Automatic learning curriculum generation

An apparatus, comprising:

• one or more processors configured to receive input data and a plurality of instructions from a group of users and automatically generate curricula, the one or more processors comprising:
• an assessor unit configured to integrate the input data from the group of users and store the input data, wherein the users comprise one or more instructors, a class of students attending a course, and one or more educational institutions;
• a searcher unit configured to search for a collection of course materials pertaining to the course from a plurality of sources, wherein the sources comprise a collection of substances from the one or more instructors, a plurality of institutional repositories, and a collection of educational databases publicly available;
• an analyzer unit configured to determine a plurality of blended-learning contents for a plurality of sub-groups of the class respectively by searching for a collection of course materials from a plurality of sources with a web-search engine or search Application Programming Interface (API), wherein the searching for the collection of course materials from the plurality of sources is performed using a set of keywords relevant to one or more core topics and a plurality non-core topics provided by the one or more instructors, wherein the plurality of blended-learning contents comprise a plurality of common topics for the class as a whole and a plurality of private topics for each of the sub-groups from output of the searcher, and wherein each of the sub-groups comprises at least one of the students in the class;
• and
• an optimizer unit configured to output a plurality of curricula based on the blended-learning contents for the plurality of sub-groups of the class, wherein the curricula are optimized with respect to a set of constraints, wherein the optimizer unit is further configured to perform operations comprising:
• choosing a second set of the common topics from the first set of the common topics to fit within a common course length;
• choosing a plurality of second sets of the private topics from the first sets of the private topics to fit within a private course length;
• associating each of the second sets of the private topics to each of the students in the class;
• organizing a common class schedule based on the second set of the common topics for the class as a whole;
• organizing a plurality of private class schedules with the second sets of the private topics for each of the sub-groups of the class;
• and
• generating a plurality of curricula each of which generated for a respective sub-group of the sub-groups of the class by combining the common class schedule, each of the private class schedules, a number of assignments, a number of projects, and a number of exams.

A non-transitory computer readable storage medium tangibly embodying a computer readable program code having computer readable instructions that, when executed, causes a processor to carry out a method of optimizing curricula, the method comprising:

• integrating, by a processor, input data from a group of users, wherein the users provide a plurality of instructions for processing the input data, and wherein the users comprise one or more instructors, a class of students attending a course, and one or more educational institutions;
• constructing, by the processor, a course boundary according to the input data of the users, wherein the course boundary comprises a plurality of topics of the course from the input data of the users;
• determining, by the processor, a plurality of blended-learning contents pertaining to the course boundary for a plurality of sub-groups of the class respectively by searching for a collection of course materials from a plurality of sources with a web-search engine or search Application Programming Interface (API), wherein the searching for the collection of course materials from the plurality of sources is performed using a set of keywords relevant to one or more core topics and a plurality non-core topics provided by the one or more instructors, wherein the blended-learning contents comprise a plurality of common topics for the class as a whole and a plurality of private topics for each of the sub-groups, and wherein each of the sub-groups comprises at least one of the students in the class;
• and
• optimizing, by the processor, a plurality of curricula based on the blended-learning contents for the plurality of sub-groups of the class with respect to a set of constraints,
• wherein the constructing of the course boundary according to the input data of the users comprises:
• assembling a concept graph of the course, wherein the concept graph comprises one or more core topics and a plurality of non-core topics provided by the one or more instructors;
• and
• for each of the non-core topics, determining whether to exclude the respective non-core topic by analyzing the input data;
• and
• wherein the determining of whether to exclude the respective non-core topic by analyzing the input data comprises:
• receiving a plurality of responses from the students concerning each of the non-core topics;
• assigning a plurality of thresholds for each of the responses concerning each of the non-core topics;
• and
• comparing each of the thresholds with a corresponding response for the respective non-core topic to exclude the respective non-core topic when the threshold is not exceeded.