Optuna

{{Short description|Hyperparameter optimization framework}} {{copyedit|date=July 2025}} {{Infobox software | name = Optuna | logo = Optuna-logo.png | logo size = | screenshot size = | caption = Optuna | developer = [[Preferred Networks]] | released = Version 1.0 (January, 2020) | latest release version = Version 4.4.0 (June 16, 2025) | programming language = [[Python (programming language)|Python]] | platform = | repo = {{URL|https://github.com/optuna/optuna}} | language = English | genre = | license = [[MIT License]] | website = {{URL|https://optuna.org/}} }}

'''Optuna''' is an [[Open source|open-source]] [[Python (programming language)|Python]] library for automatic [[Hyperparameter (machine learning)|hyperparameter]] tuning of [[machine learning]] models.<ref name=":1" /> It was first introduced in 2018 by [[Preferred Networks]], a Japanese startup that works on practical applications of [[deep learning]] in various fields.<ref name=":3">{{Cite journal |last1=Wang |first1=Zhaofei |last2=Li |first2=Hao |last3=Wang |first3=Qiuping |date=2025 |title=An Approach to Truck Driving Risk Identification: A Machine Learning Method Based on Optuna Optimization |journal=IEEE Access |volume=13 |pages=42723–42732 |doi=10.1109/ACCESS.2025.3547445 |bibcode=2025IEEEA..1342723W |issn=2169-3536|doi-access=free }}</ref><ref name=":4">{{Cite journal |last1=Li |first1=Yifan |last2=Cao |first2=Yanpeng |last3=Yang |first3=Jintang |last4=Wu |first4=Mingyu |last5=Yang |first5=Aimin |last6=Li |first6=Jie |date=2024-06-01 |title=Optuna-DFNN: An Optuna framework driven deep fuzzy neural network for predicting sintering performance in big data |journal=Alexandria Engineering Journal |volume=97 |pages=100–113 |doi=10.1016/j.aej.2024.04.026 |issn=1110-0168|doi-access=free }}</ref><ref name=":5">{{Cite journal |last1=Pinichka |first1=Chayut |last2=Chotpantarat |first2=Srilert |last3=Cho |first3=Kyung Hwa |last4=Siriwong |first4=Wattasit |date=2025-07-01 |title=Comparative analysis of SWAT and SWAT coupled with XGBoost model using Optuna hyperparameter optimization for nutrient simulation: A case study in the Upper Nan River basin, Thailand |journal=Journal of Environmental Management |volume=388 |article-number=126053 |doi=10.1016/j.jenvman.2025.126053 |pmid=40460531 |bibcode=2025JEnvM.38826053P |issn=0301-4797|doi-access=free }}</ref><ref name=":2" /> The beta version of Optuna was released at the end of the year, with the subsequent first major stable release announced in January 2020.<ref name=":1">{{Cite book |last1=Akiba |first1=Takuya |last2=Sano |first2=Shotaro |last3=Yanase |first3=Toshihiko |last4=Ohta |first4=Takeru |last5=Koyama |first5=Masanori |chapter=Optuna: A Next-generation Hyperparameter Optimization Framework |date=2019-07-25 |title=Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining |chapter-url=https://dl.acm.org/doi/10.1145/3292500.3330701 |series=KDD '19 |location=New York, NY, USA |publisher=Association for Computing Machinery |pages=2623–2631 |doi=10.1145/3292500.3330701 |isbn=978-1-4503-6201-6}}</ref>

==Hyperparameter optimization== {{See also|Hyperparameter optimization}} [[File:Hyperparameter Optimization using Grid Search.svg|thumb|upright=1.2|Hyperparameter Optimization using Grid Search]] Hyperparameter optimization involves finding the optimal value of non-trainable parameters, defined by the user. Examples of hyperparameters are [[learning rate]], number of layers or neurons, regularization strength, and tree depth. However, they strongly depend on the specific algorithm (e.g., [[classification]], [[Regression analysis|regression]], [[Cluster analysis|clustering]], etc.).<ref>{{Cite journal |last1=Yang |first1=Li |last2=Shami |first2=Abdallah |date=2020-11-20 |title=On hyperparameter optimization of machine learning algorithms: Theory and practice |url=https://www.sciencedirect.com/science/article/pii/S0925231220311693 |journal=Neurocomputing |volume=415 |pages=295–316 |doi=10.1016/j.neucom.2020.07.061 |arxiv=2007.15745 |issn=0925-2312}}</ref>

Hyperparameter optimization can be especially relevant when dealing with large-scale problems or limited resources, as it improves model accuracy, reduces overfitting, and decreases training time.<ref name=":1" /> However, when the hyperparameter space increases, it may become computationally expensive. Hence, there are methods (e.g., [[grid search]], [[random search]], or [[bayesian optimization]]) that considerably simplify this process.<ref name=":1" />

==Features== Optuna is designed to optimize the model hyperparameters by searching large spaces and discarding combinations that show no significant improvements in the model. Moreover, it can parallelize the hyperparameter search over multiple threads or processes. Optuna works with a high degree of modularity, allowing the definition of different configurations for searching hyperparameters. It is possible to choose which hyperparameters to optimize and how to optimize them. Additionally, it permits parameter customization at runtime, meaning that the values to explore for the hyperparameters (i.e., the search space) can be defined while writing the code, rather than being defined in advance.<ref name=":1" /> [[File:Hyperparameter Optimization using Tree-Structured Parzen Estimators.svg|thumb|upright=1.2|Hyperparameter Optimization using Tree-Structured Parzen Estimators]]

===Sampling=== Optuna exploits well-established algorithms to perform hyperparameter optimization, progressively reducing the search space in light of objective values. Examples are gaussian-process-based algorithms (i.e., a gaussian process to model the objective function<ref>{{Cite journal |last1=Tanim |first1=Ahad Hasan |last2=Smith-Lewis |first2=Corinne |last3=Downey |first3=Austin R. J. |last4=Imran |first4=Jasim |last5=Goharian |first5=Erfan |date=2024-08-01 |title=Bayes_Opt-SWMM: A Gaussian process-based Bayesian optimization tool for real-time flood modeling with SWMM |url=https://www.sciencedirect.com/science/article/pii/S136481522400183X |journal=Environmental Modelling & Software |volume=179 |article-number=106122 |doi=10.1016/j.envsoft.2024.106122 |bibcode=2024EnvMS.17906122T |osti=2378088 |issn=1364-8152}}</ref>), [[Tree-structured Parzen estimators|tree-structured parzen estimator]] (TPE) (i.e., a model-based optimization method that estimates the objective function and selects the best hyperparameters<ref>{{Cite journal |last1=Ozaki |first1=Yoshihiko |last2=Tanigaki |first2=Yuki |last3=Watanabe |first3=Shuhei |last4=Nomura |first4=Masahiro |last5=Onishi |first5=Masaki |date=2022-04-08 |title=Multiobjective Tree-Structured Parzen Estimator |url=https://www.jair.org/index.php/jair/article/view/13188 |journal=Journal of Artificial Intelligence Research |language=en |volume=73 |pages=1209–1250 |doi=10.1613/jair.1.13188 |issn=1076-9757|doi-access=free }}</ref>), and [[random search]] (i.e., a basic optimization approach used for benchmarking<ref>{{Cite journal |last1=Bergstra |first1=James |last2=Bengio |first2=Yoshua |date=2012-02-01 |title=Random search for hyper-parameter optimization |url=https://dl.acm.org/doi/abs/10.5555/2188385.2188395 |journal=J. Mach. Learn. Res. |volume=13 |pages=281–305 |issn=1532-4435}}</ref>).

===Early stopping=== [[File:GpParBayesAnimationSmall.gif|thumb|upright=1.2|Bayesian optimization of a function (black) with a gaussian process (purple). Three acquisition functions (blue).]] Optuna includes a pruning feature to stop trials early when the results show no significant improvement in the model. This allows for the prevention of unnecessary computation, and it is used for models with long training times in order to save time and computational resources. Specifically, Optuna exploits techniques such as median and threshold [[pruning]].<ref>{{Cite book |last1=Wistuba |first1=Martin |last2=Schilling |first2=Nicolas |last3=Schmidt-Thieme |first3=Lars |chapter=Hyperparameter Search Space Pruning – A New Component for Sequential Model-Based Hyperparameter Optimization |series=Lecture Notes in Computer Science |date=2015 |volume=9285 |editor-last=Appice |editor-first=Annalisa |editor2-last=Rodrigues |editor2-first=Pedro Pereira |editor3-last=Santos Costa |editor3-first=Vítor |editor4-last=Gama |editor4-first=João |editor5-last=Jorge |editor5-first=Alípio |editor6-last=Soares |editor6-first=Carlos |title=Machine Learning and Knowledge Discovery in Databases |chapter-url=https://link.springer.com/chapter/10.1007/978-3-319-23525-7_7 |language=en |location=Cham |publisher=Springer International Publishing |pages=104–119 |doi=10.1007/978-3-319-23525-7_7 |isbn=978-3-319-23525-7}}</ref><ref>{{Cite book |last1=Akiba |first1=Takuya |last2=Sano |first2=Shotaro |last3=Yanase |first3=Toshihiko |last4=Ohta |first4=Takeru |last5=Koyama |first5=Masanori |chapter=Optuna: A Next-generation Hyperparameter Optimization Framework |date=2019-07-25 |title=Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining |chapter-url=https://dl.acm.org/doi/10.1145/3292500.3330701 |series=KDD '19 |location=New York, NY, USA |publisher=Association for Computing Machinery |pages=2623–2631 |doi=10.1145/3292500.3330701 |isbn=978-1-4503-6201-6}}</ref>

===Scalability=== Optuna is designed to scale with distributed computing resources, supporting [[Parallel execution units|parallel execution]]. This feature allows users to run optimization trials across multiple processors or machines.<ref>{{Cite book |last1=Parra-Ullauri |first1=Juan |last2=Zhang |first2=Xunzheng |last3=Bravalheri |first3=Anderson |last4=Nejabati |first4=Reza |last5=Simeonidou |first5=Dimitra |chapter=Federated Hyperparameter Optimisation with Flower and Optuna |date=2023-06-07 |title=Proceedings of the 38th ACM/SIGAPP Symposium on Applied Computing |chapter-url=https://dl.acm.org/doi/10.1145/3555776.3577847 |series=SAC '23 |location=New York, NY, USA |publisher=Association for Computing Machinery |pages=1209–1216 |doi=10.1145/3555776.3577847 |isbn=978-1-4503-9517-5 |url=https://research-information.bris.ac.uk/en/publications/49495d0b-4f4e-4d9e-a81e-ef18ed54ea8c }}</ref>

===Integration with third-part libraries=== Optuna integrates with various ML libraries and frameworks:<ref name=":1" /> * [[CatBoost|Catboost]] * [[Dask (software)|Dask]] * [[Fast.ai]] * [[Keras]] * [[LightGBM]] * MLflow * [[PyTorch]] * PyTorch Ignite * [[PyTorch Lightning]] * [[TensorBoard]] * [[TensorFlow]] * tf.keras * Weights & Biases * [[XGBoost]] * skrub Moreover, Optuna offers a real-time dashboard that allows a user to monitor, through graphs and tables, the optimization history and the hyperparameter importance.<ref>{{Cite journal |last1=Almarzooq |first1=Hussain |last2=Waheed |first2=Umair bin |date=2024-05-21 |title=Automating hyperparameter optimization in geophysics with Optuna: A comparative study |url=https://www.earthdoc.org/content/journals/10.1111/1365-2478.13484 |journal=Geophysical Prospecting |language=en |volume=72 |issue=5 |pages=1778–1788 |doi=10.1111/1365-2478.13484 |bibcode=2024GeopP..72.1778A |issn=1365-2478}}</ref>

==Applications== Optuna was designed to be framework agnostic so that it can be used with any machine-learning (ML) or deep-learning (DL) framework.<ref name=":1" /> [[File:Standard pipeline for the optimization of hyperparameters for the training of a Decision Tree Classfier using Optuna.png|thumb|upright=2|Standard pipeline for the optimization of hyperparameters for the training of a Decision Tree Classfier using Optuna]]

===Machine learning=== {{Overly detailed|section|date=July 2025}} Optuna can be used to optimize hyperparamenters of ML models. Examples are: *[[Random forest]]: number of trees, maximum depth, and minimum samples per leaf.<ref>{{Citation |last1=Rokach |first1=Lior |title=Decision Trees |date=2005 |work=Data Mining and Knowledge Discovery Handbook |pages=165–192 |editor-last=Maimon |editor-first=Oded |place=Boston, MA |publisher=Springer US |language=en |doi=10.1007/0-387-25465-x_9 |isbn=978-0-387-25465-4 |last2=Maimon |first2=Oded |editor2-last=Rokach |editor2-first=Lior}}</ref>

*[[Gradient Boosting Machines|Gradient boosting machines]] (GBM): learning rate, number of estimators, and maximum depth.<ref>{{Cite book |last1=R |first1=Shyam |last2=Ayachit |first2=Sai Sanjay |last3=Patil |first3=Vinayak |last4=Singh |first4=Anubhav |chapter=Competitive Analysis of the Top Gradient Boosting Machine Learning Algorithms |date=2020-12-18 |title=2020 2nd International Conference on Advances in Computing, Communication Control and Networking (ICACCCN) |pages=191–196 |doi=10.1109/ICACCCN51052.2020.9362840 |isbn=978-1-7281-8337-4 }}</ref>

*[[Support vector machine|Support vector machines]] (SVM): regularization parameter (C), kernel type (e.g., linear, radial basis function), and gamma (gamma).<ref>{{Citation |last1=Pisner |first1=Derek A. |title=Chapter 6 - Support vector machine |date=2020-01-01 |work=Machine Learning |pages=101–121 |editor-last=Mechelli |editor-first=Andrea |url=https://www.sciencedirect.com/science/article/pii/B9780128157398000067 |access-date=2025-07-08 |publisher=Academic Press |isbn=978-0-12-815739-8 |last2=Schnyer |first2=David M. |editor2-last=Vieira |editor2-first=Sandra}}</ref>

*[[K-nearest neighbors algorithm|K-nearest neighbors]] (KNN): number of neighbors (k), distance metrics (e.g., Euclidean or Manhattan), and weight function.<ref>{{Cite journal |last=Zhang |first=Zhongheng |date=2016-06-14 |title=Introduction to machine learning: k-nearest neighbors |journal=Annals of Translational Medicine |volume=4 |issue=11 |page=218 |doi=10.21037/atm.2016.03.37 |doi-access=free |issn=2305-5839 |pmc=4916348 |pmid=27386492}}</ref>

*[[Linear regression|Linear]] and [[Logistic regression|logistic]] regression: alpha in Ridge Regression or C in Logistic Regression.<ref>{{Cite journal |last1=Tripepi |first1=G. |last2=Jager |first2=K. J. |last3=Dekker |first3=F. W. |last4=Zoccali |first4=C. |date=2008-04-01 |title=Linear and logistic regression analysis |url=https://www.sciencedirect.com/science/article/pii/S0085253815530895 |journal=Kidney International |volume=73 |issue=7 |pages=806–810 |doi=10.1038/sj.ki.5002787 |pmid=18200004 |issn=0085-2538|url-access=subscription }}</ref>

*[[Naive bayes classifier|Naive Bayes]]: smoothing coefficients.<ref>{{Cite book |last=Yang |first=Feng-Jen |chapter=An Implementation of Naive Bayes Classifier |date=2018-12-12 |title=2018 International Conference on Computational Science and Computational Intelligence (CSCI) |pages=301–306 |doi=10.1109/CSCI46756.2018.00065 |isbn=978-1-7281-1360-9 }}</ref>

===Deep learning=== In the context of deep learning, Optuna can be exploited during the training of [[Neural network|neural networks]] (NN) to optimize learning rate, batch size, and the number of hidden layers. For example, it can be used for:

*[[Convolutional neural network|Convolutional neural networks]] (CNNs), for image classification, object detection, and semantic-segmentation tasks.<ref>{{Cite journal |last1=Li |first1=Zewen |last2=Liu |first2=Fan |last3=Yang |first3=Wenjie |last4=Peng |first4=Shouheng |last5=Zhou |first5=Jun |date=2021-06-10 |title=A Survey of Convolutional Neural Networks: Analysis, Applications, and Prospects |journal=IEEE Transactions on Neural Networks and Learning Systems |volume=33 |issue=12 |pages=6999–7019 |doi=10.1109/TNNLS.2021.3084827 |pmid=34111009 |hdl=10072/405164 |issn=2162-2388|hdl-access=free }}</ref>

*[[Recurrent neural network|Recurrent neural networks]] (RNNs), for sequence-based tasks such as time-series forecasting and natural language processing.<ref>{{Citation |last1=Caterini |first1=Anthony L. |title=Recurrent Neural Networks |date=2018-03-23 |work=Deep Neural Networks in a Mathematical Framework |pages=59–79 |editor-last=Caterini |editor-first=Anthony L. |place=Cham |publisher=Springer International Publishing |language=en |doi=10.1007/978-3-319-75304-1_5 |isbn=978-3-319-75304-1 |last2=Chang |first2=Dong Eui |editor2-last=Chang |editor2-first=Dong Eui}}</ref>

*[[Transformers]], for NLP tasks such as text classification, sentiment analysis, and machine translation.<ref>{{Cite journal |last1=Gillioz |first1=Anthony |last2=Casas |first2=Jacky |last3=Mugellini |first3=Elena |last4=Khaled |first4=Omar Abou |date=2020-09-06 |title=Overview of the Transformer-based Models for NLP Tasks |journal=2020 15th Conference on Computer Science and Information Systems (FedCSIS) |series=Proceedings of the 2020 Federated Conference on Computer Science and Information Systems |volume=21 |pages=179–183 |doi=10.15439/2020F20 |isbn=978-83-955416-7-4 |doi-access=free }}</ref>

==Domains== Optuna has found applications in various research studies and industry implementations across different applicative domains.<ref name=":3" /><ref name=":4" /><ref name=":5" /><ref name=":2" />

===Healthcare=== In [[Health care|healthcare]], Optuna is currently utilized for [[medical image analysis]] to optimize DL models for tumor detection,<ref>{{Cite book |last1=Bhuvanya |first1=R. |last2=T.Kujani |last3=Kumaran |first3=S.Manoj |last4=Lokesh Kumar |first4=N. |chapter=OptNet: Innovative Model for Early Lung Cancer Diagnosis integrating TabNet and Optuna |date=2024-11-22 |title=2024 International Conference on Electronic Systems and Intelligent Computing (ICESIC) |pages=174–179 |doi=10.1109/ICESIC61777.2024.10846378 |isbn=979-8-3315-2298-8 }}</ref><ref>{{Cite journal |last1=Kumar Sahu |first1=Prabhat |last2=Fatma |first2=Taiyaba |date=2025-02-07 |title=Optimized Breast Cancer Classification Using PCA-LASSO Feature Selection and Ensemble Learning Strategies With Optuna Optimization |journal=IEEE Access |volume=13 |pages=35645–35661 |doi=10.1109/ACCESS.2025.3539746 |bibcode=2025IEEEA..1335645K |issn=2169-3536|doi-access=free }}</ref> disease identification,<ref>{{Citation |last1=Deepan |first1=P. |title=Maximizing Accuracy in Alzheimer's Disease Prediction: A Optuna Hyper Parameter Optimization Strategy Using MRI Images |date=2024 |work=Revolutionizing Healthcare 5.0: The Power of Generative AI: Advancements in Patient Care Through Generative AI Algorithms |pages=77–91 |editor-last=Bhattacharya |editor-first=Pronaya |place=Cham |publisher=Springer Nature Switzerland |language=en |doi=10.1007/978-3-031-75771-6_5 |isbn=978-3-031-75771-6 |last2=Prabhakar Reddy |first2=G. |last3=Arsha Reddy |first3=M. |last4=Vidya |first4=R. |last5=Dhiravidaselvi |first5=S. |editor2-last=Liu |editor2-first=Haipeng |editor3-last=Dutta |editor3-first=Pushan Kumar |editor4-last=Rodrigues |editor4-first=Joel J. P. C.}}</ref> multi-organ semantic segmentation,<ref>{{Cite journal |last=Vaiyapuri |first=Thavavel |date=2025-08-01 |title=An Optuna-Based Metaheuristic Optimization Framework for Biomedical Image Analysis |url=https://etasr.com/index.php/ETASR/article/view/11234 |journal=Engineering, Technology & Applied Science Research |language=en |volume=15 |issue=4 |pages=24382–24389 |doi=10.48084/etasr.11234 |issn=1792-8036|doi-access=free }}</ref> and radiological image analysis. It can be also implemented for [[disease prediction]], in particular for the improvement in the accuracy of predictive models for disease diagnosis and treatment outcomes.<ref>{{Cite journal |last1=Srinivas |first1=Polipireddy |last2=Katarya |first2=Rahul |date=2022-03-01 |title=hyOPTXg: OPTUNA hyper-parameter optimization framework for predicting cardiovascular disease using XGBoost |url=https://www.sciencedirect.com/science/article/pii/S1746809421010533 |journal=Biomedical Signal Processing and Control |volume=73 |article-number=103456 |doi=10.1016/j.bspc.2021.103456 |issn=1746-8094|url-access=subscription }}</ref> Its application can be also found in [[genomics]] to predict genetic diseases and to identify genetic variations.<ref>{{Cite journal |last1=Shen |first1=Zijie |last2=Shen |first2=Enhui |last3=Zhu |first3=Qian-Hao |last4=Fan |first4=Longjiang |last5=Zou |first5=Quan |last6=Ye |first6=Chu-Yu |date=2023 |title=GSCtool: A Novel Descriptor that Characterizes the Genome for Applying Machine Learning in Genomics |journal=Advanced Intelligent Systems |language=en |volume=5 |issue=12 |article-number=2300426 |doi=10.1002/aisy.202300426 |issn=2640-4567|doi-access=free }}</ref>

===Finance=== In finance, Optuna is used to optimize models for [[algorithmic trading]]. It allows for predicting market movements due to its ability to handle wide parameter ranges and complex model structures.<ref>{{Cite book |last1=Shui |first1=Hongyi |last2=Sha |first2=Xinye |last3=Chen |first3=Baizheng |last4=Wu |first4=Jiajie |chapter=Stock weighted average price prediction based on feature engineering and Lightgbm model |date=2024-08-26 |title=Proceedings of the 2024 International Conference on Digital Society and Artificial Intelligence |chapter-url=https://dl.acm.org/doi/10.1145/3677892.3677945 |series=DSAI '24 |location=New York, NY, USA |publisher=Association for Computing Machinery |pages=336–340 |doi=10.1145/3677892.3677945 |isbn=979-8-4007-0983-8}}</ref><ref name=":0">{{Cite book |last1=Garg |first1=Deepak |last2=Shelke |first2=Nitin Arvind |last3=Kitukale |first3=Gauri |last4=Mehlawat |first4=Nishka |chapter=Leveraging Financial Data and Risk Management in Banking sector using Machine Learning |date=2024-04-05 |title=2024 IEEE 9th International Conference for Convergence in Technology (I2CT) |pages=1–6 |doi=10.1109/I2CT61223.2024.10544336 |isbn=979-8-3503-9445-0 }}</ref> In particular, it is exploited for financial risk analysis and forecasting. Topics addressed are credit, market, and operational risks.<ref name=":0" />

===Autonomous systems=== Optuna is used for real-time applications in autonomous systems for [[robotics]],<ref name=":2">{{Cite journal |last1=Huang |first1=Jiancong |last2=Rojas |first2=Juan |last3=Zimmer |first3=Matthieu |last4=Wu |first4=Hongmin |last5=Guan |first5=Yisheng |last6=Weng |first6=Paul |date=2021-03-08 |title=Hyperparameter Auto-Tuning in Self-Supervised Robotic Learning |journal=IEEE Robotics and Automation Letters |volume=6 |issue=2 |pages=3537–3544 |doi=10.1109/LRA.2021.3064509 |arxiv=2010.08252 |bibcode=2021IRAL....6.3537H |issn=2377-3766}}</ref> supporting decision making in dynamic environments. It is also used in the context of [[Self-driving car|self-driving cars]] to optimize the model to navigate safely in complex environments. For example, Optuna can be used in scenarios where there is the need to evaluate the rate and the severity of accidents<ref>{{Cite journal |last1=Rezashoar |first1=Soheil |last2=Kashi |first2=Ehsan |last3=Saeidi |first3=Soheila |date=2024-07-26 |title=A hybrid algorithm based on machine learning (LightGBM-Optuna) for road accident severity classification (case study: United States from 2016 to 2020) |journal=Innovative Infrastructure Solutions |language=en |volume=9 |issue=8 |page=319 |doi=10.1007/s41062-024-01626-y |bibcode=2024InnIS...9..319R |issn=2364-4184}}</ref> or to address the issue of network intrusion attacks due to potential vulnerabilities.<ref>{{Cite book |last1=Jha |first1=Jayesh |last2=Yadav |first2=Jatin |last3=Naqvi |first3=Haider |chapter=MILCCDE: A Metaheuristic Improved Decision-Based Ensemble Framework for Intrusion Detection in Autonomous Vehicles |series=Lecture Notes in Networks and Systems |date=2025 |volume=1295 |editor-last=Bansal |editor-first=Jagdish Chand |editor2-last=Jamwal |editor2-first=Prashant K. |editor3-last=Hussain |editor3-first=Shahid |title=Sustainable Computing and Intelligent Systems|chapter-url=https://link.springer.com/chapter/10.1007/978-981-96-3311-1_21 |language=en |location=Singapore |publisher=Springer Nature |pages=255–267 |doi=10.1007/978-981-96-3311-1_21 |isbn=978-981-96-3311-1}}</ref><ref>{{Cite book |last1=Parekh |first1=Nishank |last2=Sen |first2=Arzob |last3=Rajasekaran |first3=P. |last4=Jayaseeli |first4=J. D. Dorathi |last5=Robert |first5=P. |chapter=Network Intrusion Detection System Using Optuna |date=2024-12-17 |title=2024 International Conference on IoT Based Control Networks and Intelligent Systems (ICICNIS) |pages=312–318 |doi=10.1109/ICICNIS64247.2024.10823298 |isbn=979-8-3315-1809-7 }}</ref>

===Natural language processing (NLP)=== Optuna has also been applied in NLP. For example, it has been used for [[Text Classification|text classification]] to classify a piece of written text into categories (e.g., spam vs. not spam, topic modeling).<ref>{{Cite journal |last1=Rahmi |first1=Nadya Alinda |last2=Defit |first2=Sarjon |last3=Okfalisa |date=2024-12-31 |title=The Use of Hyperparameter Tuning in Model Classification: A Scientific Work Area Identification |url=http://joiv.org/index.php/joiv/article/view/3092 |journal=International Journal on Informatics Visualization |volume=8 |issue=4 |page=2181 |doi=10.62527/joiv.8.4.3092 |issn=2549-9904|doi-access=free }}</ref> Another task is [[sentiment analysis]], namely the detection of feelings expressed in text, particularly exploited for the analysis of content from social media and customer reviews.<ref>{{Cite book |last1=Efendi |first1=Akmar |last2=Fitri |first2=Iskandar |last3=Nurcahyo |first3=Gunadi Widi |chapter=Improvement of Machine Learning Algorithms with Hyperparameter Tuning on Various Datasets |date=2024-08-07 |title=2024 International Conference on Future Technologies for Smart Society (ICFTSS) |pages=75–79 |doi=10.1109/ICFTSS61109.2024.10691354 |isbn=979-8-3503-7384-4 }}</ref>

===Reinforcement learning (RL)=== For what concerns RL, Optuna is utilized in the [[Gaming computer|gaming]] field to improve the model performance in games and virtual environments. In [[robotics]] it has been used to optimize decision-making processes in robotic systems for tasks like manipulation and navigation. Additionally, in the field of [[autonomous vehicles]] Optuna can optimize RL models to obtain enhanced safety, and a more efficient navigation strategy.<ref>{{Cite journal |last1=Kiran |first1=B Ravi |last2=Sobh |first2=Ibrahim |last3=Talpaert |first3=Victor |last4=Mannion |first4=Patrick |last5=Sallab |first5=Ahmad A. Al |last6=Yogamani |first6=Senthil |last7=Pérez |first7=Patrick |date=2021-02-09 |title=Deep Reinforcement Learning for Autonomous Driving: A Survey |journal=IEEE Transactions on Intelligent Transportation Systems |volume=23 |issue=6 |pages=4909–4926 |doi=10.1109/TITS.2021.3054625 |arxiv=2002.00444 |bibcode=2022ITITr..23.4909K |issn=1558-0016}}</ref>

==See also== *[[Hyperparameter tuning]] *[[Bayesian optimization]] *[[Machine learning]] *[[Deep learning]]

==References== <references />

==Further reading== *{{Cite journal |last1=Wang |first1=Xilu |last2=Jin |first2=Yaochu |last3=Schmitt |first3=Sebastian |last4=Olhofer |first4=Markus |date=2023-07-13 |title=Recent Advances in Bayesian Optimization |url=https://dl.acm.org/doi/10.1145/3582078 |journal=ACM Comput. Surv. |volume=55 |issue=13s |pages=287:1–287:36 |doi=10.1145/3582078 |issn=0360-0300|url-access=subscription }} *{{Cite book |last1=Shinde |first1=Pramila P. |last2=Shah |first2=Seema |chapter=A Review of Machine Learning and Deep Learning Applications |date=2018-08-16 |title=2018 Fourth International Conference on Computing Communication Control and Automation (ICCUBEA) |pages=1–6 |doi=10.1109/ICCUBEA.2018.8697857 |isbn=978-1-5386-5257-2 }}

==External links== * {{Official website|https://optuna.org/}} * {{GitHub|Optuna}} * [https://optuna.readthedocs.io/en/stable/ Optuna Documentation] * [https://optuna.readthedocs.io/en/stable/tutorial/ Tutorial] * [https://github.com/optuna/optuna-dashboard/ Optuna Dashboard] * [https://optuna-dashboard.readthedocs.io/en/stable/getting-started.html/ Optuna Dashboard Documentation]

[[Category:Software using the MIT license]] [[Category:Deep learning software]] [[Category:Python (programming language) libraries]]