INTRODUCTION:

Artificial intelligence is the application of computer technology to imitate human reasoning and problem-solving abilities. In 1956, Marvin Minsky and John McCarthy organized a symposium where the idea was initially put forth¹.

In general, the process of AI development involves four essential steps: gathering and preparing data, creating AI models, running simulations and tests, and implementing the solution². A segment of AI, machine learning leverages algorithms to uncover patterns in data, enhancing decision-making capabilities³. Examples of decision-making include operational decisions in healthcare⁴ and risk forecasting decisions^5,6. As layered structure algorithms, ANNs are a key component of deep learning within machine learning. ANNs, enthused by the neural structural design of the individual brain, exhibit greater computational and predictive abilities than traditional machine learning approaches⁷. Furthermore, a wide range of applications, including picture classification⁸, object recognition^9,10, image segmentation¹¹, natural language processing^12,13, and medical image investigation¹⁴, has extensively used deep learning. Drug development and drug discovery can also be performed by AI¹⁵.

In the field of pharmaceuticals, artificial intelligence has the potential to revolutionize multitudes of aspects related with pharmaceutical field. In this article, we provide an overview of the benefits and applications of artificial intelligence in the pharmaceutical industry, including drug discovery, clinical trial design, personalized medicine, streamlining drug development, and enhancing drug safety¹⁶.

This review introduces artificial intelligence (AI) as a methodology with the capacity for induction and deduction. AI is further classified into 3 categories¹⁷.

Passive AI: This method employs deep learning along with information adaptability, streamlining data into easy-to-understand outputs, such as recommendations or informed alerts. Predominantly used in consumer markets, these systems typically follow a basic cycle of gathering information, applying it for guidance, and recommending changes to improve future iterations.

Predictive AI: This passive AI subset specializes in modeling and organizing data to create forecasts or predictions. Typically used in consumer markets, healthcare, and finance, predictive AI can generate data through deduction when evaluating models, potentially outperforming standard databases. This capability aids the development of AI in domains with sparse data and a lack of available expertise¹⁸.

Active AI: An artificial intelligence (AI) system controls a basic or sophisticated robotic system to sense, probe, and autonomously explore the real world (or a virtual description). The engagement of the AI with the physical environment serves as a source of direct confirmation, self-obtaining the knowledge and modifying the program, in contrast to the passive method¹⁹. This approach may actively close knowledge gaps. As a result, these programs cover factual machine learning experiments for situations for which there is no prior experience^20-22.

A cutting-edge hybrid framework has been created, utilizing state-of-the-art natural verbal communication processing and self-supervised learning techniques^23-25. It is becoming more popular for computer vision applications^26-29, and it may be crucial in future studies involving medical imaging applications. Given that it operates on unlabeled data, self-supervised learning is comparable to unsupervised learning. Self-monitored algorithms operate in two stages. Initially, a "pretext task" is given to the model, and its goal is to extract such decision-making signals from the data by pre-training. Subsequent, the "downstream task" is solved by using the learned information and optimizing the model. The high-resolution data can be analyzed with advance Machine Learning (ML) algorithms, it will not only help to discover the disease patterns but also an real-time and objective monitoring of bio-signals can help to discover the unknown patterns linked with CV³⁰.

There are multiple hybrid learning frameworks demonstrating how the lines separating supervised and unsupervised learning have gradually become hazier and mixed methodologies that can deal with real-world issues and data sets.

Figure 1: In artificial intelligence, the three traditional learning frameworks are supervised, semi-supervised, and unsupervised learning.^23-25

Neural Networks:

Three strategies that can be used to correct the situation have been developed with recent developments in mathematics and computer science:

(1) Fuzzy logic: aims to replicate the human brain's ability to make decisions and produce responses based on imprecise or incomplete information ³⁵

(2) Neural Networks: Simulate the functioning of the human brain, incorporating genetic algorithms and evolutionary strategies that facilitate self-organization and adaptation in biological systems. ³⁶

Features of Neural Network³⁷:

They are capable of handling an array of tasks through their versatile applications–

1) Associating: By training on various patterns, neural networks can associate unfamiliar patterns with the most comparable versions they have memorized, switching to those versions when required.

2) Classification: Patterns or data are categorized using neural networks into predefined groups.

3) Clustering: Without being aware of the data beforehand, they classify the data by determining a distinctive feature.

4) Prophecy: They get the predictable outcome from the provided input.

Types of Neural Networking in Deep Learning:

a. Convolutional Neural Network: At their core, convolutional neural networks (CNNs) rely on multilayer perceptron and consist of single or additional convolutional layers that can be either joint or fully linked. The convolutional layers process input through convolution operations, allowing for deeper structures with a reduced number of parameters. CNNs excel in domain such as recommender systems, natural language dealing out, figure and video recognition, paraphrase identification, and semantic processing [38, 39]. Moreover, they are applied in signal processing and agriculture, leveraging satellite meteorological data to anticipate growth and yield on specific land plots.

b. Feed forward Neural Network – Artificial Neuron: A crucial component of artificial neural networks, the artificial neuron is particularly important in feed-forward neural networks. In these structures, data travels from input to output nodes in a single direction, following a path called front propagation that often involves an activation function for classification [40]. Feed-forward networks can include multiple layers, and their outputs are derived from the product of inputs and their weights. While they are used in computer visualization and face detection, they may face challenges in classifying target classes. However, they handle noisy data effectively.

c. Recurrent Neural Network (RNN) – Long Short-Term Memory (LSTM) ⁴¹: It is the reuses its layer outputs as inputs to forecast future outcomes. The first layer operates similarly to a feed-forward network, but RNNs start processing later, with nodes holding onto information from prior steps. Every nodule functions as a recollection unit during computations. The arrangement recalls relevant data and proceeds with forward propagation, using backpropagation to refine predictions when necessary. RNNs excel in applications such as text-to-speech conversion ⁴².

d. Radial Basis Function Neural Network: These processes evaluate the space from each point to the center and consist of two layers. The inner layer combines features using the radial basis function, while the output layer calculates results based on these features ⁴³. These networks are extensively used in power restoration systems, which have become critical as power networks grow larger and more complex, increasing the risk of blackouts. They enable rapid electricity restoration during outages.

e. Multilayer Perceptron: Comprising three or more layers, a multilayer perceptron is used for classifying non-linearly separable data. This fully connected artificial neural network connects all node in one layer to every node in the subsequent layer ⁴⁴, employing nonlinear activation functions like the hyperbolic tangent or logistic function.

f. Modular Neural Network: Composed of multiple independent networks, a modular neural network handles various sub-tasks without interacting with one another during the computing process. This separation allows for quicker task completion, as breaking down a complex computational task into smaller components enhances processing speed ⁴⁵.

g. Sequence-To-Sequence Models: It includes an encoder that processes input data and a decoder for generating output. These components may have different settings, making the model effective when input and output lengths differ. Common applications include chatbots, machine translation, and Q&A systems ⁴⁶.

Neural networks operate in different ways to accomplish a range of tasks, designed to function like human brain neurons. They can learn and adapt as more data is introduced, allowing for growth beyond the limitations of traditional machine learning algorithms, which often plateau. Different types of neural networks are expected to serve as the foundation for future AI innovations.

Table 2: The purpose of artificial neural networks (ANNs) is on the rise in the turf of formulation design along with optimization.

Formulation Type	Study	References
Liposomes	Optimizing Cytarabine Liposome Formulation Parameters: The Role of ANN Compared to Multiple Regression Analysis	47
Hydrogels	Simultaneous Multi-Objective Optimization Using ANN for a Ketoprofen Hydrogel Formulation Enhanced with o-Ethylmenthol for Percutaneous Absorption	48
Hydrogels	Artificial Neural Network-Driven Simultaneous Optimization in a Ketoprofen Hydrogel preparation with o-Ethyl-3-Butylcyclohexanol for Enhanced Percutaneous Absorption.	49
Tablets	Assessing the Impact of investigational plan on the Modeling of Tablet Coating Formulations with ANN	50, 51
	Leveraging ANNs for the range of Key Formulation and handling Variables to Anticipate dissolution Release Profiles of Sustained Release Minitablets.	52
	Harnessing ANN and PK Simulations in the Development of CR Drug Formulations.	53
	Employing comprehensive Regression Neural Networks for the Optimization and Modeling of Aspirin ER Tablets Using Eudragit RS PO Matrix.	54
	The Development and Optimization of Theophyline Controlled-Release Tablets Through ANN Techniques.	55
	ANNs in the Design and Optimization of Aspirin Extended-Release Tablets with Eudragit L 100 Matrix.	56
	Assessing the Differences Between ANN and Traditional Modeling Techniques by means of Data from a Galenic Study on Dosage Forms and Various Experimental Designs.	57
	Identifying the Pitfalls of ANN representing for Datasets with Outliers Through a Study of combination Properties in Directly Compressed Dosage Forms.	58
	Employing ANN for the Optimization of Sustained Release Formulation to Improve Diclofenac	59
Powders	Illustrating the characteristics of solid dosage form Through Artificial Neural Networks and Regression Approaches.	60
Powders	Utilizing ANN for Effective Modeling of Powder Flow Dynamics.	61
Pellets	Predicting Drug Dissolution Profiles with ANNs and Evaluating Model Performance Through Similarity Factor Analysis.	62
Gelispheres	Texture Assessment and Statistical Optimization of Gelisphere Matrices Made from Cross-Linked Calcium Alginate, Pectinate, and Cellulose Acetopthalate.	63
Transdermal	Forecasting Skin Penetration Using ANN Techniques.	64
Transdermal	ANNs for optimizing a vehicle mixture using response surface method for Transdermal delivery of melatonin	65
Granules	Investigating the Fluidized Bed Granulation Process Through Neural Network Modeling.	66
Emulsion	Neural network modeling for predicting viscoelastic behavior and physical stability of Lipophilic semisolid emulsion system	67

REFERENCES:

1. McCarthy J, Minsky ML, Rochester N, Shannon CE. A proposal for the dartmouth summer research project on artificial intelligence, august 31, 1955. AI Magazine. 2006; Dec 15; 27(4): 12-12. https://doi.org/10.1609/aimag.v27i4.1904

2. Jiang J, Ma X, Ouyang D, Williams III RO. Emerging artificial intelligence (AI) technologies used in the development of solid dosage forms. Pharmaceutics. 2022 Oct 22; 14(11): 2257. https://doi.org/10.3390/pharmaceutics14112257

3. Ethem Alpaydin. Introduction to Machine Learning, fourth edition. Adaptive computation and machine learning series, MIT Press, 2020

4. Zain Amin, M.; Ali, A. Performance Evaluation of Supervised Machine Learning Classifiers for Predicting Healthcare Operational Decisions; Technical Report; University of California: Irvine, CA, USA, 2017. https://doi.org/10.3390/pharmaceutics14112257

5. Berk R. An impact assessment of machine learning risk forecasts on parole board decisions and recidivism. Journal of Experimental Criminology. 2017; Jun; 13: 193-216. https://doi.org/10.1007/s11292-017-9286-2

6. Berk RA, Sorenson SB, Barnes G. Forecasting domestic violence: A machine learning approach to help inform arraignment decisions. Journal of Empirical Legal Studies. 2016; Mar; 13(1): 94-115. https://doi.org/10.1111/jels.12098

7. LeCun Y, Bengio Y, Hinton G. Deep learning. nature. 2015; May 28; 521(7553): 436-44. https://doi.org/10.1038/nature14539

8. Affonso, C.; Rossi, A.L.D.; Vieira, F.H.A.; de Carvalho, A.C.P.d.L.F. Deep Learning for Biological Image Classification. Expert Syst. Appl. 2017, 85, 114–122. https://doi.org/10.1016/j.eswa.2017.05.039.

9. Ajay I. Patel, Pooja K. Khunti, Amit J. Vyas, Ashok B. Patel. Explicating Artificial Intelligence: Applications in Medicine and Pharmacy. Asian Journal of Pharmacy and Technology. 2022; 12(4): 401-6.

10. Liu L, Ouyang W, Wang X, Fieguth P, Chen J, Liu X, Pietikäinen M. Deep learning for generic object detection: A survey. International Journal of Computer Vision. 2020; Feb; 128: 261-318. https://doi.org/10.1007/s11263-019-01247-4.

11. Suyash Ingale, Nikhil Shrisunder, Ganesh Gophane, Avinash Birajdar. Ascent of Artificial Intelligence (AI) in Pharmacy. International Journal of Technology. 2024; 14(1): 54-8.

12. Garcia-Garcia A, Orts-Escolano S, Oprea S, Villena-Martinez V, Garcia-Rodriguez J. A review on deep learning techniques applied to semantic segmentation. arXiv preprint arXiv:1704.06857. 2017 Apr 22. https://doi.org/10.48550/arXiv.1704.06857

13. Xu Z, Sun J. Model-driven deep-learning. National Science Review. 2018; Jan 1; 5(1): 22-4. https://academic.oup.com/nsr/article/5/1/22/4093908

14. Chan HP, Samala RK, Hadjiiski LM, Zhou C. Deep learning in medical image analysis. Deep learning in medical image analysis: challenges and applications. 2020: 3-21. https://doi.org/10.1007/978-3-030-33128-3_1.

15. Shaikh Habeeba. Use of Artificial Intelligence in Drug Discovery and its Application in Drug Development. Asian Journal of Research in Chemistry. 2023; 16(1): 83-0.

16. Prasad Patil, Nripesh Kumar Nrip, Ashok Hajare, Digvijay Hajare, Mahadev K. Patil, Rajesh Kanthe, Anil T. Gaikwad. Artificial Intelligence and Tools in Pharmaceuticals: An Overview. Research Journal of Pharmacy and Technology. 2023; 16(4): 2075-2.

17. Mak KK, Pichika MR. Artificial intelligence in drug development: present status and future prospects. Drug Discovery Today. 2019: Mar 1; 24(3): 773-80. https://doi.org/10.1016/j.drudis.2018.11.014.

18. Bhushan S. Mahajan, Bhupendra Sing P. Mahale, Amol R. Pawar, Vikas V. Patil, Pankaj S. Patil, Jayesh Songire. A Review on Artificial Intelligence in Pharmacy. Research Journal of Science and Technology. 2024; 16(2):129-6.

19. R. R. Kulkarni, P. S. Pawar. Artificial Intelligence in Pharmacy. Asian Journal of Pharmacy and Technology. 2023; 13(4): 304-6.

20. Schmidhuber J. Deep learning in neural networks: An overview. Neural Networks. 2015; Jan 1; 61: 85-117. https://doi.org/10.1016/j.neunet.2014.09.003

21. Ha D, Schmidhuber. J. World models. arXiv preprint arXiv:1803.10122. 2018; Mar 27. https://doi.org/10.48550/arXiv.1803.10122

22. Colombo S. Applications of artificial intelligence in drug delivery and pharmaceutical development. In Artificial Intelligence in Healthcare 2020 Jan 1 (pp. 85-116). Academic Press.https://doi.org/10.1016/B978-0-12-818438-7.00004-6

23. Lan Z, Chen M, Goodman S, Gimpel K, Sharma P, Soricut R. Albert: A lite bert for self-supervised learning of language representations. arXiv preprint arXiv:1909.11942. 2019; Sep 26. https://doi.org/10.48550/arXiv.1909.11942

24. Devlin J, Chang MW, Lee K, Toutanova K. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv: 1810.04805. 2018; Oct 11. https://doi.org/10.48550/arXiv.1810.04805

25. Conneau A, Khandelwal K, Goyal N, Chaudhary V, Wenzek G, Guzmán F, Grave E, Ott M, Zettlemoyer L, Stoyanov V. Unsupervised cross-lingual representation learning at scale. arXiv preprint arXiv: 1911.02116. 2019; Nov 5.https://doi.org/10.18653/v1/2020.acl-main.747

26. Jing L, Tian Y. Self-supervised visual feature learning with deep neural networks: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2020; May 4; 43(11): 4037-58. https://doi.org/10.48550/arXiv.1902.06162

27. Kolesnikov A, Zhai X, Beyer L. Revisiting self-supervised visual representation learning. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2019; (pp. 1920-1929). https://doi.org/10.48550/arXiv.1901.09005

28. He K, Fan H, Wu Y, Xie S, Girshick R. Momentum contrast for unsupervised visual representation learning. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2020; (pp. 9729-9738). https://doi.org/10.1109/CVPR42600.2020.00975

29. Goyal P, Caron M, Lefaudeux B, Xu M, Wang P, Pai V, Singh M, Liptchinsky V, Misra I, Joulin A, Bojanowski P. Self-supervised pretraining of visual features in the wild. arXiv preprint arXiv:2103.01988. 2021; Mar 2. https://doi.org/10.48550/arXiv.2103.01988.

30. Sahil Mahajan, Heemani Dave, Santosh Bothe, Debarshikar Mahpatra, Sandeep Sonawane, Sanjay Kshirsagar, Santosh Chhajed. Objective Monitoring of Cardiovascular Biomarkers using Artificial Intelligence (AI). Asian Journal of Pharmaceutical Research. 2022; 12(3): 229-4.

31. Peikari M, Salama S, Nofech-Mozes S, Martel AL. A cluster-then-label semi-supervised learning approach for pathology image classification. Scientific Reports. 2018; May 8; 8(1): 1-3. https://doi.org/10.1038/s41598-018-24876-0

32. Taleb A, Loetzsch W, Danz N, Severin J, Gaertner T, Bergner B, Lippert C. 3d self-supervised methods for medical imaging. Advances in Neural Information Processing Systems. 2020; 33: 18158-72.

33. Hatamizadeh A, Tang Y, Nath V, Yang D, Myronenko A, Landman B, Roth HR, Xu D. Unetr: Transformers for 3d medical image segmentation. InProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision 2022 (pp. 574-584). https://doi.org/10.1109/WACV51458.2022.00181

34. Chen L, Bentley P, Mori K, Misawa K, Fujiwara M, Rueckert D. Self-supervised learning for medical image analysis using image context restoration. Medical Image Analysis. 2019; Dec 1; 58: 101539. https://doi.org/10.1016/j.media.2019.101539.

35. Pardeep Kumar Sharma, Amit Sachdeva, Cherry Bhargava. Fuzzy logic: A tool to predict the Renal diseases. Research Journal of Pharmacy and Technology. 2021; 14(5): 2598-2.

36. Valdes G, Luna JM, Eaton E, Simone CB, Ungar LH, Solberg TD. MediBoost: a patient stratification tool for interpretable decision making in the era of precision medicine. Scientific Reports. 2016; Nov 30; 6(1): 37854. https://doi.org/10.1038/srep37854.

37. Mehrotra K, Mohan CK, Ranka S. Elements of Artificial Neural Networks. MIT Press; 1997.

38. Thakur A, Konde A. Fundamentals of neural networks. International Journal for Research in Applied Science and Engineering Technology. 2021; 9(VIII):407-26.

39. Thakur A, Rizvi H, Satish M. White-box cartoonization using an extended gan framework. arXiv preprint arXiv:2107.04551. 2021 Jul 9.https://doi.org/10.33564/IJEAST.2021.v05i12.049

40. Fine TL. Feedforward neural network methodology. Springer Science & Business Media; 2006 Apr 6.

41. Sanjay S. Patel, Sparsh A. Shah. Artificial Intelligence: Comprehensive Overview and its Pharma Application. Asian Journal of Pharmacy and Technology. 2022; 12(4):337-8

42. Sherstinsky A. Fundamentals of recurrent neural network (RNN) and long short-term memory (LSTM) network. Physica D: Nonlinear Phenomena. 2020; Mar 1; 404: 132306.https://doi.org/10.1016/j.physd.2019.132306

43. Vt SE, Shin YC. Radial basis function neural network for approximation and estimation of nonlinear stochastic dynamic systems. IEEE Transactions on Neural Networks. 1994; Jul; 5(4): 594-603. doi: 10.1109/72.298229

44. Pal SK, Mitra S. Multilayer perceptron, fuzzy sets, classifiaction. doi: 10.1109/72.159058

45. Happel BL, Murre JM. Design and evolution of modular neural network architectures. Neural networks. 1994; Jan 1; 7(6-7): 985-1004.https://doi.org/10.1016/S0893-6080(05)80155-8

46. Chiu CC, Sainath TN, Wu Y, Prabhavalkar R, Nguyen P, Chen Z, Kannan A, Weiss RJ, Rao K, Gonina E, Jaitly N. State-of-the-art speech recognition with sequence-to-sequence models. In2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2018 Apr 15 (pp. 4774-4778). IEEE. doi: 10.1109/ICASSP.2018.8462105.

47. Subramanian N, Yajnik A, Murthy RS. Artificial neural network as an alternative to multiple regression analysis in optimizing formulation parmaeters of cytarabine liposomes. AAPS PharmSciTech. 2004; Mar; 5: 11-9. https://doi.org/10.1208/pt050104

48. Takahara J, Takayama K, Isowa K, Nagai T. Multi-objective simultaneous optimization based on artificial neural network in a ketoprofen hydrogel formula containing O-ethylmenthol as a percutaneous absorption enhancer. International Journal of Pharmaceutics. 1997; Dec 8; 158(2): 203-10. https://doi.org/10.1016/S0378-5173(97)00260-3

49. Wu PC, Obata Y, Fujikawa M, Li CJ, Higashiyama K, Takayama K. Simultaneous optimization based on artificial neural networks in ketoprofen hydrogel formula containing O-ethyl-3-butylcyclohexanol as percutaneous absorption enhancer. Journal of Pharmaceutical Sciences. 2001; Aug 1; 90(8): 1004-14. https://doi.org/10.1002/jps.1053.

50. Shraddha Jain, Sanket Jain, Dr. Sujit Pillai, Rampal Singh Mandloi. Review article on Role of Artificial Intelligence in Radiology. Research Journal of Pharmacognosy and Phytochemistry. 2023; 15(3): 264-0

51. Plumb AP, Rowe RC, York P, Doherty C. The effect of experimental design on the modeling of a tablet coating formulation using artificial neural networks. European Journal of Pharmaceutical Sciences. 2002; Sep 1; 16(4-5): 281-8. https://doi.org/10.1016/S0928-0987(02)00112-4

52. Leane MM, Cumming I, Corrigan OI. The use of artificial neural networks for the selection of the most appropriate formulation and processing variables in order to predict the in vitro dissolution of sustained release minitablets. Aaps Pharmscitech. 2003; Jun; 4: 129-40. doi: 10.1208/pt040226

53. Chen Y, McCall TW, Baichwal AR, Meyer MC. The application of an artificial neural network and pharmacokinetic simulations in the design of controlled-release dosage forms. Journal of Controlled Release. 1999; May 1; 59(1): 33-41. https://doi.org/10.1016/S0168-3659(98)00171-0

54. Ibrić S, Jovanović M, Djurić Z, Parojčić J, Solomun L. The application of generalized regression neural network in the modeling and optimization of aspirin extended release tablets with Eudragit® RS PO as matrix substance. Journal of Controlled Release. 2002; Aug 21; 82(2-3): 213-22. https://doi.org/10.1016/S0168-3659(02)00044-5

55. Takayama K, Morva A, Fujikawa M, Hattori Y, Obata Y, Nagai T. Formula optimization of theophylline controlled-release tablet based on artificial neural networks. Journal of Controlled Release. 2000; Aug 10; 68(2): 175-86. https://doi.org/10.1016/S0168-3659(00)00248-0

56. Ibrić S, Jovanović M, Djurić Z, Parojčić J, Petrović SD, Solomun L, Stupar B. Artificial neural networks in the modeling and optimization of aspirin extended release tablets with Eudragit L 100 as matrix substance. Aaps Pharmscitech. 2003; Mar; 4: 62-70. https://doi.org/10.1208/pt040109

57. Bourquin J, Schmidli H, van Hoogevest P, Leuenberger H. Comparison of artificial neural networks (ANN) with classical modelling techniques using different experimental designs and data from a galenical study on a solid dosage form. European Journal of Pharmaceutical Sciences. 1998; Oct 1; 6(4): 287-300.https://doi.org/10.1016/S0928-0987(97)10025-2

58. Bourquin J, Schmidli H, van Hoogevest P, Leuenberger H. Pitfalls of artificial neural networks (ANN) modelling technique for data sets containing outlier measurements using a study on ixture properties of a direct compressed dosage form. European Journal of Pharmaceutical Sciences. 1998; Dec 1; 7(1): 17-28. https://doi.org/10.1016/S0928-0987(97)10027-6

59. Bozič DZ, Vrečer F, Kozjek F. Optimization of diclofenac sodium dissolution from sustained release formulations using an artificial neural network. European Journal of Pharmaceutical Sciences. 1997; May 1;5(3):163-9. https://doi.org/10.1016/S0928-0987(97)00273-X

60. Zolotariov E, Anwar J. Modelling properties of powders using artificial neural networks and regression: the case of limited data. Journal of Pharmacy and Pharmacology. 1998; Sep; 50(Supplement_9):190. https://doi.org/10.1111/j.2042-7158.1998.tb02390.x

61. Ibric S, Jovanovic M, Djuric Z, Parojcic J, Solomun L. Artificial neural networking (ANN) and modeling of powder flow. J Cont Release 2002;82:213e22. https://doi.org/10.1211/jpp.59.5.0017

62. Peh KK, Lim CP, Quek SS, Khoh KH. Use of artificial neural networks to predict drug dissolution profiles and evaluation of network performance using similarity factor. Pharmaceutical Research. 2000; Nov; 17: 1384-9. https://doi.org/10.1023/A:1007578321803

63. Pillay V, Danckwerts MP. Textural profiling and statistical optimization of crosslinked calcium‐alginate‐pectinate‐cellulose acetophthalate gelisphere matrices. Journal of Pharmaceutical Sciences. 2002; Dec; 91(12): 2559-70.https://doi.org/10.1002/jps.10251

64. Değim T, Hadgraft J, İlbasmiş S, Ozkan Y. Prediction of skin penetration using artificial neural network (ANN) modeling. Journal of Pharmaceutical Sciences. 2003; Mar; 92(3): 656-64. https://doi.org/10.1002/jps.10312

65. Kandimalla KK, Kanikkannan N, Singh M. Optimization of a vehicle mixture for the transdermal delivery of melatonin using artificial neural networks and response surface method. Journal of Controlled Release. 1999; Aug 27; 61(1-2): 71-82. https://doi.org/10.1016/S0168-3659(99)00107-8

66. Murtoniemi E, Yliruusi J, Kinnunen P, Merkku P, Leiviskä K. The advantages by the use of neural networks in modelling the fluidized bed granulation process. International Journal of Pharmaceutics. 1994; Aug 1; 108(2): 155-64. https://doi.org/10.1016/0378-5173(94)90327-1

67. Gasperlin M, Tusar L, Tusar M, Kristl J, Smid-Korbar J. Lipophilic semisolid emulsion systems: viscoelastic behaviour and prediction of physical stability by neural network modelling. International Journal of Pharmaceutics. 1998;168(2):243-54. Doi:10.1016/S0378-5173(98)00099-4

Received on 23.10.2024 Revised on 09.12.2024

Accepted on 18.01.2025 Published on 03.03.2025

Available online from March 10, 2025

Res. J. Pharma. Dosage Forms and Tech.2025; 17(1):31-36.

DOI: 10.52711/0975-4377.2025.00005

This work is licensed under a Creative Commons Attribution-Non Commercial-Share Alike 4.0 International License. Creative Commons License.

Time of Supervision		Completeness of supervision
		Complete	Partial	Absent
	Anticipate	Reinforcement learning Delayed rewards from the environment
	Immediate	Supervised learning Complete, immediate supervision Ensemble learning: community of models	Semi-supervised learning Incomplete, immediate supervision	Unsupervised learning No supervision at all, only input data
	Recycle	Transfer learning Pretrain model by reusing results of past similar supervised problem	Self-supervised learning Pretrain model on dummy supervised task created from input data