Machine learning applied to human learning

In order to memorize information, humans use various methods: mnemonics, visualization and relating the information to already known information. Despite these efforts, memorizing information is a difficult task for humans. On the other hand, a computer can easily memorize information. Using this ability of computers, our research tries to make memorizing word meanings easier for humans. Humans need to be reminded of the information at intervals (that need not be equally spaced) in order to register the information in the memory. Keeping this in mind, we use vocabulary building as the area of this research. Vocabulary building can be contextualized as in reading a book or de-contextualized as in learning word meanings from a list. We focus on de-contextualized vocabulary building, which is used in vocabulary based tests. In order to achieve the goal of building the de-contextualized vocabulary for a student, we need to assess the student at intervals of time, and using the information thus obtained, maximize the probability that the student will be successful in his/her next assessment. Thus, the aim is to find the assessment intervals and the number of times the word and its meaning have to be reiterated, using the data from the previous assessments, so that the word and its meaning get registered in the student's memory. Also, learning is a cognitive process, thus, adaptability for each student comes into the picture. In order to account for adaptability we use machine learning. The result of each assessment of the student is recorded in order to determine the probability of success in the next assessment. The research thus tries to solve the difficulties involving memory faced by humans in learning new vocabulary: be it in a native or non-native language.