Sentence Meaning Similarity Detector Using FAISS
Premanand Ghadekar, Sahil Mohite, Omkar More, Praiwal Patil, Sayantika, Shubham Mangrule
Abstract
Sentence meaning similarity in machine learning (ML), which is automatically done by calculating the semantic similarity between sentences. In this study, word embeddings are used to depict sentences as numerical vectors in a high-dimensional space. It is then discussed how different similarity measures may be applied to evaluate these vectors and determine how similar the sentences are. The FAISS (Facebook AI Similarity Search) method is covered in the study. An open-source library developed by Facebook AI Research for similarity search and clustering of dense vectors, can be used to efficiently calculate sentence similarity by leveraging its efficient nearest neighbor search capabilities. Searching exhaustively increases the number of comparisons to decrease that we proposed in this paper.