DISCO’s tag predictions help lawyers review documents faster and more accurately. Read through the steps of this guide and learn how to turn on, sort by, and use this feature to its fullest capabilities.
The general workflow we will show is to:
- Specify the tags for which you would like to see predictions
- Begin reviewing and tagging documents
- Leverage tag predictions in your review
- Sort your search results by predictions
- See which tags are recommended when reviewing documents
- Prioritize the documents in your Workflow Stage based on a tag’s predictions
Specifying tags for predictions
To get started, navigate to SETUP > Tags within the main DISCO menu. Next, you will turn the ‘PREDICTIONS’ toggle to ON next to the desired tag(s) you wish to initiate machine learning for.
For this example, we will focus on the tag predictions for the tag ‘Broadband.’ As shown in the image below, to jump start machine learning for ‘Broadband’, turn the PREDICTIONS toggle on for that particular tag.
Note that it is okay if you already have applied a particular tag to documents before you enable it for predictions. Those tagging activities are not lost and will be used to make predictions once you enable the toggle.
Begin reviewing and tagging documents
Now you will want to start reviewing documents as you normally would do. As you apply, or don’t apply, tags to each document, the DISCO ML engine learns from those actions silently. Once it has learned enough from your tagging activities, DISCO will recommend which documents should or should not be tagged with the ‘Broadband’ tag and the strength of that recommendation.
Tag predictions are reflected as a number/score ranging from -100 to 100 applied to every document in your database. A positive number indicates the likelihood a tag should be applied to a document, while a negative number indicates the likelihood a tag should not be applied to a document. This process of learning, training and scoring documents based on your normal tagging activities happens continuously in the background. This architecture is unique to DISCO!
Sort search results by tag predictions
First, you may want to add specific columns to your view so you can see the tag predictions and recommended tags in your search results. To do so, in your search results, go to View > Customize to add one or more of the tags you have enabled for predictions. Create a new view, or copy an existing one, and add the predictive-enabled tags from Tags > Specific Tag Predictions. You may also want to add the column named ‘Recommended Tags’, which will display a list of the tags which DISCO ML strongly recommends for each document. In this example, we will be adding the ‘Broadband’ tag as a column which we enabled for predictions earlier. You may also want to rearrange your columns for a better visual (please read Custom Columns User Guide for a more detailed step-by-step guide on how to set your custom views.) Once you’re done with configuring your view, save and apply it.
Now you should see the columns you added to your view related to tag predictions. You can sort the results of any search by a particular tag’s prediction in one of two ways: 1) clicking within the ‘Broadband Prediction’ column to sort documents or 2) selecting to sort by ‘Broadband Predictions’ straight from the ‘Sort by’ menu. You can also change the sort order to ascending or descending based on the tag prediction score for each document.
Before clicking within the ‘BROADBAND PREDICTION’ column. *Notice how your documents are still sorted by Relevance.
After clicking within the ‘BROADBAND PREDICTION’ column. *Notice how your documents are now sorted by ‘Broadband Prediction.’
Sorting by ‘Broadband Prediction’ from the ‘Sort by’ drop-down menu.
What do the scores and colors mean?
DISCO ML will serve tagging suggestions and scores start appearing. DISCO’s Tag Predictions score how likely a tag is to be applied to a given document on a scale of +100 to -100, +100 being most likely to be applied and -100 being least likely to be applied. These numerical scores are also accompanied by a color code which mimics that of a traffic light; green being likely to highly likely, yellow being somewhat likely or neutral, and red being unlikely to highly unlikely that a certain tag should be applied to a document.
High scores (green) = high likely this tag should be applied to this document
*Note - you may notice some of the color squares are solid and some are not. A solid color square means that tag has already been applied to the document, while a hollow square means that tag has not been applied to the document.
Medium scores (lime green/yellow)= Somewhat likely or neutral that a tag should be applied to a document
Low scores (Orange or red)= Unlikely to highly unlikely that a certain tag should be applied to a document
Tag predictions and recommended tags while reviewing documents
When you are viewing a document in the Document Viewer for review, you can see Recommended Tags as well as the tag predictions. By default, predictions are not displayed when viewing a document. You must select “Show Predictions” to display which tags are recommended for that document. You can choose to hide these predictions as well. The recommended tags are only tags which have been enabled for predictions and have a score higher than 50 for that document, indicating a higher likelihood of being applied to a particular document.
Select the Show Predictions button. Click on the + icon next to the recommended tags to apply them to the document, if you think the recommendation is correct.
*Note that DISCO ML does not apply any tags to documents automatically. Instead, it recommends tags to you to help reduce errors and find relevant documents faster.
Can hide Predictions as well.
You can also see the predictive score and color for all the tags predictive was turned on for when you click into the apply tag area.
From the prediction scores and colors displayed, you are able to see DISCO’s ML predict that the tags ‘Broadband’ and ‘Enron Outside USA’ are relevant to this document, while the other three tags show very low relevancy. You are able to confirm DISCO’s predictions are correct by reading the document these predictions were coded for.
Prioritized workflow review stage using tag predictions
You can leverage tag predictions to power a prioritized review in Workflow Review Stages. You can configure your Review Stage to order the documents based on the predictions for a particular tag when they are batched for review. This could allow you to find the more relevant documents earlier in your review than if you performed a linear review. For a more detailed explanation on prioritized reviews with tag predictions, please read DISCO’s data sheet, ‘Review Stage Management’.