Modification of base distance representation using dynamic objective

A computer-implemented method, comprising:

• in a training phase:
• obtaining training data pertaining to roles of members of a social networking service, the training data including, for each of a plurality of members of the social networking service, member attribute information, action information regarding one or more actions taken in the social networking service, and an indication as to whether the corresponding member achieved an objective;
• for each of the plurality of members of the social networking service, feeding the corresponding training data into a machine learning algorithm to train an objective prediction model to output a prediction value for an input member and an objective based on member attribute information and action information;
• in a prediction phase:
• obtaining an identification of a first member of the social networking service;
• retrieving, using the identification, a first member profile corresponding to the first member of the social networking service;
• for each of a plurality of different potential objectives, pass one or more attributes from the first member profile, one or more actions performed by the first member in the social networking service, and the potential objective to the objective prediction model to generate a prediction value for the first member and potential objective, wherein a potential objective is an objective that has not vet been selected for the first member;
• automatically selecting one of the plurality of different potential objectives for the first member based on the generated prediction values;
• and
• using the selected one of the plurality of different potential objectives to optimize coordinates, in a representation space by mapping each of a plurality of entities in the social network structure to a different set of a plurality of coordinates in an n-dimensional space and altering one or more of the sets of plurality of coordinates based on the selected one of the plurality of different potential objectives.

A non-transitory machine-readable storage medium comprising instructions, which when implemented by one or more machines, cause the one or more machines to perform operations comprising:

• in a training phase:
• obtaining training data pertaining to roles of members of a social networking service, the training data including, for each of a plurality of members of the social networking service, member attribute information, action information regarding one or more actions taken in the social networking service, and an indication as to whether the corresponding member achieved an objective;
• for each of the plurality of members of the social networking service, feeding the corresponding training data into a machine learning algorithm to train an objective prediction model to output a prediction value for an input member and an objective based on member attribute information and action information;
• in a prediction phase:
• obtaining an identification of a first member of the social networking service;
• retrieving, using the identification, a first member profile corresponding to the first member of the social networking service;
• for each of a plurality of different potential objectives, pass one or more attributes from the first member profile, one or more actions performed by the first member in the social networking service, and the potential objective to the objective prediction model to generate a prediction value for the first member and potential objective, wherein a potential objective is an objective that has not vet been selected for the first member;
• automatically selecting one of the plurality of different potential objectives for the first member based on the generated prediction values;
• and
• using the selected one of the plurality of different potential objectives to optimize coordinates, in a representation space by mapping each of a plurality of entities in the social network structure to a different set of a plurality of coordinates in an n-dimensional space and altering one or more of the sets of plurality of coordinates based on the selected one of the plurality of different potential objectives.