Simple ML is an AI-powered platform that provides automatic machine learning model generation and deployment for businesses and individuals. Its aim is to simplify the process of developing and deploying machine learning models, irrespective of one's expertise in coding and data science.
Its AutoML functionality offers data cleaning, data exploration, feature scaling and selection, best model selection, hyperparameter tuning, and model deployment, along with an intuitive user interface that simplifies the process of creating custom models. Simple ML integrates seamlessly with popular applications such as Excel, Google Sheets, Google Slides, and Microsoft Power BI, and offers unique features such as Jupyter Notebook integration, API access, workflow adjustment, and model training, ensuring users have control over the machine learning model development process.
The platform offers pre-built solutions for various business domains, such as sales forecasting, churn analysis, image classification, text classification, ensuring cost-effective custom machine learning solutions.
TLDR
Simple ML offers an AI-powered platform that provides automatic machine learning model generation and deployment. Its intuitive AutoML functionality simplifies the process of developing and deploying custom models, irrespective of one's expertise in coding and data science, along with popular application integrations such as Excel, Google Sheets, Google Slides, and Microsoft Power BI.
The platform provides advanced features such as Jupyter Notebook integration, API access, workflow adjustment, and unique model training, offering users full control over the machine learning model development process. Additionally, Simple ML offers pre-built solutions for various business domains at cost-effective prices, such as sales forecasting, churn analysis, image classification, and text classification.
Company Overview
Simple ML is an AI-powered platform that offers automatic machine learning model generation and deployment for businesses and individuals. The platform is designed to simplify the process of developing and deploying machine learning models without requiring high levels of expertise in coding and data science. Simple ML integrates easily with other popular applications, including Excel, Google Sheets, Google Slides, and Microsoft Power BI.
The platform's main focus is on reducing the time needed to develop and deploy machine learning models. Simple ML, therefore, offers an automated machine learning (AutoML) service that handles the entire process of data cleaning, feature selection, training, and deployment of the best model.
The generated models are designed to be fair, explainable, and auditable to meet the ethical AI criteria. Simple ML also provides an intuitive user interface that simplifies the process of creating custom models without coding experience.
The main features of Simple ML include data cleaning, data exploration, feature scaling and selection, model selection, hyperparameter tuning,and model deployment. The platform also provides users with an opportunity to perform advanced customization on their models with AI techniques and a high degree of settings flexibility. Simple ML allows users to customize their experiences by providing different functionalities such as Jupyter Notebook, API access, workflow adjustment and unique model training.
Simple ML solutions range from sales forecasting, churn analysis, image classification, and text classification to other various problem types in multiple domains. The platform offers cost-effective AI solutions for business users who may not have huge development budgets or require custom models for their needs.
Features
Automatic Machine Learning (AutoML)
Data Cleaning
Simple ML's AutoML service offers data cleaning functionality to ensure that the data used for the machine learning process is free of errors, inconsistencies, and redundancies. This feature automatically identifies and removes invalid data entries, corrects formatting errors, and handles missing data points through imputation with mean, median, or regression-based algorithms.
Feature Scaling and Selection
Simple ML's AutoML service includes feature scaling and selection functionality to ensure that the machine learning model's inputs are optimized for efficient analysis. This feature scales the data with normalization and centering techniques, selects the most relevant features through advanced statistical algorithms, and performs dimensionality reduction for visually challenging datasets.
Model Selection and Hyperparameter Tuning
Simple ML's AutoML service offers a range of model selection and hyperparameter tuning functionalities to ensure that the machine learning model is optimized for accuracy, precision, and performance. This feature includes a comprehensive suite of classic and modern algorithms, such as random forests, gradient boosting, neural networks, and support vector machines. Furthermore, hyperparameter optimization is done using cutting-edge optimization algorithms, such as Bayesian optimization, gradient-based optimization, and random search optimization.
Custom Machine Learning Model Generation
Jupyter Notebook Integration
Simple ML's custom machine learning model generation functionality integrates Jupyter Notebook to offer advanced data exploration, transformation, and visualization options for users who are confident with coding experience. This feature provides users with a variety of kernels that include popular languages such as Python, R, and Julia, along with a collection of common machine learning libraries such as scikit-learn, TensorFlow, and PyTorch.
API Access and Workflow Adjustment
Simple ML's custom machine learning model generation functionality includes API access and workflow adjustment features to ensure that users have full control over the model's development process. The API access feature allows users to use Simple ML within their existing software stack, while the workflow adjustment feature enables users to customize the machine learning development process for their requirements, such as creating validation datasets, defining model metrics, and managing the training process.
Unique Model Training
Simple ML's custom machine learning model generation functionality provides users with unique model training features to ensure that the machine learning model is optimized for the individual requirements. This feature includes advanced techniques such as transfer learning, multi-task learning, and semi-supervised learning. In particular, the semi-supervised learning process can be used to train a model on a large number of unlabeled data points combined with a few labeled examples to drastically improve model accuracy for complex datasets.
Integrations with Popular Applications
Excel Integration
Simple ML integrates with Excel to offer users a streamlined way to manage datasets and create machine learning models. The integration allows the data to be imported directly from Excel, and the machine learning models output can be visualized in Excel through powerful graphing and charting features.
Google Workspace Integration
Simple ML integrates with Google Sheets, Google Slides and Power BI to offer users a streamlined way to create and visualize machine learning models. The integration allows the data to be imported directly from the Google Workspace applications, and the machine learning models output can be visualized in Google Sheets and Power BI through powerful graphing and charting features. Furthermore, Simple ML's AutoML service can be used within Google Slides to offer detailed insights and predictions within presentations.
Other Application Integrations
Simple ML supports integrations with other popular data analysis and visualization applications, such as Tableau and PowerApps, to offer users a comprehensive way to create machine learning models. The integration allows the data to be imported directly from these applications, and the machine learning models output can be visualized in these applications through powerful graphing and charting features.
Diverse ML Solutions
Sales Forecasting
Simple ML's AutoML service is tailored for sales forecasting use cases, allowing users to gain predictive insights into sales trends, patterns and variations. This functionality builds robust models taking into account seasonality, promotional campaigns, and market competition.
Churn Analysis
Simple ML's AutoML service provides predictive analytics and insights for churn analysis use cases, allowing businesses to identify why they lose customers, and how to prevent it from happening. This functionality helps businesses segment customers based on demographics, purchasing habits, customer loyalty, and satisfaction scores in order to send targeted marketing messages.
Image Classification
Simple ML's AutoML service provides users with advanced image classification functionality, enabling them to automate tasks that previously required manual review. With the help of modern convolutional neural networks (CNNs), object recognition tasks, facial recognition tasks, and handwritten text recognition tasks can now be automated with a high degree of accuracy.
Text Classification
Simple ML's AutoML service provides users with advanced text classification functionality, enabling them to analyze, classify, and cluster their unstructured text data in a variety of domains. With the help of modern machine learning techniques, text classification tasks such as sentiment analysis, topic modeling, and spam detection can now be automated with a high degree of accuracy.
Cost-Effective Custom Solutions
Simple ML provides cost-effective custom machine learning solutions to ensure that businesses and individuals can benefit from the power of AI without having to incur the high costs related to building custom machine learning models. This functionality offers pre-built solutions for common business use-cases, allowing businesses to quickly and easily extract insights and make critical decisions.
FAQ
What is Simple ML for Sheets?
Simple ML for Sheets is an Addon for Google Sheets that allows you to use machine learning without being an ML expert, without coding, and without sharing data with third parties. It is developed by the TensorFlow Decision Forests team in Google Zurich.
All the operations (including model training operations) are executed on your computer in your web browser. That’s why your data is not exported to any third-party server. Models are saved on your Google Drive in the folder simple_ml_for_sheets.
What kind of models does Simple ML use?
Simple ML uses decision forest models (e.g., Random Forests, Gradient Boosted Trees) as those models are particularly well suited for tabular datasets. For more information about Decision Forests, we recommend our Decision Forests class on MLCC. Under the hood, Simple ML is powered by Yggdrasil Decision Forests, a fast, flexible and powerful library for state-of-the-art Decision Forest algorithms.
To learn more about the library, check out the YDF documentation.
Can I export my model to Colab?
Yes. After you trained a model, select the “Export model” task, select the “Colab” option, and click on “Export”.
You will see a snippet of Python code to copy/paste into Colab to run your model. The resulting model is both a TensorFlow Decision Forests model and an Yggdrasil Decision Forests model. It is compatible with TensorFlow Serving and the Yggdrasil Serving APIs.
Where can I find help if I run into issues?
The best way to find help is through the Simple ML for Sheets User group. If you found a bug, please report it. In last resort, contact us directly.
To report a bug, you can use the report form.
How does Simple ML handle missing values?
Simple ML trains a model on the non-missing values and uses this model to predict the missing ones. Simple ML trains multiple models to predict the existing values using cross-validation.
The predictions of those models are then compared to the actual values. If the predicted and actual values differ, the existing value is tagged as abnormal. The abnormality is not a yes/no question.
Instead, the abnormality is as a probability between 0% (normal) and 100% (abnormal). The computation of the abnormal score depends on the task of the model (e.g. classification, regression). However, in all cases, it is between 0 and 1.
