Multi-Sigma Version History

v2.2.5+1 2026-02-02 #

#General Features: 

・A new feature to share tasks between users or with all users will be deployed. 

・A space-filling  method will be added to the experimental design methods. 

・A new feature of visualization of results will be added. 

・An issue in the calculation of EI and PI when performing optimization using Gaussian Process Regression will be fixed. 

v2.2.5+0 2025-11-27 #

#General Features:

・Python version was updated from 3.8 to 3.10.

・The performance of hyperparameter tuning was improved in Bayesian optimization.

・Two-factor and multi-factor authentication were introduced.

・In the experiment-design feature, Latin hypercube sampling was added, along with a new function that can generate experimental conditions based on past experimental data.

・To address precision loss, numerical values will be displayed in exponential notation instead of decimal notation.

v2.2.4+3 2025-10-01 #

#General Features:

・When the browser locale is set to English, the registration email, Privacy Policy, Terms of Service, and White Paper will be provided in English.

v2.2.3+2 2025-06-18 #

# General Features:

・New functions for data transformation (Quantile transformation, Log, and Log1p) will be added.

・It will become possible to run Profile function repeatedly and download the results without reloading the browser.

・Some wording in Multi-Sigma and the startup guide will be revised.

v2.2.3+1 2025-05-19 #

# General Features:

・Reduce analysis time based on the type of data preprocessing.

・Users will be able to confirm all values for explanatory variables and objective variables in prediction and optimization plots.

・Users will be able to confirm project name, task name, and analysis method on the analysis page.

・Users will be able to download the result of validation and generate scatter plots comparing predicted and actual values at the validation stage of AI prediction.

・Users will be able to access the startup guide and help after logging in Multi-Sigma.

v2.2.3 2025-04-04 #

# General Features:

・Header information is now added to the “Normalized CSV” files generated after preprocessing.

・We have fixed an issue that prevented the upload of training data when using the Safari browser.

v2.2.2 2025-02-28 #

# General Features:

・You will be able to download the results of the Monte Carlo simulation as a CSV file.

・The threshold values will be displayed in the Violin plot and ROC curve when the output is either 0 or 1.

v2.2.1 2024-11-15 #

# General Features:

The selection of analysis methods will be made when creating a task, rather than when creating a project. This allows you to use various analysis methods within a single project.

v2.2.0 2024-10-18 #

# General Features:

– The feature for missing value interpolation will be available.

– You will be able to evaluate errors using multiple methods in the error validation function for your test data.

v2.1.0 2024-07-31 #

# General Features:

– Chain analysis combining AI will be available.

– Monte Carlo simulation analysis becomes optional.

– In Bayesian analysis, setting the kernel and hyperparameters for Gaussian process regression will be possible.

v2.0.14 2024-06-13 #

v2.0.13 2024-04-08 #

v2.0.12 2024-03-11 #

v2.0.11 2024-02-21 #

v2.0.10 2024-02-02 #

v2.0.9 2023-12-15 #

v2.0.8 2023-11-14 #

v2.0.7 2023-11-08 #

v2.0.0 2023-08-22 #

Multi-Sigma V2 release