Academic Journal of Science and Technology ISSN: 2771-3032 | Vol. 12, No. 1, 2024 145 Strengthen the Foundation and Practice Orientation: Exploration of Teaching Reform in Machine Learning Courses Yan Wang * Department of Computer, North China Electric Power University, Huadian Road, Baoding, China * Corresponding author Email: wangyan1206@126.com Abstract: Machine learning has become a core technology in the field of artificial intelligence with its excellent performance and wide range of applications. Traditional machine learning course teaching models often focus on theoretical explanations and lack sufficient practical guidance, making it difficult for students to apply their knowledge to solve practical problems. This article explores feasible ways to reform the teaching of machine learning courses, by strengthening the close integration of basic knowledge and practical skills, as well as optimizing course evaluation methods, significantly improving students' learning effectiveness and practical application abilities. Keywords: Machine Learning; Practice; Teaching Reform. 1. Introduction In the current rapidly developing era, machine learning technology has become the core technology in the field of artificial intelligence due to its excellent performance and wide application areas. Currently, the challenges and problems in teaching machine learning courses [1] are gradually becoming prominent and urgently require attention. On the one hand, with the rapid progress of technology, the knowledge system in the field of machine learning continues to update and iterate. However, traditional teaching content and methods often find it difficult to keep up with this rapid development. According to relevant statistical data, up to 30% of the teaching content in current machine learning courses has become outdated and cannot fully meet the actual needs of industry development. This leads to students often finding a significant disconnect between their knowledge and actual job demands after completing their studies, which in turn affects their competitiveness in the job market. Therefore, we must face this problem head-on and actively seek solutions to better adapt to the needs of industry development, enhance students' practical application abilities and employment competitiveness. On the other hand, the practical aspects in current machine learning course teaching are generally weak. In classroom teaching, there is a general emphasis on theoretical teaching, but insufficient opportunities for project practice are not provided. This to some extent restricts the improvement of students' practical operational and problem-solving abilities, making it difficult to comprehensively exercise their comprehensive application abilities. In addition, the singularity of teaching methods is also a major issue in the current teaching of machine learning courses. Adopting traditional lecture style teaching, lacking interaction and discussion with students. This teaching method not only fails to stimulate students' interest in learning, but may also lead to a low level of understanding and mastery of the course content. We need to explore more diverse teaching methods to stimulate students' interest and creativity in learning. In summary, the current challenges and problems in teaching machine learning courses are mainly reflected in the lagging teaching content, insufficient practical aspects, and single teaching methods. To address these issues, we need to strengthen teaching reform, strengthen theoretical foundations, enhance practical aspects, innovate teaching methods, and improve the evaluation system. Only in this way can we cultivate more machine learning talents with solid theoretical foundations and strong practical abilities, and promote the sustainable development of the field of machine learning. Figure 1. Course Teaching Reform Framework 2. Theoretical Foundation Enhancement: Building a Solid Machine Learning Knowledge System 2.1. Select Core Course Content and Strengthen the Transmission of Basic Knowledge In the reform of machine learning curriculum teaching, selecting core course content and strengthening the transmission of basic knowledge is a crucial part. To this end, Theoretical Foundation Enhancement Innovative teaching methods Strengthening practical activities Improvement of Evaluation System 146 we have conducted an in-depth analysis of the core knowledge points in the current field of machine learning, and combined with practical application scenarios, selected the most representative and practical content as the teaching focus. For example, in the data preprocessing section, we focused on key technologies such as data cleaning, feature selection, and data transformation, and demonstrated the key role of these technologies in improving model performance through practical cases. At the same time, classic algorithms in machine learning such as decision trees, neural networks, ensemble learning, etc. were introduced, and their principles, application scenarios, advantages and disadvantages were elaborated in detail. By imparting these core contents, students can establish a solid machine learning knowledge system, laying a solid foundation for subsequent practical applications. In order to further strengthen the transmission of basic knowledge, we focus on cultivating students' logical thinking ability and problem-solving ability. During the teaching process, interactive teaching methods such as case analysis and group discussions are used to guide students to deeply reflect on the principles and application scenarios behind machine learning algorithms. At the same time, encourage students to actively participate in practical projects and deepen their understanding of theoretical knowledge through practical operations. These measures not only enhance students' interest and enthusiasm in learning, but also effectively enhance their hands-on ability and problem- solving ability. In addition, we also focus on introducing cutting-edge theories and technologies to broaden students' academic horizons. By focusing on the latest research achievements and development trends in the field of machine learning, we continuously update and optimize course content to ensure that students are exposed to the latest knowledge and technology. At the same time, we also encourage students to actively participate in academic exchanges and discussions, interact and collaborate with peer experts, in order to further enhance their academic literacy and innovation abilities. 2.2. Introducing Cutting-edge Theories to Broaden Students' Academic Horizons Introducing cutting-edge theories is a key step in broadening students' academic horizons in the reform of machine learning curriculum teaching. With the rapid development of technology, new machine learning algorithms and theories emerge one after another, such as deep learning, reinforcement learning, etc. These cutting-edge theories not only bring new breakthroughs to the field of machine learning, but also provide students with broader academic exploration space. By introducing these cutting-edge theories, students can be exposed to the latest research results, understand the development trends in the field of machine learning, and thus stimulate their learning interest and innovative spirit. Taking deep learning as an example, significant achievements have been made in areas such as image recognition and natural language processing in recent years. By introducing theories related to deep learning, students can learn advanced models such as convolutional neural networks and recurrent neural networks, and attempt to apply them to practical problems. This not only helps to enhance students' practical abilities, but also enables them to have a deeper understanding of the essence and potential of machine learning. At the same time, introducing cutting-edge theories can also help students establish interdisciplinary knowledge systems, combine machine learning with other fields, and cultivate talents with innovative spirit and cross-border thinking. In addition, introducing cutting-edge theories can also promote academic exchange between teachers and students. Teachers can guide students to pay attention to cutting-edge issues in the field of machine learning and encourage them to conduct in-depth research by sharing the latest research results and academic trends. Students can engage in academic discussions and practical activities, engage in intellectual collisions and knowledge exchange with teachers and other classmates, and continuously improve their academic literacy and innovation abilities. In summary, introducing cutting-edge theories is an indispensable part of the reform of machine learning curriculum teaching. By introducing the latest research findings and academic trends, students can broaden their academic horizons, stimulate their learning interests and innovative spirit, and promote academic exchange and cooperation between teachers and students. This will help cultivate more talents with innovative spirit and cross-border thinking, injecting new vitality into the development of machine learning. 3. Strengthening Practical Activities: Enhancing Students' Hands-On Ability and Problem-Solving Ability 3.1. Design Diverse Practical Projects to Stimulate Student Interest In the reform of machine learning curriculum teaching, designing diverse practical projects[2,3,4] is crucial for stimulating student interest. We have carefully planned a series of practical projects based on the characteristics of the course and the needs of students, aiming to deepen their theoretical knowledge and improve their problem-solving abilities in practice. For example, we designed a practical project based on image recognition, requiring students to use machine learning algorithms to classify and recognize images. Through practical operation, students not only master the basic principles and algorithms of image recognition, but also learn how to apply these algorithms to solve practical problems. In addition, we have introduced a practical project based on natural language processing, allowing students to use machine learning techniques to process and analyze text data. These practical projects not only enrich the learning experience of students, but also stimulate their interest and enthusiasm for learning. In order to further improve the quality and effectiveness of practical projects, we also focus on introducing practical cases and data analysis models. For example, in the image recognition project, we provided a large number of image datasets and required students to use machine learning algorithms for image preprocessing, feature extraction, and classification recognition. By comparing the performance and effectiveness of different algorithms, students can gain a deeper understanding of the principles and applications of machine learning algorithms. At the same time, students are encouraged to combine practical projects with practical applications, such as participating in scientific research projects, competitions, etc., to further enhance their practical and innovative abilities. 147 By designing diverse practical projects, we have successfully stimulated students' interest in learning, enabling them to master machine learning knowledge and skills in a relaxed and enjoyable atmosphere. At the same time, these practical projects also provide students with a platform to showcase themselves and exercise their abilities, which helps to cultivate their teamwork spirit and innovation ability. In the future, we will continue to explore and practice more effective teaching methods and means, and contribute to the cultivation of more outstanding machine learning talents. 3.2. Strengthen Experimental Procedures and Cultivate Students' Practical Operational Abilities Strengthening the experimental stage is crucial in the reform of machine learning curriculum teaching. By designing diverse experimental projects, students can gain a deeper understanding of theoretical knowledge and improve their practical operational abilities[5,6]. For example, when teaching deep learning algorithms, we designed an experimental project based on image recognition, requiring students to use deep learning frameworks to build models and classify images. Through practical operation, students not only master the basic principles of deep learning algorithms, but also learn how to use tools for model training and optimization. In addition, we have introduced competitive experiments to encourage students to participate in machine learning competitions, and to exercise their practical operation and problem-solving abilities by solving practical problems. According to statistics, students who participate in competitions exhibit higher enthusiasm and creativity in subsequent course learning and project practice. In the design of the experimental phase, we also pay special attention to cultivating students' teamwork ability and innovative thinking. By collaborating in groups to complete experimental projects, students have learned how to effectively communicate with others, divide tasks and work together to solve problems. At the same time, we also encourage students to put forward their innovative ideas and try and verify them in experiments. This open experimental environment not only stimulates students' creativity, but also cultivates their critical thinking and problem-solving abilities. By strengthening the experimental process, we are committed to stimulating students' creativity and cultivating them into machine learning talents with practical operation ability and innovative spirit. 4. Innovative Teaching Methods: Adopting Diversified Teaching Methods to Enhance Teaching Effectiveness 4.1. Introducing Case Teaching to Improve Classroom Interactivity In the reform of machine learning curriculum teaching, introducing case teaching is an effective strategy that can significantly improve classroom interactivity[6]. By selecting representative and practical cases, teachers can enable students to gain a deeper understanding of machine learning theory and exercise their practical application abilities in the process of analyzing, discussing, and solving problems. For example, when teaching decision tree algorithms, a real classification problem case can be introduced, such as predicting customer churn. Students need to use decision tree algorithms to process and analyze data, and attempt to construct effective classification models. During this process, the teacher guides students to discuss and share, encouraging them to come up with their own insights and solutions. Through the introduction of case teaching, classroom interactivity has been significantly improved, and student participation and enthusiasm have also been effectively stimulated. In addition, case teaching can also help students establish an intuitive understanding of machine learning technology and enhance their practical skills. By analyzing and solving real cases, students can gain a deeper understanding of the application of machine learning algorithms in practical problems, and learn how to choose appropriate algorithms and parameters based on specific needs. This practice oriented teaching method not only helps to improve students' hands-on abilities, but also cultivates their innovative thinking and problem-solving abilities. At the same time, case teaching can also promote communication and interaction between teachers and students, providing teachers with opportunities to understand student learning situations and provide feedback, thereby further optimizing teaching content and methods. 4.2. Utilize Online Resources to Build a Blended Learning Model In the reform of machine learning curriculum teaching, utilizing online resources to construct a blended learning model has become a key measure to improve teaching effectiveness[7]. By introducing online learning platforms and online programming environments, students can learn independently outside of class and consolidate their classroom knowledge. According to statistics, after adopting the blended learning mode, the average learning time of students increased by 30%, and course satisfaction also significantly improved. This teaching model not only improves the learning efficiency of students, but also cultivates their self-learning ability and teamwork ability. In addition, the blended learning model fully utilizes the richness and interactivity of online resources. By introducing online testing, online discussion areas, and real-time feedback systems, teachers can timely understand students' learning situations and problems, and provide targeted guidance and assistance. Meanwhile, students can also access more learning materials and cases through online resources, broaden their academic horizons, and improve their problem- solving abilities. This teaching model not only conforms to the learning habits of modern students, but also conforms to the development trend of educational informatization. Through the organic combination of online and offline, students can explore and innovate in a more flexible and autonomous learning environment, laying a solid foundation for future machine learning research and application. 5. Improvement of Evaluation System: Establishing a Scientific and Reasonable Curriculum Assessment System In the teaching reform of machine learning courses, we focus on adopting diversified assessment methods to comprehensively evaluate students' abilities [8]. We have introduced various assessment methods such as project 148 reports, classroom discussions, and practical operations. Through project reports, we require students to apply the knowledge they have learned to practical problem-solving, and to test their understanding and application abilities through practical operations. Classroom discussions encourage students to actively participate, express their opinions and insights, and cultivate their critical thinking and expression abilities. Practical operation tests students' hands- on ability and problem-solving ability through programming experiments, data analysis, and other methods. This diversified assessment method can more comprehensively reflect the ability level of students and avoid the one- sidedness and limitations that may arise from a single evaluation method. Taking project reports as an example, during the project process, students need to collect data, design algorithms, write code, and ultimately submit a detailed report. Through this project, we can not only evaluate students' understanding and mastery of machine learning algorithms, but also assess their comprehensive abilities such as teamwork and project management. At the same time, we have also introduced a peer review mechanism to allow students to evaluate each other and improve their critical thinking and self reflection abilities. In addition, we also focus on using data analysis tools to conduct in-depth analysis of assessment results. By conducting statistical analysis on the various assessment scores of students, we can discover their strengths and weaknesses in different aspects, providing strong support for subsequent teaching improvements. At the same time, we can also adjust and optimize the assessment methods based on student feedback, making them more in line with their actual needs and learning characteristics. Through the application of diversified assessment methods, we can not only comprehensively evaluate students' abilities, but also stimulate their interest and creativity in learning, cultivate their comprehensive qualities and innovative abilities. This is of great significance for cultivating high- quality talents who can meet the needs of future social development. 6. Summary After a series of teaching reforms and explorations, machine learning courses have achieved significant results in strengthening foundations and practical orientation. Students have not only significantly improved their mastery of theoretical knowledge, but also gained good exercise in practical operation and innovative thinking. At the same time, the innovation of teaching methods and the improvement of evaluation systems also provide strong guarantees for the improvement of course quality. We will continue to deepen teaching reform and explore teaching models that are more suitable for student needs and social development trends. We will further strengthen the experimental process, provide more practical opportunities in real-life scenarios, and help students better apply theoretical knowledge to practical problems. At the same time, we will also focus on cultivating students' critical thinking and innovative abilities, encouraging them to be brave enough to try new methods and ideas, and promoting the development of machine learning. Acknowledgments This work was supported by the Association of Fundamental Computing Education in Chinese Universities (No: 2023-AFCEC-270). References [1] Simard, Patrice Y., et al. "Machine Teaching: A New Paradigm for Building Machine Learning Systems." ArXiv, 2017, /abs/ 1707. 06742. Accessed 3 Aug. 2024. [2] Shieh R S, Chang W. Fostering student’s creative and problem- solving skills through a hands-on activity[J]. Journal of Baltic Science Education, 2014, 13(5): 650. [3] Karan E, Brown L. Enhancing Student's Problem-Solving Skills through Project-Based Learning[J]. Journal of Problem Based Learning in Higher Education, 2022, 10(1): 74-87. [4] Prianto A, Qomariyah U N, Firman F. Does student involvement in practical learning strengthen deeper learning competencies? [J]. International Journal of Learning, Teaching and Educational Research, 2022, 21(2): 211-231. [5] Hang L T, Van V H. Building Strong Teaching and Learning Strategies through Teaching Innovations and Learners' Creativity: A Study of Vietnam Universities[J]. International Journal of Education and Practice, 2020, 8(3): 498-510. [6] Tashtoush M A, Wardat Y, Aloufi F, et al. The effectiveness of teaching method based on the components of concept-rich instruction approach in students achievement on linear algebra course and their attitudes towards mathematics[J]. Journal of Higher Education Theory and Practice, 2022, 22(7). [7] Dakhi O, JAMA J, IRFAN D. Blended learning: a 21st century learning model at college[J]. International Journal of Multi Science, 2020, 1(08): 50-65. [8] Zhai X, Yin Y, Pellegrino J W, et al. Applying machine learning in science assessment: a systematic review[J]. Studies in Science Education, 2020, 56(1): 111-151.