Implementing Deep Learning Approaches for Students with

Special Needs: A Systematic Literature Review

Marlina

, Endang Sri Handayani

, Syari Yuliana

, Yosa Yulia Nasri

, Rindia Nengsih

Selvi Rahmawati

, Nurmalika Ulfa

, Elma Diana

and Al Shaffaat Ronvy

Department of Special Education, Universitas Negeri Padang, Indonesia

Department of Special Education, Universitas Adzkia Padang, Indonesia

Keywords: Implementation, Deep Learning, Students with Special Needs, Systematic Literature Review, Inclusive

Pedagogy.

Abstract: The purpose, this systematic literature review (SLR) is to explore how deep learning approaches have been

implemented to support students with special educational needs (SEN). Method, following the PRISMA 2020

protocol, a comprehensive search was conducted across major academic databases including Scopus, Web of

Science, and ERIC. The selection process resulted in 56 peer-reviewed articles published between 2015 and

2025 that met the inclusion criteria. The data were analysed thematically to identify implementation strategies,

common practices, benefits, challenges, and contextual factors influencing deep learning in inclusive

classrooms. The review reveals that deep learning in inclusive education for students with SEN primarily

involves strategies such as project-based learning, collaborative inquiry, and metacognitive scaffolding.

Teachers play a pivotal role in facilitating deep learning through adaptive instruction and emotional support.

However, challenges such as limited teacher training, rigid curricula, and inadequate school support systems

persist. Studies also highlight the significance of school culture and curriculum flexibility in sustaining deep

learning practices. Despite positive impacts on student engagement, critical thinking, and social-emotional

growth, the integration of deep learning remains uneven across contexts. Conclusion: Deep learning holds

promising potential for enhancing educational experiences and outcomes for students with special needs in

inclusive settings. However, successful implementation requires systemic alignment between teacher

competencies, curriculum design, and institutional support.

1 INTRODUCTION

The transformation of 21st-century education

emphasises learning that focuses not only on

academic achievement but also on developing higher-

order thinking skills, creativity, collaboration, and

problem-solving. The deep learning approach

responds to this need, emphasising conceptual

understanding, critical reflection, the

interconnectedness of concepts, and the application

of knowledge in real-world contexts (Zebua, 2025).

In practice, deep learning positions students as active

learners who construct knowledge through

meaningful experiences, rather than simply

memorising information (MacFarlane et al., 2017).

https://orcid.org/0000-0003-3265-8045

In Indonesia, the direction of educational policy

that upholds similar values is reflected in the

Independent Curriculum. This curriculum

emphasises differentiated, project-based learning and

focuses on strengthening competencies and character

(Marlina, 2019). Students are encouraged to have

agency (control over their learning process), and

teachers act as facilitators, adapting approaches to

individual learners' needs. This is particularly

relevant for students with special needs, who require

a flexible, meaningful learning approach that is

oriented toward their unique potential (Jauhari &

Idhartono, 2022).

Both deep learning and the Independent

Curriculum share a common thread in promoting

Marlina, , Handayani, E. S., Yuliana, S., Nasri, Y. Y., Nengsih, R., Rahmawati, S., Ulfa, N., Diana, E. and Ronvy, A. S.

Implementing Deep Learning Approaches for Students with Special Needs: A Systematic Literature Review.

DOI: 10.5220/0014069200004935

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 7th International Conference on Early Childhood Education (ICECE 2025) - Meaningful, Mindful, and Joyful Learning in Early Childhood Education, pages 81-90

ISBN: 978-989-758-788-7; ISSN: 3051-7702

transformative, participatory, and student-centered

learning. However, to date, few studies have

specifically evaluated how deep learning principles

are implemented in the context of special education

in Indonesia, particularly within the framework of the

Independent Curriculum. Empirical evidence remains

limited regarding the effectiveness of deep learning

strategies in improving engagement, conceptual

understanding, and the social-emotional development

of students with special needs, both in inclusive and

special schools (Andayanie, et al, 2025).

One approach that is increasingly being studied

in 21st-century education is deep learning. Unlike

surface learning, which focuses solely on

memorization and repetition of information, deep

learning encourages students to think critically, solve

problems, understand the meaning of concepts, and

connect new information to existing knowledge. This

approach is capable of improving the quality of the

teaching and learning process. It is relevant to the

demands of 21st-century competencies such as

collaboration, communication, creativity, and critical

thinking (Thornhill-Miller et al., 2023).

In the context of students with special needs

studying in inclusive schools, implementing

immersive learning presents unique challenges. Their

limitations in cognitive, social, and adaptive abilities

require teachers to adapt learning strategies to ensure

they remain meaningful and accessible to all students

(Darwish et al., 2025). Meanwhile, literature that

discusses the practical implementation of immersive

learning for students with special needs in inclusive

classrooms remains limited and scattered.

Therefore, this study aims to conduct a

Systematic Literature Review (SLR) to collect,

analyse, and synthesise relevant research findings

related to the implementation of immersive learning

for students with special needs in inclusive schools.

This study uses the PRISMA 2020 guidelines for

literature identification and analysis, as well as the

PICo framework for determining the focus

population, interventions, and context. The results of

this systematic review are expected to provide a more

comprehensive understanding, identify good

practices, and uncover challenges and

recommendations in implementing inclusive and

adaptive immersive learning for students with special

needs. In this systematic review, the research

questions formulated are as follows:

1. How is deep learning implemented in inclusive

schools to support students with special

educational needs (SEN)?

2. What types of deep learning strategies are most

commonly used for students with special

educational needs (SEN)?

3. What benefits and challenges are reported in the

implementation of deep learning approaches for

SEN students?

4. What contextual factors (e.g., teacher roles,

curriculum design, school environment) influence

the success of deep learning implementation in

inclusive classrooms?

5. To what extent do existing studies integrate

inclusive principles into deep learning practices

for students with special needs?

2 METHOD

2.1 Research Design and Data Sources

This study uses a systematic literature review

approach to examine the implementation of deep

learning in education for students with special needs.

The goal is to systematically identify, evaluate, and

synthesise relevant literature to obtain a

comprehensive overview of the practices, challenges,

and effectiveness of its implementation. The review

process follows the PRISMA (Preferred Reporting

Items for Systematic Reviews and Meta-Analyses)

guidelines developed by Page et al. (2021) to ensure

transparency, objectivity, and rigour in literature

selection. The PRISMA approach includes three main

stages: identification, screening (including eligibility

assessment), and inclusion. Each stage is carefully

conducted to ensure that the analysed sources are

valid, relevant, and contribute to the development of

deep learning strategies that are appropriate to the

needs of students with special needs. This process

flow is depicted in the PRISMA diagram in Figure 1.

2.2 Databases

A literature search was conducted using the academic

database Scopus.com, as it aggregates articles from

various scientific journals, offering broad and diverse

coverage. This database selection ensured access to

multidisciplinary research, allowing for a

comprehensive overview of the topics explored in this

study. This timeframe ensured the inclusion of the

most recent and relevant studies.

A structured literature search was conducted

using the Scopus.com database, using targeted search

terms. A series of structured search terms was used to

capture a comprehensive range of studies related to

immersive learning in Special Education, including

ICECE 2025 - The International Conference on Early Childhood Education

forms, types, benefits, challenges, and factors

influencing the implementation of immersive

learning for students with special needs.

The search queries included:

1. [“Deep Learning”] OR [“Immersive Learning]

2. [“Students with Intelectual disabilities”] OR

[“Mentally Retarded”]

3. [“Meaningful Learning”] OR [“Joyful

Learning”] OR [“Mindful Learning”]

2.3 Inclusion and Exclusion Criteria

Data

This study establishes inclusion and exclusion criteria

to ensure that the study conducted is relevant and

academically credible. Exclusion criteria include: (1)

Document type: book, book chapter, conference

paper, conference review, (2) Language: other than

English, (3) Publication year: other than 2015-July

2025. At the same time, inclusion criteria include: (1)

Document type: article, (2) Language: written in

English, (3) Source type: only articles, (4) Article

publication year from 2015 to 2025.

Figure 1: PRISMA Flowchart.

2.4 Data Extraction

Information collected from each article: (1) author

name and year, (2) country of origin of the study, (3)

research objectives, (4) participant characteristics

(age, stage, diagnosis), (5) methodology, (6) deep

learning strategy used, and (7) main findings.

2.5

Documents Included in the Analysis

After screening and eligibility, 56 documents meeting

the inclusion criteria were selected for further

analysis. Figure 1 shows the selection process. The

datasets were combined into CSV files for

bibliometric analysis using Rstudio and

Bibliometrics. This approach combines empirical and

systematic literature, providing information on the

implementation of immersive learning for students

with special needs.

2.6 Quality and Credibility

Data collection in this study was conducted

rigorously to ensure only high-quality articles were

analyzed. Each metadata entry was carefully

reviewed to ensure the accuracy and completeness of

information, including title, author(s), year of

publication, keywords, and number of citations. This

verification process ensures the validity and

reliability of the bibliometric analysis results. Table 1

shows that most metadata elements, such as abstract,

author(s), DOI, document type, journal, language,

year of publication, title, total citations, references,

affiliations, keywords, corresponding author(s),

additional keywords, and scientific category, were

fully available. This completeness provides a strong

foundation for bibliometric analysis, as it allows for

accurate mapping of citations and the influence of

authors or journals. However, deficiencies in some

metadata elements can still impact the overall

completeness of the analysis results.

The keyword data completion rate (DE) was

only 8.93% and publication year (PY) data was

recorded at 0%, indicating data gaps, albeit minimal.

The absence of keyword data can reduce the accuracy

of thematic and co-occurrence analyses, which are

crucial for identifying research trends. Meanwhile,

the completeness of corresponding author (RP) data

was only 7.14%, potentially limiting the analysis of

inter-researcher collaboration networks and mapping

institutional relationships in deep learning

implementation studies.

In contrast, several metadata elements exhibit

significant deficiencies, such as keywords plus (ID)

with 39.29% missing data, cited references (CR) with

1.79% missing data, and science categories (WC)

with 100% missing data. These deficiencies are

critical because they significantly limit citation

network analysis and thematic exploration, thus

hindering the identification of key themes and

broader scientific relationships within the literature.

The absence of data on ID and WC particularly

Implementing Deep Learning Approaches for Students with Special Needs: A Systematic Literature Review

reduces the ability to map interrelated concepts and

research dynamics in this field.

Table 1: Completeness of Deep Learning Bibliographic

Metadata.

Although some metadata elements, such as

keywords plus (ID) and scientific categories (WC),

experienced data loss, this deficiency could be

addressed by utilizing the author keywords (DE)

available in the metadata. Co-occurrence analysis

could still be conducted with limited scope to map

key themes in research on implementing deep

learning for students with special needs. In the

context of this research, mapping trends, citation

patterns, and researcher collaborations was crucial.

Therefore, while the absence of WC and ID limited

certain aspects of the analysis, the availability of other

metadata still allowed the research to proceed without

compromising the integrity of the analysis (Aria &

Cuccurullo, 2017)

2.7 Data Analysis

Data were analyzed using thematic analysis with a

narrative synthesis approach. This was because the

data obtained from the included studies were

qualitative and quantitative descriptive,

demonstrating diversity in methodology, context, and

forms of in-depth learning interventions for students

with special needs.

3 RESULT

3.1 Demographic Distribution Results

This study analyzed 56 articles on the implementation

of immersive learning for students with special needs

published between 2015 and 2025. The publication

growth rate reached 7.18% annually, reflecting the

increasing attention to this topic. A total of 225

authors were involved, with two articles written

individually. Each article had an average of four

authors, indicating a strong collaborative nature.

International collaboration was reflected in 28.57%

of articles involving authors from different countries.

Keyword analysis identified 211 unique terms,

indicating broad thematic and conceptual diversity.

The average age of the documents was 3.27 years,

indicating the novelty of the reviewed literature. An

average of 26.05 citations per article indicates

significant academic impact. Figure 2 visually

presents demographic data and publication

characteristics.

Figure 2:

Demographic

Distribution Results of Deep

Learning Data.

3.2

Annual Publication Trends

Figure 3 shows a slow growth in publications related

to the implementation of deep learning for students

with special needs from 2015-2025, with each article

published. In 2015-2017, there was an increase in

publications of 3 articles, and in 2018, there was a

decrease in publications of 0 articles. In 2019-2021,

there was an increase in publications of 21 articles. In

2022, there was a decrease in publications of 8

articles. In 2023, there was an increase in publications

of 13 articles. In 2024-2025, there was a decreased in

publications of 11 articles.

Figure 3: Annual Publication Trends in Deep Learning.

ICECE 2025 - The International Conference on Early Childhood Education

3.3

Country Level Analysis

The USA leads research publications on

implementing immersive learning for students with

special needs, with 26 papers published between

2015-2025.

Figure 4: Country-Level Analysis of Deep Learning

Figure 4 provides information on the number of

citations and average citations of articles by country.

The USA published 95 articles. Other countries, such

as China, published 41 articles, Australia published

60, Spain published 50, and the Netherlands

published 58.

Table 2: Most Cited Countries.

3.4 Key Contributors and Influential

Institutions

3.4.1 Most Relevant Authors

This study identified several academics who

consistently published articles on immersive learning

and students with special needs between 2015 and

2025. Table 3 presents the 10 most relevant authors

and their number of publications. The most prominent

authors on this topic are Alkhurayyif, Yazeed.

Table 3: Most Relevant Authors in the Research.

3.4.2 Author Productivity

Most authors (99.6%) published only one article,

while only 0.4% contributed more than one

publication related to implementing deep learning for

students with special needs.

Table 4: Author Productivity.

3.4.3 Author Impact: h-index, g-index, and

m-index

In addition to assessing the number of publications,

researchers evaluated the impact of each author's h-

index, g-index, and m-index. These indices help

measure an author's contributions' overall quality and

impact. Table 2 shows that the authors' h-index

averaged 1.

Tabel 5: Impact of Author Indexing.

3.4.4 The Most Influential Institutions

Some affiliates that emerged as the main actors in the

research on the implementation of in-depth learning

for students with special needs from 2015 to 2025, the

data shows that Universidad De Deusto, with eight

articles, as depicted in Table 6.

Implementing Deep Learning Approaches for Students with Special Needs: A Systematic Literature Review

Table 6: Most Relevant Affiliates.

4 DISCUSSION

4.1 Deep Learning in an Inclusive

Context

Deep learning is an instructional approach that

encourages students to think critically through

analyzing, synthesizing, and evaluating information,

rather than simply memorizing it. In inclusive

education, this approach is tailored to meet the needs

of students with intellectual disabilities, specific

learning disabilities, and communication and social

disabilities. This approach aligns with Universal

Design for Learning (UDL) principles, which

emphasize diversity in representation, expression,

and engagement in learning (CAST, 2018).

Most studies indicate that deep learning is

implemented through project-based learning,

problem-based learning, and collaborative activities.

These strategies are designed to engage students in

meaningful activities that connect knowledge to real-

world experiences. Some of the methods identified

include:

a. Scaffolded inquiry with visual and concrete

supports.

b. Story-based learning to enhance understanding of

abstract concepts.

c. Peer tutoring and cooperative learning that

encourage social interaction.

4.2 Deep Learning Implementation

Strategies for Students with Special

Needs

a. Project-Based Learning (PjBL) and Problem-

Based Learning (PBL)

This strategy is widely adopted because it

facilitates collaboration, problem-solving, and

reflection, which are the core of deep learning. For

example, students with special needs are

encouraged to complete real-life projects with

visual support, scaffolding, and adaptive

technology (Schuelka et al., 2019); (Prystiananta

& Noviyanti, 2025). A study (Fernández-

Batanero et al., 2022) showed that multimedia-

based PjBL is effective in developing critical

thinking skills in students with learning

disabilities.

b. Adaptive Learning Technology

This technology automatically detects student

needs and adapts learning content in real time.

Students with special needs can access audio,

visual, or interactive materials. The AI system can

adjust the difficulty level based on student

responses, including for those with mild

intellectual disabilities (Togni, 2025).

c. Collaborative Peer Mentoring and Guided

Discussions

Collaborative activities allow students with

special needs to interact with peers through

guidance and structured discussions, with the

support of teachers as facilitators (Marlina et al.,

2023). Dialogic learning in deep learning is

crucial, especially for students with

communication barriers (Luckin & Holmes,

2016).

d. The Role of Teachers and Inclusive Learning

Design

Successful implementation depends heavily on

teachers' ability to design adaptive and reflective

learning. Teachers need to use ongoing formative

assessment, design reflective activities, and

provide meaningful feedback. An inclusive

approach based on deep learning requires detailed

instructional planning and teacher training in

reflective and collaborative learning strategies

(Mitchell & Sutherland, 2020).

Table 7: Summary of Deep Learning Strategies.

4.3 Benefits and Challenges of

Implementing a Deep Learning

Approach for Students with Special

Needs

Deep learning encourages students to go beyond

memorizing information to understanding,

connecting, and applying knowledge. In students with

special needs, this approach has been shown to

improve their ability to solve simple problems,

understand cause-and-effect relationships, and

ICECE 2025 - The International Conference on Early Childhood Education

communicate thoughts verbally or visually. Project-

based approaches have been shown to strengthen

memory and conceptual connections (Nguyen &

Nguyen, 2021; Lee & Kim, 2019).

Through project-based and cooperative learning

methods, students actively engage with their peers,

developing social and collaborative skills. Reported

social benefits include increased empathy and

tolerance from classmates and a sense of belonging

for students with disabilities, who feel valued through

their roles within the group (CAST, 2018).

Furthermore, several studies have shown that deep

learning encourages students to express themselves

through portfolios, journals, or group discussions,

which builds self-awareness and responsibility for the

learning process (Lim & Tan, 2021).

Despite its benefits, implementing deep learning

in the context of students with special needs is not

without challenges. Four main challenges reported

are: (1) limited teacher capacity, (2) minimal

curriculum and policy support, (3) limited resources

and technology, and (4) unequal participation in

group activities.

Many teachers have not received adequate

training in implementing deep learning strategies,

particularly in differentiating learning for students

with special needs (Marlina, 2021). Teachers often

struggle to design assignments that are appropriate to

students' cognitive levels and face increased

workloads when preparing adapted materials. As

many as 68% of teachers feel a lack of confidence in

implementing project-based learning for students

with special needs (Yusuf et al., 2022). In developing

countries, including Indonesia, the national

curriculum does not fully support the flexibility

needed to implement deep learning in inclusive

classrooms. As a result, teachers focus more on

achieving formal academic targets than on

developing deep thinking processes. Furthermore,

there is no evaluation system that assesses higher-

order thinking skills in students with disabilities. In

practice, teachers often adapt deep learning strategies

informally without the support of clear curriculum

guidelines (Lee & Kim, 2019).

Implementing deep learning relies heavily on

media, visual aids, technology, and sufficient time.

Many inclusive schools face limitations: (1) access to

interactive digital devices, (2) appropriate visual

teaching materials, and (3) high teacher-student ratios

(García-Robles et al., 2024). Despite collaborative

efforts, some students with special needs are excluded

from group discussions, especially with no clear

roles. The impact is increased reliance on peers and

shallow participation without cognitive engagement

(Rahman et al., 2023).

4.4 Factors Influencing the Success of

Deep Learning Implementation for

Students with Special Needs

The successful implementation of a deep learning

approach in inclusive classrooms for students with

special needs is influenced by various factors,

including teacher competency and perception,

curriculum design and differentiation, a supportive

inclusive learning environment, and the availability

of resources and technology (Marlina, 2015). Studies

show that success is highly dependent on teachers'

abilities to: (1) design structured and adapted deep

learning, (2) use approaches such as project-based

learning, problem-based learning, and cooperative

learning, and (3) manage heterogeneous classes based

on needs. Teachers who have received training in

inclusive pedagogy and higher-order thinking

strategies are more successful in effectively

implementing the deep learning approach (Yusuf et

al., 2022). Teachers who have positive beliefs about

the abilities of students with special needs are more

likely to provide challenging yet accessible learning

experiences (Lee & Kim, 2019).

Flexible modification of learning objectives,

materials, and methods encourages the

implementation of deep learning strategies. Good

practices found include the use of UDL and the

adoption of multi-tiered instruction based on student

abilities. Multi-level planning within a single learning

theme is crucial. Implementing deep learning relies

on teachers' ability to differentiate content, processes,

and products according to student profiles (Liang et

al., 2024). Critical thinking task designs should be

developed in multiple difficulty levels to ensure all

students are meaningfully engaged (CAST, 2018).

Schools that instill values of diversity, respect for

differences, and encourage collaboration among

students be more effective in implementing deep

learning strategies (Lim & Tan, 2021). Support from

school leaders and a collaborative work culture

among teachers directly impact pedagogical

innovation in inclusive classrooms, enabling positive

interactions between students to become a catalyst for

the active participation of students with special needs

in complex tasks (Rahman et al., 2023). Access to

visual aids, interactive technologies (such as

AR/VR), and image-based learning materials

significantly impact student engagement in deep

learning. The use of VR-based simulations improves

the understanding of abstract concepts (such as

Implementing Deep Learning Approaches for Students with Special Needs: A Systematic Literature Review

gravity and weather) for students with special needs

(García-Robles et al., 2024); (Tsai et al., 2020).

Table 8: Factors Determining the Success of Deep Learning.

4.5 To What Extent Do Existing

Studies Address the Role of

Teachers, Curriculum, and School

Environment in Facilitating Deep

Learning for Students with Special

Educational Needs?

The main findings indicate that most studies

explicitly or implicitly acknowledge that the success

of deep learning implementation is heavily influenced

by three components of the education system:

teachers as the primary facilitators, the curriculum as

the pedagogical framework, and the school

environment as the context for its implementation.

However, the depth and focus of discussion on these

three aspects vary across studies.

1. The Role of Teachers as Deep Learning

Facilitators

Teachers are the primary actors in deep learning

strategies for students with special needs.

Adaptations are made through project-based

assignment modifications, visuals and concrete

tools, group-based learning, and peer tutoring

(Marlina, 2014). Creative teachers can

transform problem-based learning into concrete

and collaborative experiences for students with

special needs (Lee & Kim, 2019); (Nguyen &

Nguyen, 2021).

The teacher's role as a scaffolding provider is

crucial in helping students think reflectively,

make connections, and develop deep

understanding. Teachers' use of open-ended

questions, step-by-step reflection, and verbal

feedback can enhance students with special

needs' ability to explore ideas (Lee & Kim,

2019). Studies also note that the success of deep

learning is closely related to teachers' beliefs in

the learning potential of students with special

needs (Yusuf et al., 2022).

2. The Role of the Curriculum as a Driving

Framework

A curriculum implementing the UDL approach

allows for various forms of representation and

expression of learning (Lim & Tan, 2021); an

open thematic curriculum provides space for

teachers to develop collaborative and reflective

learning (Zahra et al., n.d.). Conversely, several

studies highlight that the national curriculum's

overly academic and linear nature makes it

difficult to implement deep learning, especially

in inclusive classrooms. The absence of specific

indicators for critical thinking or reflection skills

in the curriculum makes deep learning a low

priority (Tsai et al., 2020).

3. The Role of the School Environment in

Supporting Deep Learning

Schools that embrace a collaborative culture,

value diversity, and allow for pedagogical

innovation can provide a learning environment

conducive to deep thinking. School

environments open to innovative technologies

(e.g., VR and interactive simulations) enable

implementing more inclusive deep learning

strategies (García‐Robles et al., 2024).

Principals and other educational leaders also

significantly encourage innovation, training,

and resources (Rahman et al., 2023)

5 CONCLUSION

An analysis of 56 articles shows that deep learning

approaches significantly positively impact students

with special needs, particularly in improving

cognitive engagement, social interaction, contextual

understanding, and higher-order thinking skills. The

most commonly used strategies include project-based

learning (PjBL), problem-based learning (PBL),

cooperative learning, peer tutoring, and multimedia

and AR/VR technologies. These approaches

effectively bridge cognitive limitations through

concrete, collaborative, and real-life learning

experiences.

The success of implementation is heavily

influenced by four main factors: (1) teacher

competency and perception in designing in-depth

learning, (2) curriculum flexibility that supports

differentiation, (3) an inclusive learning environment

that encourages collaboration, and (4) the availability

of appropriate resources and technology. Despite the

strong benefits, challenges remain, such as limited

teacher training, an unadaptive curriculum, and

limited access to technology.

Overall, the analyzed studies underscore the

importance of integrating deep learning into inclusive

education, supported by teacher training, an inclusive

curriculum, and collaborative learning environments.

Systemic efforts are needed to improve teacher

ICECE 2025 - The International Conference on Early Childhood Education

capacity, expand access to technology, and promote

flexible education policies to enable students with

intellectual disabilities to experience meaningful and

immersive learning.

REFERENCES

Andayanie, L. M., Adhantoro, M. S., Purnomo, E., &

Kurniaji, G. T. (2025). Implementation of Deep

Learning in Education: Towards Mindful, Meaningful,

and Joyful Learning Experiences. Journal of Deep

Learning, 47-56.

Aria, M., & Cuccurullo, C. (2017). bibliometrix: An R-tool

for comprehensive science mapping analysis. Journal

of Informetrics, 11(4), 959–975.

CAST, C. (2018). Universal design for learning guidelines

version 2.2. Http://Udlguidelines. Cast. Org.

Darwish, S., Alodat, A., Al-Hendawi, M., & Ianniello, A.

(2025). General Education Teachers’ Perspectives on

Challenges to the Inclusion of Students with Intellectual

Disabilities in Qatar. Education Sciences, 15(7), 908.

Fernández-Batanero, J. M., Montenegro-Rueda, M.,

Fernández-Cerero, J., & García-Martínez, I. (2022).

Assistive technology for the inclusion of students with

disabilities: A systematic review. Educational

Technology Research and Development, 70(5), 1911–

1930.

García‐Robles, P., Cortés‐Pérez, I., Nieto‐Escámez, F. A.,

García‐López, H., Obrero‐Gaitán, E., & Osuna‐Pérez,

M. C. (2024). Immersive virtual reality and augmented

reality in anatomy education: a systematic review and

Meta‐Analysis. Anatomical Sciences Education, 17(3),

514–528.

Jauhari, M. N., & Idhartono, A. R. (2022). Pengoptimalan

Aksesibilitas Fisik Bagi Siswa Berkebutuhan Khusus

Disekolah Inklusi. Kanigara, 2(1), 264-269.

Lee, J.-A., & Kim, C.-J. (2019). Teaching and learning

science in authoritative classrooms: Teachers’ power

and students’ approval in Korean elementary

classrooms. Research in Science Education, 49(5),

1367–1393.

Liang, W., To, J., & Lo, Y. Y. (2024). A journey towards

teacher empowerment in differentiated instruction:

implications for a sustainable teacher professional

development model. Asia Pacific Education Review, 1–

13.

Lim, F. V., & Tan, J. M. (2021). Curriculum and assessment

mismatch: Examining the role of images in literacy

assessments. The Australian Journal of Language and

Literacy, 44(3), 22–34.

Luckin, R., & Holmes, W. (2016). Intelligence unleashed:

An argument for AI in education.

MacFarlane, J., Aladjem, D., & Russel, C. A. (2017).

Instructional practices for deeper learning: Lessons for

educators. Policy Studies Associates. https://asiasocie

ty.org/Files/Instructional-Practices-for-Deeper-Learn

ing-Lessons-for-Educators-May-2017-Report. Pdf.

Marlina. (2014). Kelompok Teman Sebaya, Popularitas,

dan Preferensi Sosial Anak Berkebutuhan Khusus

Belajar Di Sekolah Dasar Inklusi. Jurnal Pendidikan

Luar Biasa FIP Universitas Negeri Padang, 1(1), 6–14.

Marlina, M. (2015). Asesmen Anak Berkebutuhan Khusus:

Pendekatan Psikoedukasional Edisi Revisi.

Marlina, M. (2019). Panduan Pelaksanaan Model

Pembelajaran Berdiferensiasi di Sekolah Inklusif.

Universitas Negeri Padang. http://repository

.unp.ac.id/id/eprint/23547

Marlina, M. (2021). Strategi Pembelajaran Berdiferensiasi

di Sekolah Inklusif. Afifa Utama. http://

repository.unp.ac.id/32203/1/Marlina_2020_Buku_Str

ategi_Pembelajaran_Berdiferensiasi_di_Sekolah_Inkl

usif_ok.pdf

Marlina Marlina; Rahmahtrisilvia; Grahita Kusumastuti;

Pujaningsih. (2023). Peer Mentorship Mahasiswa

Disabilitas di Perguruan Tinggi. Afifa Utama.

Mitchell, D., & Sutherland, D. (2020). What really works in

special and inclusive education: Using evidence-based

teaching strategies. Routledge.

Nguyen, L. T., & Nguyen, D. Q. (2021). PhoNLP: A joint

multi-task learning model for Vietnamese part-of-

speech tagging, named entity recognition and

dependency parsing. ArXiv Preprint ArXiv:21

01.01476.

Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I.,

Hoffmann, T. C., Mulrow, C. D., Shamseer, L.,

Tetzlaff, J. M., Akl, E. A., & Brennan, S. E. (2021). The

PRISMA 2020 statement: an updated guideline for

reporting systematic reviews. Bmj, 372.

Prystiananta, N. C., & Noviyanti, A. I. (2025). Enhancing

EFL Instruction in Special Needs Education:

Integrating Multimodal Digital Tools and Deep

Learning Strategies. Voices of English Language

Education Society, 9(1), 96–108.

Rahman, A. A., Santosa, T. A., Nurtamam, M. E., Widoyo,

H., & Rahman, A. (2023). Meta-Analysis: The Effect

of Ethnoscience-Based Project Based Learning Model

on Students’ Critical Thinking Skills. Jurnal Penelitian

Pendidikan IPA, 9(9), 611–620.

Schuelka, M. J., Artiles, A. J., Thomas, G., & Johnstone, C.

J. (2019). The SAGE handbook of inclusion and

diversity in education.

Thornhill-Miller, B., Camarda, A., Mercier, M., Burkhardt,

J.-M., Morisseau, T., Bourgeois-Bougrine, S., Vinchon,

F., El Hayek, S., Augereau-Landais, M., & Mourey, F.

(2023). Creativity, critical thinking, communication,

and collaboration: Assessment, certification, and

promotion of 21st century skills for the future of work

and education. Journal of Intelligence, 11(3), 54.

Togni, J. (2025). Development of an Inclusive Educational

Platform Using Open Technologies and Machine

Learning: A Case Study on Accessibility Enhancement.

ArXiv Preprint ArXiv:2503.15501.

Tsai, S.-C., Chen, C.-H., Shiao, Y.-T., Ciou, J.-S., & Wu,

T.-N. (2020). Precision education with statistical

learning and deep learning: a case study in Taiwan.

International Journal of Educational Technology in

Higher Education, 17(1), 12.

Implementing Deep Learning Approaches for Students with Special Needs: A Systematic Literature Review

Yusuf, M. O., Ansah, S. D., Ahmed, T. F., & Yusuf, H. T.

(2022). Professional development in technology

integration among teacher educators in Ghana.

Indonesian Journal of Educational Research and

Review, 5(1), 88–99.

Zahra, A., Soumia, T., & Mohamed, R. (n.d.). The impact

of continuous teacher training based on the flipped

classroom on teaching practices and learner

performance. Int J Eval & Res Educ ISSN, 2252(8822),

8822.

Zebua, N. (2025). Education Transformation: Implemen-

tation of Deep Learning in 21st-Century Learning.

Harmoni Pendidikan: Jurnal Ilmu Pendidikan, 2(2),

146–152.

ICECE 2025 - The International Conference on Early Childhood Education