Exciting Update: ML Assisted Labeling Now Supports Multiple Input Columns!

We are excited to announce a new capability in our ML-Assisted Labeling feature: support for multiple input columns. This enhancement allows you to use multiple features (from multiple columns) to improve the accuracy of your model predictions.

Why is This Important?

Previously, our ML-Assisted Labeling could only process one input column at a time. Now, you can use multiple columns, making it easier to handle complex datasets and improve your model's performance. This is especially useful for datasets that have many features, like the Titanic dataset or bank marketing data.

Key Benefits:

• Use Multiple Features: Select multiple columns as features for your model.
• Handle More Use Cases: Cover a wider range of dataset use cases.
• Improve Accuracy: Improve model performance by incorporating multiple features.

How to Get Started

• Enable ML-Assisted Labeling from your extension and select a provider. For details, please visit this link.
• Select a provider. For example, you can choose LLM Labs.
• Select target columns.

• Get the predictions.

Real-World Applications

Many machine learning tasks require multiple features, and in this case, one column represents one feature. Tabular datasets, such as the Titanic dataset or bank marketing dataset, have many features used to predict the target variable. These datasets represent real-world problems where users often have many features to process. For ML-Assisted Labeling, the model is already trained with many features, so limiting the model to one input column will significantly reduce its accuracy.

By utilizing multiple input columns, you can leverage the full potential of your data, ensuring that the model performs at its best.

Try It Now!

Ready to see the difference? Access your NLP project and enable ML-Assisted Labeling to start experimenting with multiple input columns. This update will help you achieve better model outcomes with less effort.

If you have any questions or need help getting started, feel free to reach out support@datasaur.ai and Happy labeling!