2021/10/07 by Pierre Colombo, Chouchang Yang, Colombo, Pierre +5 · 11 citations
Computer Science · #Algorithm #Artificial Intelligence (cs.AI) #Artificial intelligence #Artificial neural network #Beam search #Computation and Language (cs.CL) #Computer science #FOS: Computer and information sciences #Language model #Multimodal Machine Learning Applications #Natural Language Processing Techniques #Order (exchange) #Path (computing) #Search algorithm #Selection (genetic algorithm) #Sentence #Sequence (biology) #Similarity (geometry) #Topic Modeling #cs.AI #cs.CL
paper · pdf · doi:10.48550/arxiv.2110.03389
published in arXiv (Cornell University) (Cornell University)
arxiv created 2021/10/07 · openalex publication_date 2021/10/07 · arxiv updated 2021/10/08 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Sequence-to-sequence neural networks have been widely used in language-based applications as they have flexible capabilities to learn various language models. However, when seeking for the optimal language response through trained neural networks, current existing approaches such as beam-search decoder strategies are still not able reaching to promising performances. Instead of developing various decoder strategies based on a "regular sentence order" neural network (a trained model by outputting sentences from left-to-right order), we leveraged "reverse" order as additional language model (a trained model by outputting sentences from right-to-left order) which can provide different perspectives for the path finding problems. In this paper, we propose bidirectional strategies in searching paths by combining two networks (left-to-right and right-to-left language models) making a bidirectional beam search possible. Besides, our solution allows us using any similarity measure in our sentence selection criterion. Our approaches demonstrate better performance compared to the unidirectional beam search strategy.