paper-with-me

홈 › Papers

Generating Diverse and High-Quality Texts by Minimum Bayes Risk Decoding

2024-01-10 · Yuu Jinnai, Ukyo Honda, Tetsuro Morimura, Peinan Zhang

One of the most important challenges in text generation systems is to produce outputs that are not only correct but also diverse. Recently, Minimum Bayes-Risk (MBR) decoding has gained prominence for generating sentences of the highest quality among the decoding algorithms. However, existing algorithms proposed for generating diverse outputs are predominantly based on beam search or random sampling, thus their output quality is capped by these underlying methods. In this paper, we investigate an alternative approach -- we develop diversity-promoting decoding algorithms by enforcing diversity objectives to MBR decoding. We propose two variants of MBR, Diverse MBR (DMBR) and $k$-medoids MBR (KMBR), methods to generate a set of sentences with high quality and diversity. We evaluate DMBR and KMBR on a variety of directed text generation tasks using encoder-decoder models and a large language model with prompting. The experimental results show that the proposed method achieves a better trade-off than the diverse beam search and sampling algorithms.

📄 PDF Abstract BibTeX arXiv:2401.05054

Code (1)

CyberAgentAILab/diverse-mbr 공식 구현 pytorch

Tasks

DecoderDiversityLanguage ModelingLanguage ModellingLarge Language ModelText Generation

Methods 이 논문이 사용한 방법론

SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

Cascading CMA-ES Instances for Generating Input-diverse Solution Batches

2025-02-19 · Maria Laura Santoni, Christoph Dürr, Carola Doerr, Mike Preuss 외

Rather than obtaining a single good solution for a given optimization problem, users often seek alternative design choices, because the best-found solution may perform poorly with respect to additional objectives or cons…

Diversity

Necessary and Sufficient Watermark for Large Language Models

2023-10-02 · Yuki Takezawa, Ryoma Sato, Han Bao, Kenta Niwa 외

In recent years, large language models (LLMs) have achieved remarkable performances in various NLP tasks. They can generate texts that are indistinguishable from those written by humans. Such remarkable performance of LL…

ArticlesMachine Translation

Summarize, Outline, and Elaborate: Long-Text Generation via Hierarchical Supervision from Extractive Summaries

2020-10-14 · COLING 2022 10 · Xiaofei Sun, Zijun Sun, Yuxian Meng, Jiwei Li 외

The difficulty of generating coherent long texts lies in the fact that existing models overwhelmingly focus on predicting local words, and cannot make high level plans on what to generate or capture the high-level discou…

Text Generation

Exploring Precision and Recall to assess the quality and diversity of LLMs

2024-02-16 · Florian Le Bronnec, Alexandre Verine, Benjamin Negrevergne, Yann Chevaleyre 외

We introduce a novel evaluation framework for Large Language Models (LLMs) such as \textsc{Llama-2} and \textsc{Mistral}, focusing on importing Precision and Recall metrics from image generation to text generation. This …

DiversityImage GenerationText Generation

Diverse Text Generation via Variational Encoder-Decoder Models with Gaussian Process Priors

2022-04-04 · Wanyu Du, Jianqiao Zhao, LiWei Wang, Yangfeng Ji

Generating high quality texts with high diversity is important for many NLG applications, but current methods mostly focus on building deterministic models to generate higher quality texts and do not provide many options…

DecoderDiversityGaussian ProcessesParaphrase Generation+5