-scale iterated singing experiments reveal oral transmission
mechanisms underlying music evolution,” Current Biology, vol. 33, no. 8, pp. 1472-1486.e12, Apr. 2023, doi:
10.1016/j.cub.2023.02.070.
[41] D. M. Hofmann, “A genetic programming approach to generating musical compositions,” in Evolutionary and Biologically
Inspired Music, Sound, Art and Design, vol. 9027, 2015, pp. 89–100, doi: 10.1007/978-3-319-16498-4_9.
[42] L. Dong, “Using deep learning and genetic algorithms for melody generation and optimization in music,” Soft Computing, vol. 27,
no. 22, pp. 17419–17433, Nov. 2023, doi: 10.1007/s00500-023-09135-3.
[43] M. Youngblood, “Cultural transmission modes of music sampling traditions remain stable despite delocalization in the digital
age,” PLoS ONE, vol. 14, no. 2, p. e0211860, Feb. 2019, doi: 10.1371/journal.pone.0211860.
[44] E. Nakamura and K. Kaneko, “Statistical evolutionary laws in music styles,” Scientific Reports, vol. 9, no. 1, Nov. 2019, doi:
10.1038/s41598-019-52380-6.
[45] D. Sellin and T. Seppälä, “Digital music industry,” Music Industry Research Association Conference, pp. 899–909, 2018, doi:
10.13140/RG.2.2.27536.71681.
[46] R. R. A. Faria, R. A. Ruschioni, and J. A. Zuffo, “Wavelets in music analysis and synthesis: timbre analysis and perspectives,” in
Wavelet Applications in Signal and Image Processing IV, Oct. 1996, vol. 2825, pp. 950–961, doi: 10.1117/12.255309.
[47] R. M. MacCallum, M. Mauch, A. Burt, and A. M. Leroi, “Evolution of music by public choice,” Proceedings of the National
Academy of Sciences of the United States of America, vol. 109, no. 30, pp. 12081–12086, Jul. 2012, doi:
10.1073/pnas.1203182109.
[48] A. R. Brown, “Exploring rhythmic automata,” in Workshops on Applications of Evolutionary Computation, vol. 3449, 2005, pp.
551–556, doi: 10.1007/978-3-540-32003-6_57.
[49] A. K. Hoover, P. A. Szerlip, and K. O. Stanley, “Functional scaffolding for composing additional musical voices,” Computer
Music Journal, vol. 38, no. 4, pp. 80–99, Dec. 2014, doi: 10.1162/COMJ_a_00269.
[50] J. H. Jensen and P. C. Haddow, “Evolutionary music composition based on Zipf’s law,” in Genetic and Evolutionary
Computation Conference, GECCO’11 - Companion Publication, Jul. 2011, pp. 41–42, doi: 10.1145/2001858.2001883.
[51] E. H. H. Law and S. Phon-Amnuaisuk, “Learning and generating folk melodies using MPF-inspired hierarchical self-organising
maps,” in Simulated Evolution and Learning, vol. 7673, 2012, pp. 371–380, doi: 10.1007/978-3-642-34859-4_37.
[52] R. Dirksen, “Mixed modes and performance codes of political demonstrations and carnival in Haiti,” in Ethnomusicology: A
Contemporary Reader, Routledge, 2017, doi: 10.4324/9781315439167.
[53] M. Lumaca, A. Ravignani, and G. Baggio, “Music evolution in the laboratory: Cultural transmission meets neurophysiology,”
Frontiers in Neuroscience, vol. 12, Apr. 2018, doi: 10.3389/fnins.2018.00246.
[54] N. Komatović, “‘Modulate! Modulate! Modulate! But do not change the key.’ the development and transformation of the term
modulation in the 19th-century French theory,” Rasprave Instituta za Hrvatski Jezik i Jezikoslovlje, vol. 44, no. 2, pp. 517–526,
2018, doi: 10.31724/rihjj.44.2.12.
[55] D. Park, J. Nam, and J. Park, “Novelty and influence of creative works, and quantifying patterns of advances based on
probabilistic references networks,” EPJ Data Science, vol. 9, no. 1, Dec. 2020, doi: 10.1140/epjds/s13688-019-0214-8.
[56] P. Ferrero, “The ‘other lives’ of Harry Smith’s anthology of American folk music,” in Harry Smith’s Anthology of American Folk
Music: America Changed Through Music, Oxon, New York City: Routledge, 2017, pp. 246–259, doi: 10.4324/9781315586250-
24.
[57] L. J. Tardón, I. Barbancho, C. Roig, E. Molina, and A. M. Barbancho, “Music learning: automatic music composition and singing
voice assessment,” in Springer Handbooks, 2018, pp. 873–883, doi: 10.1007/978-3-662-55004-5_42.
[58] I. Márquez, “Dancing with zizek: Origins and evolution of the digital cumbia in buenos aires,” America Latina Hoy, vol. 78, pp.
87–104, Apr. 2018, doi: 10.14201/alh20187887104.
[59] A. Ravignani, B. Thompson, T. Grossi, T. Delgado, and S. Kirby, “Evolving building blocks of rhythm: How human cognition
creates music via cultural transmission,” Annals of the New York Academy of Sciences, vol. 1423, no. 1, pp. 176–187, Jul. 2018,
doi: 10.1111/nyas.13610.
[60] A. M. -Rioja, “‘We are the dead’: the war poets, metal music and chaos control,” Metal Music Studies, vol. 7, no. 2, pp. 299–315,
Jun. 2021, doi: 10.1386/mms_00050_1.
[61] A. U. Igamberdiev, “Reflexive structure of the conscious subject and the origin of language codes,” BioSystems, vol. 231, Sep.
2023, doi: 10.1016/j.biosystems.2023.104983.
[62] P. Cohen and G. Nitschke, “Evolving music with emotional feedback,” in GECCO 2019 Companion - Proceedings of the 2019
Genetic and Evolutionary Computation Conference Companion, Jul. 2019, pp. 135–136, doi: 10.1145/3319619.3321883.
[63] S. Sun, “A college music teaching system designed based on android platform,” Scientific Programming, vol. 2021, pp. 1–16,
Sep. 2021, doi: 10.1155/2021/7460924.
[64] N. Chen and G. Wen, “Music composition feasibility using a quality classification model based on artificial intelligence,”
Aggression and Violent Behavior, Jun. 2021, doi: 10.1016/j.avb.2021.101632.
[65] A. S. Ryzhinsky, “Choral music in the works of hans werner henze,” Music Scholarship, no. 1, pp. 46–60, 2021, doi:
10.33779/2587-6341.2021.1.046-060.
[66] A. Ravignani and T. Verhoef, “Which melodic universals emerge from repeated signaling games? a note on lumaca and baggio
(2017),” Artificial Life, vol. 24, no. 2, pp. 149–153, May 2018, doi: 10.1162/artl_a_00259.

 ISSN: 2252-8938
3726
[67] M. A. K. -Papakostas, A. Floros, and M. N. Vrahatis, “Interactive music composition driven by feature evolution,” SpringerPlus,
vol. 5, no. 1, Dec. 2016, doi: 10.1186/s40064-016-2398-8.
[68] D. Seca, J. M. -Moreira, T. M. -Neves, and R. Sousa, “Hierarchical qualitative clustering: clustering mixed datasets with critical
qualitative information,” arXiv-Computer Science, pp. 1-12, 2020.
[69] S. Ianigro and O. Bown, “Exploring transfer functions in evolved CTRNNs for music generation,” in Computational Intelligence
in Music, Sound, Art and Design, vol. 11453, 2019, pp. 234–248, doi: 10.1007/978-3-030-16667-0_16.
[70] G. Zouridakis, P. G. Simos, and A. C. Papanicolaou, “Multiple bilaterally asymmetric cortical sources account for the auditory
N1m component,” Brain Topography, vol. 10, no. 3, pp. 183–189, 1998, doi: 10.1023/A:1022246825461.
[71] H. Rafraf, “Differential music: automated music generation using LSTM networks with representation based on melodic and
harmonic intervals,” arXiv-Computer Science, pp. 1-14, 2021, doi: 10.48550/arXiv.2108.10449.
[72] N. Nedjah, H. D. Bezerra, and L. M. Mourelle, “Automatic generation of harmonious music using cellular automata based
hardware design,” Integration, vol. 62, pp. 205–223, Jun. 2018, doi: 10.1016/j.vlsi.2018.03.002.
[73] S. A. -Pizarro and N. Verelst, “Análisis bibliométrico sobre la calidad de la educación superior en Chile,” Educación, vol. 32, no.
62, pp. 5–32, May 2023, doi: 10.18800/educacion.202301.010.
[74] L. Benomar et al., “Bibliometric analysis of the structure and evolution of research on assisted migration,” Current Forestry
Reports, vol. 8, no. 2, pp. 199–213, Jun. 2022, doi: 10.1007/s40725-022-00165-y.
[75] Y. Okubo, “Bibliometric indicators and analysis of research systems methods and examples,” OECD Science, Technology and
Industry Working Papers, 1997, doi: 10.1787/208277770603.
BIOGRAPHIES OF AUTHORS
Frank Manuel Gutiérrez Paitan is a highly motivated, dedicated professional
with a solid software development foundation. He is a software development technician from
S.E.N.A.T.I. He is furthering his education as a Systems Engineering student at the
Technological University of Peru, enhancing his skills and knowledge in the constantly
evolving field of technology. His passion is programming, and he has successfully assumed
the dynamic role of a full-stack developer at several companies. Specializing in Angular+ for
front-end development has been a critical aspect of his experience, and he is currently honing
his backend skills with a focus on NestJS. Eager to expand his knowledge, he is exploring
areas such as Docker and A.W.S., envisioning a comprehensive skill set that spans the entire
software development lifecycle. His background reflects a commitment to continuous learning
and a drive to stay at the forefront of technological advancements, ensuring that he contribute
meaningfully to the projects and teams he is involved with. He can be contacted at email:
u20239818@utp.edu.pe.
María Acuña Meléndez is professor at the Faculty of Engineering of
Engineering of the Technological University of Peru, Lima-Peru. She holds a Ph.D. in
Systems Engineering. Her research areas are information systems and communications, as
well as systems auditing and information security. She has participated in several research
projects, as well as in thesis advising. She has registered several patents on software
copyrights. Her research interests include information systems, data science, information
security, data mining, artificial intelligence, and knowledge management among other related
lines of research. She can be contacted at email: c21584@utp.edu.pe.
Christian Ovalle is an associate professor at the Faculty of Engineering of the
Technological University of Peru, Lima-Peru. He has a Ph.D. in Systems Engineering with a
specialization in artificial intelligence. His research areas are process mining, business data
analysis, and pattern recognition. He is the CEO of the 7D consultancy dedicated to the
Investigation of intelligent solutions. He has participated in different research projects,
receiving awards from the Peruvian Ministry of Defense and the Armed Forces Army for the
best general researcher, which is a technology-based company and his innovative products
received national and international recognition. He has filed a number of patents and industrial
designs on his innovative ideas. His research interests include data mining, artificial
intelligence, image/signal processing, bibliometrics, and pattern recognition. He can be
contacted at email: dovalle@utp.edu.pe.

Application of artificial intelligence in music generation: a systematic review

More Related Content

Similar to Application of artificial intelligence in music generation: a systematic review (20)

More from IAESIJAI (20)

Recently uploaded (20)

Application of artificial intelligence in music generation: a systematic review