Method for classifying data items
Broadly speaking, the present techniques provide an automatic way of classifying data items within an environment (e.g. a business, workplace, organisation, etc.), and applying data management policies based on the classifications. This is advantageous over existing techniques which require manual classification of data items, which is time consuming in environments where hundreds of new data items may be generated in a day or week. The present techniques use an embedding machine learning, ML, model to automatically determine the relevant classification label(s) for an unlabelled data item, which is then used to select and apply the relevant data management policy(ies).
1 . A computer-implemented method for controlling actions performed with respect to a data item using a data management policy, the method comprising:
identifying a non-labelled data item within an environment in which actions performed with respect to the data item are being controlled;
generating, using an embedding machine learning (ML) model, at least one embedding vector for the non-labelled data item, where the at least one embedding vector captures characteristics of content of the non-labelled data item;
comparing the generated at least one embedding vector to a database of stored embedding vectors that correspond to classification labels for data items;
selecting, responsive to the comparing, at least one stored embedding vector that is most similar to the generated at least one embedding vector;
applying, to the non-labelled data item, at least one classification label corresponding to the selected at least one stored embedding vector, and thereby generating a labelled data item;
retrieving at least one data management policy corresponding to the at least one classification label of applied to the labelled data item; and
using the at least one retrieved data management policy to control an action performed with respect to the labelled data item.
2 . The method as claimed in claim 1 further comprising:
prior to the generating, dividing the non-labelled data item into two or more segments; and
generating the at least one embedding vector for each of the two or more segments.
3 . The method as claimed in claim 1 wherein
generating at least one embedding vector comprises:
extracting text content from the non-labelled data item; and
generating at least one embedding vector for the extracted text content.
4 . The method as claimed in claim 3 further comprising:
prior to the generating, translating the extracted text content into a pre-defined natural language.
5 . The method as claimed in claim 3 further comprising:
prior to the generating, dividing the extracted text content into two or more segments; and
generating the at least one embedding vector for each of the two or more segments.
6 . The method as claimed in claim 1 wherein comparing the generated at least one embedding vector to a database of stored embedding vectors comprises:
calculating a cosine similarity between the generated at least one embedding vector and each stored embedding vector.
7 . The method as claimed in claim 1 wherein selecting at least one stored embedding vector that is most similar to the generated at least one embedding vector comprises:
selecting at least one stored embedding vector that is within a predefined threshold distance in embedding space from the generated at least one embedding vector.
8 . The method as claimed in claim 1 wherein applying, to the non-labelled data item, the classification label corresponding to the selected at least one stored embedding vector comprises:
applying a single classification label to the non-labelled data item.
9 . The method as claimed in claim 1 wherein applying, to the non-labelled data item, the classification label corresponding to the selected at least one stored embedding vector comprises:
applying multiple classification labels to the non-labelled data item when multiple stored embedding vectors are selected.
10 . The method as claimed in claim 9 wherein retrieving at least one data management policy for the labelled data item comprises:
retrieving a data management policy corresponding to each label of the multiple classification labels applied to the non-labelled data item; and
determining which data management policy or policies to use to control actions performed with respect to the labelled data item.
11 . The method as claimed in claim 1 wherein the applying comprises applying multiple classification labels to the non-labelled data item, and wherein retrieving at least one data management policy for the labelled data item comprises:
retrieving a security policy corresponding to each classification label of the multiple classification labels applied to the non-labelled data item; and
determining which data management policy or policies to use to control actions performed with respect to the labelled data item.
12 . The method as claimed in claim 11 wherein determining which data management policy or policies to use to control actions performed with respect to the labelled data item comprises:
selecting the most strict security policy from the data management policies corresponding to the multiple labels.
13 . The method as claimed in claim 1 further comprising:
receiving an override instruction to ignore one or more of: a label applied to the labelled data item, and a data management policy associated with a label applied to the labelled data item.
14 . The method as claimed in claim 1 wherein using the at least one data management policy to control an action performed with respect to the labelled data item comprises:
receiving a request to perform an action with respect to the labelled data item;
determining, using the at least one data management policy, whether the request should be granted; and
granting the request to perform the action with respect to the labelled data item responsive to the determining.
15 . The method as claimed in claim 1 wherein using the at least one data management policy to control an action performed with respect to the labelled data item comprises controlling any one or more of: accessing, reading, modifying, editing, sharing, archiving, deleting, distributing within the environment, and distributing external to the environment.
16 . A system for controlling actions performed with respect to a data item in an environment, the system comprising:
a classification database for storing embedding vectors that correspond to classification labels for data items;
a classification engine for identifying a non-labelled data item within the system; and
an embedding machine learning, ML, model for generating at least one embedding vector for the non-labelled data item, where the at least one embedding vector represents content of the non-labelled data item;
wherein the classification engine is configured for:
comparing the generated at least one embedding vector to the database of stored embedding vectors;
selecting, responsive to the comparing, at least one stored embedding vector that is most similar to the generated at least one embedding vector;
applying, to the non-labelled data item, at least one classification label corresponding to the selected at least one stored embedding vector, and thereby generating a labelled data item;
retrieving, from storage storing a plurality of data management policies for each classification label, at least one data management policy corresponding to the at least one classification label of the labelled data item; and
using the at least one data management policy to control an action performed with respect to the labelled data item.