A B S T R A C T

By discovering and creating novel drugs to treat a range of illnesses, the discipline of drug discovery and development plays a vital role in healthcare. Conventional approaches to drug development have been costly, time-consuming, and frequently produce drugs that don't work for every patient. Precision medicine, on the other hand, seeks to customize medical care to each patient's unique needs while accounting for lifestyle, environment, and genetics. Artificial Intelligence (AI) has revolutionized drug discovery and development in recent years when it has become a potent instrument. Machine learning and deep learning are two examples of AI technologies that could drastically speed up medication discovery, lower prices, and increase treatment efficacy. Researchers can find possible therapeutic targets, create new compounds, and forecast patient response to treatment with the use of artificial intelligence (AI), which analyzes vast datasets and find patterns. This research investigates how AI can be used to find and produce drugs for precision medicine. With adding that, it gives a summary of the conventional drug discovery procedure, emphasizing its drawbacks. Later this research describes how AI technologies are being applied to solve these obstacles, with particular attention to how they are being employed in clinical trials, target identification and validation, and computational drug design. The research also looks at how AI may help to provide personalized medicine, in which each patient receives a customized course of therapy. In summary, this study attempts to present a thorough analysis of the state of AI-driven drug development today and how it can revolutionize precision medicine. Understanding the developments and difficulties in this field helps us to better grasp how AI may transform healthcare in the future and enhance patient outcomes.

DOI: https://doi.org/10.63471/jamsai24004 @ 2024 Journal of Advances in Medical Sciences and Artificial Intelligence (JAMSAI), C5K Research Publication  

1. Introduction

New drug discovery and development is a time-consuming, costly, and complex process that usually involves some crucial steps. To find possible therapeutic candidates, researchers start by going through vast libraries of peptides, chemical compounds, natural products, and other substances. Known active compounds are frequently chemically altered to produce novel analogues with enhanced characteristics. Finding new therapeutic targets involves scientists identifying molecules or biological pathways that are implicated in diseases and creating medications that can interact with these targets. Using their understanding of biological processes and receptor architecture, scientists create novel compounds to address specific illnesses. Once promising candidates are distinguished, they undergo pre-clinical testing and clinical trials in four phases to survey safety and efficacy. However, the process is ruined by high costs, low success rates, and long lines (Lee, 2023). 

AI offers a solution by accelerating screening, predicting viability and safety, and streamlining drug improvement, which can diminish costs and move forward success rates. Despite challenges such as data quality and ethical concerns, AI has the potential to revolutionize drug discovery by making it speedier, more proficient, and cost-effective. The way AI and machine learning (ML) may revolutionize statistical analysis and decision-making in healthcare is astounding. AI-powered predictive analytics, based on patient-specific data, can assist medical professionals in selecting optimal medications, dosage adjustments, and treatment plans, thereby enhancing patient care and treatment results (Vyas et al., 2022). Analyzing how AI-powered analytics could improve system efficacy and decision-making is crucial. Predictive data analysis is one circumstance where statistics prompted by AI function efficiently. Through the examination of past patient data, including genetic, wearable, genomic, medical imaging, along with electronic health record (EHR) data, artificial intelligence (AI) systems can recognize patterns of depth and anticipate important findings. As time passes, numerous innovations have been slowly incorporated into healthcare systems to enhance productivity and outcomes (Bajwa et al., 2021). From the manufacturing of medications through their transmission to patients, artificial intelligence (AI) can be highly significant. It could assist enabling rational prescription drugs design, decision-making, integrating the clinical information generated, deciding the best method of therapy for patients, incorporating PM, and utilizing it to subsequent drug development. AI may advance the development of drug frameworks, uncover advantageous particles, and accelerate the validation of targets for therapeutic use (Pal & Anjum Taqi, 2020).

 2. Literature Review:

Artificial Intelligence (AI) applications in drug development are based on Machine Learning (ML) techniques (Wang et al., 2024). These algorithms don't require detailed programming in order recognize patterns and produce predictions from enormous data sets. Quantitative Structure-Activity Correlation (QSAR) modeling, virtual screening, and molecular activity predictions are just a few of the tasks in drug discovery which employ machine learning (ML) methods (Ping et al., 2024). Predicting molecular characteristics and processes has proven to be an effective strategy for algorithms that use supervised learning like Support Vector Machines (SVM), Random Forests (RF), and Gradient Boosting Machines (GBM) (Xu et al., 2024). Using labeled training data, these algorithms generate models that can generalize to novel, undiscovered chemicals. SVM outperforms other classification algorithms in the task of predicting human intestinal absorption (HIA), demonstrating encouraging findings (Shi et al., 2024).

Artificial Intelligence (AI) in individualized therapy is a significant step toward developing more efficient and customized healthcare for every person. Customized care, also known as precision medicine, aims to modify medical care according to an individual's distinct traits, such as their environment, lifestyle, and heredity. AI is crucial in this regard because it can analyze vast volumes of data, including genetic data, medical records, and clinical trial findings, to find patterns that conventional approaches would miss. Artificial Intelligence integrates all these various sources of information with sophisticated machine learning to produce a thorough and customized treatment plan for every patient. This enables medical professionals to treat patients more precisely and successfully depending on their unique needs (Khezr et al., 2019). By evaluating data from multiple sources, including genetics, lifestyle, and diagnostic tests, artificial intelligence assists in the personalization of treatment in cardiology. AI can estimate the probability of heart issues, such as heart attacks or strokes, by examining a patient's medical history and personal risk factors. This enables medical professionals to develop individualized treatment programs that are suited to the requirements of each patient. These strategies might encompass adaptations to lifestyle, medical treatment, and surveillance strategies. AI thus contributes to better patient care by assisting in the management of cardiac diseases (Kooli & Al Muftah, 2022).

 Chronic illnesses such as diabetes, asthma, and COPD (chronic obstructive pulmonary disease) necessitate continuous management and frequent modifications to treatment regimens. To provide real-time insights into a patient's status, artificial intelligence (AI)-powered solutions can evaluate data from electronic health records (EHRs), wearable technology, and remote monitoring systems. AI can forecast blood glucose levels in diabetics, for example, by considering variables like food consumption, exercise, and medication compliance. This makes it possible to modify insulin dosage more precisely and gives patients the ability to manage their own health (Kumar et al., 2021). Artificial intelligence (AI)-powered robotics is revolutionizing surgery by increasing accuracy, lowering risks, and optimizing patient outcomes. Compared to traditional surgical techniques, the combination of artificial intelligence (AI) and robotic technology gives surgeons more control, accuracy, and efficiency. These sophisticated robotic instruments improve surgeons' abilities and change the method by which surgeries get done through rendering it less difficult for professionals to carry out complicated and less-invasive treatments (Pal & Anjum Taqi, 2020). Additionally, AIpowered robotics may lead to better surgical education and skill improvement. Robotic simulators incorporating artificial intelligence (AI) allow surgeons to hone their skills in a safe setting. Using these simulators, which offer real-time performance feedback, surgeons can hone their abilities prior to doing actual procedures. By bridging the gap between theoretical understanding and real-world experience, this leading-edge curriculum eventually improves surgeon performance and patient safety. The use of AI-driven robotics in surgery is not beyond obstacles and worries and not withstanding recent developments (Saeed et al., 2022).

Improved surgical training and skill improvement are further potential benefits of AI-driven robotics. Robotic simulators incorporating artificial intelligence (AI) allow surgeons to hone their skills in a safe setting. Using these simulators, which offer real-time performance feedback, surgeons can hone their abilities prior to doing actual procedures. By bridging the gap between theoretical understanding and practical applications, this cutting-edge training system eventually improves surgeon performance and patient safety. The use of AI-driven robotics in surgery is not without its difficulties and concerns, notwithstanding these developments (Saeed et al., 2022). One difficulty is the potentially high cost of robotic systems. Some healthcare facilities may find the initial cost of robotic technology, as well as continuing maintenance and operating expenses, to be prohibitive. Furthermore, training and modification are needed to integrate robotic systems into current surgical workflows. To integrate the technology into their practice, surgeons and surgical teams need to become adept in its use. To guarantee the efficacy and safety of AIdriven robotics, further research and development is required. Although there have been encouraging outcomes from technology, ongoing observation and assessment are required to resolve possible problems and improve algorithms. Upholding strict guidelines for patient care and safety requires that robotic systems undergo extensive testing and validation. Artificial intelligence (AI)-powered robotics is transforming surgery by improving accuracy, control, and productivity. Improved surgical results, less invasive treatments, and more sophisticated training possibilities are just a few advantages of integrating AI technology with robotic systems. The development of AI-driven robotics has enormous potential for the future of surgical practice, notwithstanding certain obstacles that must be overcome (Samarpita & Satpathy, 2022). As technology advances, the role of AI in surgery is anticipated to grow more prominent, resulting in more effective and patient-centered surgical care. By utilizing AI-powered robotics, medical professionals can attain unprecedented levels of accuracy and creativity in the operating room, ultimately improving patient outcomes and developing the surgical field.

3. Artificial Intelligence in Drug Design, Drug Recovery & Therapies for Genetic Disorders

AI has the potential to significantly impact pharmaceutical product development, from conception to patient delivery. It can help with logical medication design, selecting the best course of treatment for each patient, including PM, maintaining the clinical data generated, and applying it to the creation of new drugs. Therapeutic structure design can be optimized, therapeutic targets can be validated more quickly, and effective compounds can be found with the use of AI (Mak & Pichika, 2019). Understanding how different learning approaches are applied in machine learning (ML), with an emphasis on their use in drug development, is crucial. It emphasizes how AI may be used to gather and arrange data as well as create new chemical structures. Reducing the amount of data needed for training to mimic the way the human brain learns is one technique to lessen the difficulties. This method works well in domains like medicinal chemistry and new drug target research where large datasets are not easily accessible. Presently, scientists are investigating novel approaches to tackle the problem of restricted data accessibility in drug discovery. They are looking into one-shot learning strategies, lengthy short-term memory techniques, and memory-augmented neural networks like differentiable neural computers, for instance. With the goal of making DL more flexible in scenarios with constrained data availability, these developments show promise in improving the effectiveness and application of DL models in drug discovery. In the end, these initiatives aim to speed up the process of finding and creating new medications.

4. Vaccine, Protein Design & Small Molecule Design

Immunotherapies for cancer use the patient's immune system to combat malignancies (Koury et al., 2018). A recent consortium effort has benchmarked a wide range of features, such as the MHC-binding affinity ratio between the neoepitope and wildtype peptide, similarity to viral peptide sequences, and physicochemical properties, which are thought to be linked to neoepitope immunogenicity, in order to predict reactive immune neoepitopes. The majority of T-cell reactive neoepitopes were shown by the study to be very alien or agretopic (Wells et al., 2020). Recent advances in unsupervised representation learning of protein sequences have sparked a boom in protein engineering, effectively utilizing hundreds of millions of sequenced proteins to determine which positions co-evolve within a vast collection of evolutionary diverse datasets belonging to the same protein family [Figure 3(b)]. By using such co-evolutionary data, AI can determine which protein residues are essential to its function and which can be modified to customize its characteristics. This allows for the customization of current treatments for individual patients while lowering the risk of immune-related adverse effects (Schubert et al., 2018). Fig. 1 shows AI methodologies in drug development, supervised prediction methods & deep neutral networks. A deep residual network that forecasts the distribution of contact distances is used in this model to refine the anticipated structural contacts from co-evolutionary models.

Fig. 1. AI methodologies in drug development, supervised prediction methods & deep neutral networks 

To produce incredibly precise 3D protein structures, these are immediately reduced and evaluated as the statistical energy of the protein fold. Extensions have now been made to further eliminate the need on coevolutionary analysis (Xu et al., 2021), enabling the prediction of three-dimensional structures of synthetic proteins and protein families from a small number of sequences. Neural protein folding simulators that can directly map sequence to structure and are fully trainable in an end-to-end fashion have begun to be developed by others (AlQuraishi, 2019).

New small compounds with optimized properties can be produced by combining the learnt latent representations with optimization methods like reinforcement learning [203,204 and Bayesian optimization [193,202. Previous findings have mostly or entirely depended on userspecified scoring functions to direct the optimization process; however, more recent research incorporates molecular property predictions into the reward signal (Popova et al., 2018) and expands the concept to multicriterion optimization (Ståhl et al., 2019). A standard set of validation criteria has not yet been established, despite the rise in novel AI-based small molecule design methodologies. These problems were solved by two new benchmarking platforms. While Brown et al. (2019) investigated the models' capacity to reproduce the physicochemical property distributions of a reference set and to produce novel molecules that optimize one or more criteria simultaneously, (Polykovskiy et al., 2018) assessed generative models based on their ability to produce structures like the training set.

In conclusion, AI's ability to acquire latent representations that underline the functional characteristics of molecules and use the Euclidean structure of the resulting embedding manifold to enhance those attributes has significantly helped small molecule design. Due to the ability to accurately identify and eliminate subpar drug candidates in silico, this made it possible to create effective libraries for high-throughput drug screenings, which can ultimately result in a notable improvement in the success rate of subsequent clinical trials. Many generative models require post-hoc structural fine-tuning using molecular dynamics because they are trained end-to-end and only implicitly learn about physical plausibility. Deep learning-based de novo design may be enhanced by more physics-informed networks that limit the latent space to regions producing chemically viable solutions, much like generative models for proteins. We believe that with better pharmacogenomics screening datasets, models that can produce novel compounds conditioned on mutational alterations of their target protein will soon be produced, even though the use of such models in a customized environment has not yet been demonstrated.

5. Potentials of AI in the Development of Therapies for Genetic Disorders

AI holds great promises to find treatments for hereditary illnesses brought on by gene mutations. AI-driven methods provide novel ways to pinpoint therapeutic targets, clarify disease causes, and create precise treatments based on patient profiles. These strategies include virtual screening, drug repurposing, target discovery and prioritization, and disease mechanism elucidation. These techniques also include discovering current medications or small compounds that can be repurposed for treating genetic disorders, prioritizing possible therapeutic targets, and examining genomic, transcriptomic, and proteomic data to clarify disease mechanisms.

6. Limitation of AI Tools

The development of AI systems in pharmaceutical sciences has been severely hampered by their heavy reliance on data. AI models could lack sufficient data for certain medical conditions, such uncommon diseases. Inadequate data quality could lead to skewed projections. Providing trustworthy and high-quality data from multiple sources may raise the total cost of research and development processes for pharmaceutical businesses (Alakwaa et al., 2018). In a broad chemical space, artificial intelligence (AI) and machine learning (ML) approaches can aid in the prediction of novel chemical compounds with advantageous pharmacological properties. The intricacy of the chemical space and the dearth of labeled data for predictive model training present difficulties, though. Another challenge is making sure that the compounds that are created are distinct, diverse, and synthetically accessible. It will take interdisciplinary collaboration, creative algorithmic techniques, and strong validation methodologies to fully realize the potential of AI/ML in exploring unexplored chemical space and finding new medications. AI must be used in conjunction with humans to address ethical issues including data privacy and potential bias in drug discoveries in novel sectors. To maximize productivity while resolving technical difficulties and ethical considerations, human-AI collaboration is essential (Han et al., 2023).

7. Expert Opinion

Through developments in biomarker identification, medication repurpose, combination, and drug design, artificial intelligence has made precision medicine possible. In contrast to traditional linear drug development pipelines, the effective application of the discussed AI-based methods in drug discovery and development will allow for more nuanced and iterative processes, which could ultimately lower R&D costs and improve clinical trial success rates, thereby increasing the overall efficiency of the pharmaceutical industry. There are a lot of restrictions and difficulties with this progress, though. Validating AI-driven methods in early drug development is difficult. Even while the number of computer models for designing vaccines, proteins, and small molecules is rapidly growing, only a tiny percentage of them have experimental validations or are verified using a shared set of benchmarking samples. Frequent crowd-sourcing contests, like the Dialog for Reverse Engineering Assessments and Methods (DREAM) challenges (Stolovitzky et al., 2007), could help solve this, allowing for an objective assessment of computational approaches and encouraging cooperation in a rapidly developing scientific field. Second, data privacy is a crucial concern because the information utilized for medication development in precision medicine includes highly individualized information like a person's genetic and clinical characteristics. Since many rich data sets are unavailable and secured in data silos, it can be challenging (Rieke et al., 2020) to design AI applications because such data is frequently difficult to transfer or physically leave a single area. This issue is addressed by federated learning, which permits decentralized machine learning model training and prediction without ever requiring data to leave its physical location. Without ever exchanging sensitive data, federated learning will allow a linked healthcare network to help doctors identify similar patient cases and guide their treatment choices. Additionally, it will open new business models that allow specialized firms to provide AI-based analytics for biomedical and healthcare study without requiring direct access to the data. Its wider adoption is, however, constrained by legal and technical issues and heavily relies on uniform medical records across clinical networks. We think that as this technology advances, these difficulties will eventually be resolved as well. With so many advantages, AI technologies could completely transform the quality assurance and control systems used in the pharmaceutical business. Conventional techniques might not always be precise, but AI applications can identify subtle differences that affect the functionality and safety of products.AI can improve process accuracy and product quality in pharmaceutical quality assurance and control. To guarantee uniformity from batch to batch, manual quality control inspections are still required. In the pharmaceutical business, new technologies like intelligent quality control systems and electronic lab notebooks promise to guarantee product quality and sufficient quality assurance (GJMI et al., 2018).

With the advancement of AI and data collecting technology, it will someday be feasible to model a sizable percentage of a patient's biological systems, creating a computer model that can be used to predict how the patient will react to a particular treatment. Doctors will be able to quickly customize a treatment plan for each patient based on their unique circumstances thanks to AI-supported digital twins. The pharmaceutical industry will be impacted by the availability of such data, which will accelerate the fully automated and integrated translation from biomarker discovery to medication design, reducing potential expenses along the drug development pipeline and increasing clinical trial efficiency. Already underway in precision oncology, this shift will soon spread to other medical specialties and cause the pharmaceutical industry to undergo a significant upheaval. The integration of multi-omic data sets and the more efficient handling of various data kinds are made possible by advances in algorithms. This enables scientists to better understand the mechanics underlying diseases and find novel, more effective treatment targets. Furthermore, new opportunities for de novo drug discovery have been made possible by the development of generative AI models. These models increase the chemical space available for drug discovery by using DL approaches to create molecular structures with certain pharmacological properties. This strategy may hasten the development of ground-breaking drugs for a range of illnesses. PM and AI-driven drug development are complemented by several exciting new fields in biological research. These consist of precision oncology, gene therapy, immunotherapy, regenerative medicine, and microbiome research. Future research in the biomedical field will undoubtedly be complex, and it is important to note that these fields can advance alongside or in tandem with PM and AI-driven drug discovery. Furthermore, typical methods and practices in PM and AI-driven drug discovery will be significantly impacted by technical advancements, data integration, regulatory reforms, and cooperative initiatives. This dynamic environment, which benefits from a comprehensive approach to biomedical research, is likely to result in rapid advancements in discipline. In conclusion, without the assistance of cutting-edge AI techniques developing in drug research and development, precision medicine, especially customized medicine, cannot be realized in clinical practice. The drug development pipeline is still a sluggish and ineffective process, even with improvements in treatment approaches and the growth in genomic and molecular information available. One of the biggest obstacles to medical research and development to date is the acceptance and acceleration of standard drug development procedures for precision and customized medicine. Patients, physicians, and the pharmaceutical industry will all be significantly impacted by the move to a data-driven healthcare system. Several legislative and technical obstacles must be removed before an AI-driven precision medicine strategy can be implemented. Nonetheless, we are confident that there will be a significant positive social impact.

8. Conclusion:

The translational potential of PM is expected to be improved by integrating real-world evidence (RWE) into drug discovery procedures. Researchers can gain a thorough understanding of disease phenotypes and treatment responses in real-world situations by combining data from wearable technology, electronic health records, and patient-reported outcomes. Additionally, using AI-driven drug repurposing techniques provides an affordable way to find new uses for already-approved medications. Even while PM and  AI provide exciting prospects for drug discovery, the industry's advancement is hampered by a number of important issues and obstacles. Coordinated efforts on several fronts, such as creating data standards, encouraging interdisciplinary cooperation, creating regulatory frameworks, and promoting technical innovation, will be necessary to overcome these challenges. The development of data harmonization techniques and interoperability systems that provide the smooth integration of multi-omics datasets is crucial for ensuring more reliable analysis and interpretation. PM and AI-driven drug development are complemented by several exciting new fields in biological research. These consist of precision oncology, gene therapy, immunotherapy, regenerative medicine, and microbiome research. Future research in the biomedical field will undoubtedly be complex, and it is important to note that these fields can advance alongside or in tandem with PM and AI-driven drug discovery. AIdriven medication discovery and standard methods and practices in PM will also be significantly impacted by technical developments, data integration, regulatory changes, and cooperative initiatives. This dynamic environment, which benefits from a comprehensive approach to biomedical research, is likely to result in rapid advancements in discipline. To fully fulfill AI's promise in drug research and discovery, a few issues and concerns must be resolved. These include addressing ethical issues, creating accessible and understandable AI systems, and guaranteeing access to high-quality data.AI-driven drug discovery and development for precision medicine has a bright future despite these obstacles. We can anticipate more individualized and efficient treatments for a variety of illnesses as AI technologies develop and are more fully included into the drug discovery process. Researchers have the chance to transform medicine and enhance the lives of millions of people worldwide by utilizing AI's potential. AI-powered biopharmaceuticals have enormous potential to transform drug research and discovery in the future. By resolving existing constraints, creating strong regulatory frameworks, and encouraging cooperative ecosystems, the sector is well-positioned to achieve new heights of efficiency and innovation in the pursuit of innovative therapies.

References

Alakwaa, F. M., Chaudhary, K., & Garmire, L. X. (2018). Deep Learning Accurately Predicts Estrogen Receptor Status in Breast Cancer Metabolomics Data. J Proteome Res, 17(1), 337- 347. https://doi.org/10.1021/acs.jproteome.7b00595 

AlQuraishi, M. (2019). End-to-End Differentiable Learning of Protein Structure. Cell Systems, 8(4), 292-301.e293. https://doi.org/https://doi.org/10.1016/j.cels.2019. 03.006

Bajwa, J., Munir, U., Nori, A., & Williams, B. (2021). Artificial intelligence in healthcare: transforming the practice of medicine. Future Healthc J, 8(2), e188-e194. https://doi.org/10.7861/fhj.2021-0095

Han, R., Yoon, H., Kim, G., Lee, H., & Lee, Y. (2023). Revolutionizing Medicinal Chemistry: The Application of Artificial Intelligence (AI) in Early Drug Discovery. Pharmaceuticals (Basel), 16(9). https://doi.org/10.3390/ph16091259

Khezr, S., Moniruzzaman, M., Yassine, A., & Benlamri, R. (2019). Blockchain Technology in Healthcare: A Comprehensive Review and Directions for Future Research. Applied Sciences, 9(9).

Kooli, C., & Al Muftah, H. (2022). Artificial intelligence in healthcare: a comprehensive review of its ethical concerns. Technological Sustainability, 1(2), 121-131. https://doi.org/10.1108/TECHS-12-2021-0029

Koury, J., Lucero, M., Cato, C., Chang, L., Geiger, J., Henry, D., Hernandez, J., Hung, F., Kaur, P., Teskey, G., & Tran, A. (2018). Immunotherapies: Exploiting the Immune System for Cancer Treatment. J Immunol Res, 2018, 9585614. https://doi.org/10.1155/2018/9585614

Kumar, P., Kumar, R., & Gupta, M. (2021). Deep learning based analysis of ophthalmology: A systematic review. EAI Endorsed Transactions on Pervasive Health and Technology, 7(29).

Lee, D. (2023). AI-Driven Drug Discovery and Development for Precision Medicine. African Journal of Artificial Intelligence and Sustainable Development, 3(1), 1-11.

Mak, K.-K., & Pichika, M. R. (2019). Artificial intelligence in drug development: present status and future prospects. Drug Discovery Today, 24(3), 773-780. https://doi.org/https://doi.org/10.1016/j.drudis.2018.11.014

Pal, P., & Anjum Taqi, S. A. (2020). Advancements in Data Mining and Machine Learning Techniques for Predicting Human Diseases: A Comprehensive Review. International Journal of Research in Informative Science Application & Techniques (IJRISAT), 4(11), 19-35. https://doi.org/10.46828/ijrisat.v4i11.85

Ping, G., Wang, S. X., Zhao, F., Wang, Z., & Zhang, X. (2024). Blockchain based reverse logistics data tracking: An innovative approach to enhance ewaste recycling efficiency.

Popova, M., Isayev, O., & Tropsha, A. (2018). Deep reinforcement learning for de novo drug design. Sci Adv, 4(7), eaap7885. https://doi.org/10.1126/sciadv.aap7885

Rieke, N., Hancox, J., Li, W., Milletarì, F., Roth, H. R., Albarqouni, S., Bakas, S., Galtier, M. N., Landman, B. A., Maier-Hein, K., Ourselin, S., Sheller, M., Summers, R. M., Trask, A., Xu, D., Baust, M., & Cardoso, M. J. (2020). The future of digital health with federated learning. npj Digital Medicine, 3(1), 119. https://doi.org/10.1038/s41746-020-00323-1

Saeed, U., Shah, S. Y., Ahmad, J., Imran, M. A., Abbasi, Q. H., & Shah, S. A. (2022). Machine learning empowered COVID -19 patient monitoring using non -contact sensing: An extensive review. Journal of Pharmaceutical Analysis , 12(2), 193 -204. https://doi.org/https://doi.org/10.1016/j.jpha.2021.12.006

Samarpita, S., & Satpathy, R. N. (2022). Applications of machine learning in healthcare: an overview. 2022 1st IEEE International Conference on Industrial Electronics: Developments & Applications (ICIDeA),

 Schubert, B., Schärfe, C., Dönnes, P., Hopf, T., Marks, D., & Kohlbacher, O. (2018). Population -specific design of de -immunized protein biotherapeutics. PLoS Comput Biol , 14(3), e1005983. https://doi.org/10.1371/journal.pcbi.1005983

Shi, Y., Shang, F., Xu, Z., & Zhou, S. (2024). Emotion - driven deep learning recommendation systems: Mining preferences from user reviews and predicting scores. Journal of Artificial Intelligence and Development , 3(1), 40 -46.

Ståhl, N., Falkman, G., Karlsson, A., Mathiason, G., & Boström, J. (2019). Deep Reinforcement Learning for Multiparameter Optimization in de novo Drug Design. Journal of Chemical Information and Modeling , 59(7), 3166 -3176. https://doi.org/10.1021/acs.jcim.9b00325

Wells, D. K., van Buuren, M. M., Dang, K. K., Hubbard - Lucey, V. M., Sheehan, K. C. F., Campbell, K. M., Lamb, A., Ward, J. P., Sidney, J., Blazquez, A. B., Rech, A. J., Zaretsky, J. M., Comin -Anduix, B., Ng, A. H. C., Chour, W., Yu, T. V., Rizvi, H., Chen, J. M., Manning, P., . . . Defranoux, N. A. (2020). Key Parameters of Tumor Epitope Immunogenicity Revealed Through a Consortium Approach Improve Neoantigen Prediction. Cell , 183(3), 818 - 834.e813. https://doi.org/https://doi.org/10.1016/j.cell.2020.09.015

Xu, H., Niu, K., Lu, T., Li, S., & Lou, Q. (2024). Leveraging artificial intelligence for enhanced risk management in financial services: Current applications and future prospects. Academic Journal of Sociology and Management , 2(5), 38 -53.

Xu, J., McPartlon, M., & Li, J. (2021). Improved protein structure prediction by deep learning irrespective of co -evolution information. Nat Mach Intell , 3, 601 - 609. https://doi.org/10.1038/s42256-021-00348-5