Oct 30, 2021 · In this work, we proposed a simple but effective constraint random decoding method for back-translation, which follows an APE framework. First, ...
To alleviate this problem, we propose a simple but effective constraint random decoding method for back-translation. The proposed method is based on an ...
摘要. Back-translation has been proven to be an effective data augmentation method that translates target monolingual data .
Jan 10, 2024 · MBR decoding is shown to generate higher quality sentences than random sampling and beam search in directed text generation tasks including ...
Dec 6, 2023 · We annotate random source words with pseudo-terminology translations obtained from word alignment to first train a terminology- aware model.
Missing: Diverse | Show results with:Diverse
Oct 9, 2023 · Random constraint: Random (but presumably correct) word mappings are obtained using a word alignment tool and provided as a pseudo- terminology ...
Missing: Diverse | Show results with:Diverse
Our findings show that generating back translation using nucleus sampling ... diversity decoding rates for different inputs using reinforcement learning (RL).
Jan 2, 2021 · To take a step further, trainable greedy decoding replaces the unstructured noise with a learnable random variable, predicted by a RL agent that ...
Generating Diverse Back-Translations via Constraint Random Decoding. Abstract. Back-translation has been proven to be an effective data augmentation method ...
Sep 19, 2022 · Abstract. When generating text from probabilistic models, the chosen decoding strategy has a profound effect on the resulting text.