AI Supported Mentor-Mentee Matching Algorithm

Utilizing OpenAI, MentorEase is able to extract, classify, and understand free-form text fields that mentees and mentors enter in registration forms & resume uploads.

This means that when a whole piece of text is entered into a text box – AI is able to read it, identify common phrases the mentor and mentee used, and then add it into the matching algorithm mix.

MentorEase_mentoring_software_AI-based_matching_artificial_intelligence_3
Beyond this, MentorEase can use AI tools to automatically review all mentor and mentee resumes and CVs to find similar terms that they both use. The system can review thousands of words instantly, extracting key terms and important ‘needles in a haystack’ – terms and phrases that both the mentee and potential mentors used.

MentorEase shows these key terms in context side by side so they can be assessed by the mentee or admin selecting a match. This pattern recognition service provided by AI can be very useful and practically impossible to find without many hours of comparing resumes.

AI integrations can include:

‘Key Phrase Extraction’

AI is able to extract sentences that collectively convey the essence of the whole paragraph or document to be used as a useful addition to other matching criteria.

‘Entity Linking’

AI is able to identify important concepts in text, including key phrases and named entities such as people, locations, events, and organizations.

MentorEase__MS_Azure_Text__Key_Phrase_Analytics__Example

‘Sentiment Analysis’

There are various ways to calculate a sentiment score, but the most common method is to use a dictionary of negative, neutral, or positive words. The text is then analyzed to see how many negative and positive words it contains. This can give a good idea of the overall sentiment of the text. For example if a mentee states ‘I am having trouble getting up the nerve to present to a group of people as I find it a bit scary’, and a mentor in the program has experience in this area, the question can help to make a match of these participants.

‘Ignore List’

To train the AI tools to be more useful, it is helpful to create a list of words or phrases that it initially detects, but are useless for matching purposes and tell it to ignore them in the analysis.

‘Lemmatisation’

This is the process of merging similar words or terms that have the same meaning – to avoid having too many similar results to review and speed up the effectiveness of the process.

'Text Embeddings'

This method uses the AI tools to extract key terms from text (resume and open text fields) and categorize them into types. Next it converts the text into a mathematical description of it by assigning values to each type of result found. By placing each set of numbers on a vector space we can map the “location” of each person. Comparing the “distance” between each mentor in relation to the mentee we can find the best potential mentors for them.

MentorEase AI methodology Text Embedding Score

'Matching Algorithm Scorecard'

By adding this new tool for matching we can now display three scores for each potential match:

* Field Score – from matching key data fields such as selections, checkboxes, etc.
* AI Score – from the open text fields and resume comparisons by AI
* Overall Score – combine scores of both

When reviewing potential matches MentorEase allows filtering by each of these scores to obtain a better idea of the best possible match for each mentee.

OpenAI_MentorEase_mentoring_software
MentorEase_mentoring_software_AI-based_matching_artificial_intelligence_4

How AI Can Help Manage Mentoring Programs

FastLane-Social-Badge-2023_MentorEase_Mentoring_Software
MentorEase is a member of the Vector Institute's FastLane program for AI research
FastLane_Social__MentorEase_Mentoring_Software
Vector Institute Artificial Intelligenct AI MentorEase mentoring software