Extraction of multi-modal online resources associated with research papers

A method, comprising:

• storing a set of candidate resources and a research paper, each including one or more content fields, the set of candidate resources comprise media content and are associated with the research paper;
• encoding each of one or more first content fields in the research paper into a first vector based on a first field type associated with each of the one or more first content fields;
• parsing each of the stored set of candidate resources into one or more second content fields;
• encoding each of the one or more second content fields in each of the parsed set of candidate resources into a second vector, based on a second field type associated with each of the one or more second content fields;
• comparing the first vector for each of the encoded one or more first content fields with the second vector for each of the encoded one or more second content fields;
• determining a final set of resources based on the comparison, the determining being based on when the comparison results in the first vector matching the second vector;
• generating a machine-learning model based on training data including a plurality of resources associated with a plurality of research papers, the plurality of resources comprising at least one of:
• a research paper type, a poster type, a presentation slide type, a video type, or a program code type, the plurality of research papers include the research paper and the plurality of resources include the final set of resources;
• and
• controlling a display screen to output the determined final set of resources and the research paper, the method enabling a better and enhanced understanding of the research paper and a domain of interest associated with the research paper based at least in part on the predefined plurality of resources types.

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:

• extracting, from one or more first websites, one or more first resources based on a title associated with a research paper;
• identifying a set of resource types associated with the extracted one or more first resources;
• determining one or more first resource types from a predefined plurality of resource types based on the identified set of resource types, the set of resource types excludes the determined one or more first resource types, the predefined plurality of resource types comprises at least one of:
• a research paper type, a poster type, a presentation slide type, a video type, or a program code type;
• extracting, from one or more second websites, one or more second resources, associated with the determined one or more first resource types, based on the title associated with the research paper, wherein each of the one or more first resources and the one or more second resources comprises media content;
• determining a final set of resources based on a comparison between one or more first content fields of the research paper and one or more second content fields of the extracted one or more first resources and the extracted one or more second resources, the determining being based on when the comparison results in the one or more first content fields matching the one or more second content fields;
• and
• controlling a display screen to output the determined final set of resources and the research paper, the method enabling a better and enhanced understanding of the research paper and a domain of interest associated with the research paper based at least in part on the predefined plurality of resources types.

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:

• storing a set of candidate resources and a research paper, each including one or more content fields, the set of candidate resources comprise media content and are associated with the research paper;
• encoding each of one or more first content fields in the research paper into a first vector based on a first field type associated with each of the one or more first content fields;
• parsing each of the stored set of candidate resources into one or more second content fields;
• encoding each of the one or more second content fields in each of the parsed set of candidate resources into a second vector, based on a second field type associated with each of the one or more second content fields;
• comparing the first vector for each of the encoded one or more first content fields with the second vector for each of the encoded one or more second content fields, the determining being based on when the comparison results in the first vector matching the second vector;
• determining a final set of resources based on the comparison;
• generating a machine-learning model based on training data including a plurality of resources associated with a plurality of research papers, the plurality of resources comprising at least one of:
• a research paper type, a poster type, a presentation slide type, a video type, or a program code type, the plurality of research papers include the research paper and the plurality of resources include the final set of resources;
• and
• controlling a display screen to output the determined final set of resources and the research paper, the method enabling a better and enhanced understanding of the research paper and a domain of interest associated with the research paper based at least in part on the predefined plurality of resources types.