System and Method for Recommending Semantically Relevant Content

Abstract

A property vector derived from extractable measurable properties of a data file is mapped to semantic properties for that data file. The property vector is an output from a trained artificial neural network that, following pairwise training of the ANN using pairs of files that map pairwise similarity/dissimilarity in property space towards corresponding pairwise semantic similarity/dissimilarity in semantic space, both preserves and is representative of semantic properties of the data file. The system and method assesses, based on comparisons between generated property vectors, ranks and then recommends and/or filters semantically close or semantically disparate candidate files in a database from a query from a user that includes the data file. Applications of the categorization and recommendation system and method apply to media or search tools and social media platforms, including media in the form of music, video, images data and/or text files.

1 . (canceled) 
     
     
         2 . (canceled) 
     
     
         3 . (canceled) 
     
     
         4 . (canceled) 
     
     
         5 . (canceled) 
     
     
         6 . (canceled) 
     
     
         7 . (canceled) 
     
     
         8 . (canceled) 
     
     
         9 . (canceled) 
     
     
         10 . (canceled) 
     
     
         11 . (canceled) 
     
     
         12 . (canceled) 
     
     
         13 . (canceled) 
     
     
         14 . (canceled) 
     
     
         15 . (canceled) 
     
     
         16 . A method of providing a file recommendation based on sematic qualities, the method comprising:
 identifying a recently consumed reference data file that has been consumed by a user;   processing the reference data file to extract properties therefrom;   calculating a first file vector in property space from said extracted properties, wherein the first file vector both preserves and is representative of semantic properties of content of the reference data file;   evaluating a new data file in terms of semantic closeness to the reference data file, said evaluation based on a relative comparison between the first file vector and a different second file vector derived from properties of the new data file and where the second file vector also preserves and is representative of semantic properties of content of the new data file;   determining availability and extent of at least one of (a) user data obtained for the user, and (b) property vectors in candidate file data, said property vectors reflective of semantic qualities therein;   providing the file recommendation based on a probabilistic weighting between:
 a content-based approach of semantic closeness evaluated between the reference data file and the new data file; and 
 a predictive approach based on one of a predictive model, a reinforcement learning “RL” algorithm or heuristic processing function, wherein the predictive approach is based on sufficiency in availability of user data and property vectors in candidate file data. 
   
     
     
         17 . The method of providing a file recommendation according to  claim 16 , wherein the probabilistic weighting between the content-based approach and the predictive approach varies with time. 
     
     
         18 . The method of providing a file recommendation according to  claim 16 , wherein initially the content-based approach is absolute. 
     
     
         19 . (canceled) 
     
     
         20 . (canceled) 
     
     
         21 . (canceled) 
     
     
         22 . (canceled) 
     
     
         23 . (canceled) 
     
     
         24 . (canceled) 
     
     
         25 . (canceled) 
     
     
         26 . (canceled) 
     
     
         27 . (canceled) 
     
     
         28 . (canceled) 
     
     
         29 . (canceled) 
     
     
         30 . (canceled) 
     
     
         31 . (canceled) 
     
     
         32 . (canceled) 
     
     
         33 . (canceled) 
     
     
         34 . (canceled) 
     
     
         35 . A system containing processing intelligence arranged to provide a file recommendation based on sematic qualities, the processing intelligence arranged to:
 process a reference data file to extract properties therefrom;   calculate a first file vector in property space from said extracted properties, wherein the first file vector both preserves and is representative of semantic properties of content of the reference data file;   evaluate a new data file in terms of semantic closeness to the reference data file, said evaluation based on a relative comparison between the first file vector and a different second file vector derived from properties of the new data file and where the second file vector also preserves and is representative of semantic properties of content of the new data file;   determine availability and extent of at least one of (a) user data obtained for the user, and (b) property vectors in candidate file data, said property vectors reflective of semantic qualities therein;   provide the file recommendation based on a probabilistic weighting between:
 a content-based approach of semantic closeness evaluated between the reference data file and the new data file; and 
 a predictive approach based on one of a predictive model, a reinforcement learning “RL” algorithm or heuristic processing function, wherein the predictive approach is based on sufficiency in availability of user data and property vectors in candidate file data. 
   
     
     
         36 . The system of  claim 35 , wherein the system intelligence is arranged to vary with time the probabilistic weighting between the content based approach and the predictive approach. 
     
     
         37 . The system of  claim 36 , wherein the system intelligence initially makes the content-based approach absolute. 
     
     
         38 . The system of  claim 36 , wherein the system intelligence is a server-side component remotely and selectively connected to a user device over a network. 
     
     
         39 . The system of  claim 19 , wherein the system intelligence is located, at least in part, in a user device. 
     
     
         40 . The system of  claim 35 , wherein the processing intelligence is located, at least in part, in a user device. 
     
     
         41 . The system according to  claim 35 , wherein the file vector and each property vector is an output from a trained artificial neural network “ANN” that, following pairwise training of the ANN using pairs of training files, maps pairwise similarity/dissimilarity in property space towards corresponding pairwise semantic similarity/dissimilarity in semantic space to preserve semantic evaluation by valuing, on a pairwise basis, semantic perception reflected in quantified semantic dissimilarity distance measures over property assessment reflected by distance measures in property space, said quantified semantic dissimilarity distance measures. 
     
     
         42 . The processing system of  claim 41 , wherein:
 the ANN compares a subjectively-derived semantic vector against a property space vector, the subjectively-derived semantic vector being generated independently of the property space vector, the ANN correlating quantified semantic dissimilarity measures for the subjectively-derived semantic vector, which describes content in semantic space for each of a first data file and also a different second data file, with related property separation distances for the property space vector, which is provided in property space and which describes measurable signal quality extracted for respective content of both the first data file and the different second data file, to provide an output that is adapted, over time, to align a result in property space to a result in semantic space, and   wherein the ANN is configured, during adaptation of weights in the ANN, to value semantic dissimilarity measures over measurable properties and such that the ANN is configured to map pairwise similarity/dissimilarity in property space for the first data file and the second data file towards corresponding pairwise semantic similarity/dissimilarity in semantic space for the first data file and the second data file thereby to configure a system, in identifying and quantifying similarity or dissimilarity in audio or image-based content, to output a measure of similarity between said content of said first data file relative to content in said second data file, and   the subjectively-derived semantic vector is derived using natural language processing (NLP) of a text description of content for each of the first data file and the different second data file.