Machine learning techniques for determining predicted similarity scores for input sequences

A computing system for dynamically generating a predicted similarity score for a pair of input sequences, the computing system comprising one or more processors and memory including program code, the memory and the program code configured to, with the one or more processors, cause the computing system to at least:

• generate a token-level similarity probability score for a pair of input sequences based at least in part on a first cross-token image representation for a first input sequence of the pair of input sequences and a second cross-token image representation for a second input sequence of the pair of input sequences, wherein the token-level similarity probability score is based at least in part on a comparison between a first dimensionally-reduced image representation for the first cross-token image representation and a second dimensionally-reduced image representation for the second cross-token image representation;
• generate, using a machine learning model, a predicted similarity score for the pair of input sequences based at least in part on the token-level similarity probability score;
• and
• initiate the performance of one or more prediction-based actions based at least in part on the predicted similarity score.

A computer program product comprising at least one non-transitory computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions configured to:

• generate a token-level similarity probability score for a pair of input sequences based at least in part on a first cross-token image representation for a first input sequence of the pair of input sequences and a second cross-token image representation for a second input sequence of the pair of input sequences, wherein the token-level similarity probability score is based at least in part on a comparison between a first dimensionally-reduced image representation for the first cross-token image representation and a second dimensionally-reduced image representation for the second cross-token image representation;
• generate, using a machine learning model, a predicted similarity score for the pair of input sequences based at least in part on the token-level similarity probability score;
• and
• initiate the performance of one or more prediction-based actions based at least in part on the predicted similarity score.