ScholarWorks@숙명여자대학교

초록

The latest research has shown that hashing can decrease memory usage and computing time in many applications without significantly reducing efficiency. Recently, hashing approaches have been used in open-domain question answering tasks, where only the most basic hashing layer is employed. We propose a weighted Hamming distancebased semantic hashing method that learns optimal dimension weights to obtain effective passage retrieval results. We compared the performance of our method to those of stateof-the-art semantic hashing baselines in a passage retrieval task on open-domain question answering. The results show that our method can outperform the baselines and achieve accuracy similar to that attainable with dense vector-based retrieval. © 2024.

키워드