Seed expansion in social network using graph neural network

A method, comprising:

• extracting, from one or more websites, a first plurality of seed accounts and a second plurality of seed accounts, related to the first plurality of seed accounts, based on domain-specific information, wherein the first plurality of seed accounts and the second plurality of seed accounts are extracted for training a graph neural network (GNN);
• encoding one or more features of each of the first plurality of seed accounts and each of the second plurality of seed accounts to generate a first vector, wherein the one or more features correspond to at least one of:
• a profile description, connection information, or a profile URL;
• selecting a new seed account different from the first plurality of seed accounts and the second plurality of seed accounts;
• extracting, from the one or more websites, a plurality of neighbor seed accounts for the selected new seed account;
• generating a second vector for the selected new seed account based on the extracted plurality of neighbor seed accounts using the trained GNN;
• determining the selected new seed account as a domain-specific seed account or as a non-domain specific seed account based on the generated first vector and the generated second vector;
• and
• updating the first plurality of seed accounts and the second plurality of seed accounts based on the determined new seed account as the domain-specific seed account or as the non-domain specific seed account.

One or more non-transitory computer-readable storage media configured to store instructions that, in response to being executed, cause a system to perform operations, the operations comprising:

• retrieving, from one or more websites, a first plurality of seed accounts based on a first user input, wherein the first user input indicates domain-specific information associated with the first plurality of seed accounts;
• determining a second plurality of seed accounts based on the first plurality of seed accounts associated with the domain-specific information, wherein the second plurality of seed accounts are determined for training a graph neural network (GNN);
• updating the retrieved first plurality of seed accounts and the determined second plurality of seed accounts based on the trained GNN;
• generating a final set of seed accounts associated with the domain-specific information, based on an application of a predefined ranking criteria on the updated first plurality of seed accounts and the second plurality of seed accounts;
• and
• controlling a display screen to output the generated final set of seed accounts.