Výsledky vyhledávání - "Parnell, Jacob"

Report

SumTra: A Differentiable Pipeline for Few-Shot Cross-Lingual Summarization

Autor: Parnell, Jacob, Unanue, Inigo Jauregi, Piccardi, Massimo

Cross-lingual summarization (XLS) generates summaries in a language different from that of the input documents (e.g., English to Spanish), allowing speakers of the target language to gain a concise view of their content. In the present day, the predo

Externí odkaz: http://arxiv.org/abs/2403.13240

Zobrazit plný text záznamu

Report

A Multi-Document Coverage Reward for RELAXed Multi-Document Summarization

Autor: Parnell, Jacob, Unanue, Inigo Jauregi, Piccardi, Massimo

Multi-document summarization (MDS) has made significant progress in recent years, in part facilitated by the availability of new, dedicated datasets and capacious language models. However, a standing limitation of these models is that they are traine

Externí odkaz: http://arxiv.org/abs/2203.02894

Zobrazit plný text záznamu

Report

RewardsOfSum: Exploring Reinforcement Learning Rewards for Summarisation

Autor: Parnell, Jacob, Unanue, Inigo Jauregi, Piccardi, Massimo

To date, most abstractive summarisation models have relied on variants of the negative log-likelihood (NLL) as their training objective. In some cases, reinforcement learning has been added to train the models with an objective that is closer to thei

Externí odkaz: http://arxiv.org/abs/2106.04080

Zobrazit plný text záznamu

Report

BERTTune: Fine-Tuning Neural Machine Translation with BERTScore

Autor: Unanue, Inigo Jauregi, Parnell, Jacob, Piccardi, Massimo

Neural machine translation models are often biased toward the limited translation references seen during training. To amend this form of overfitting, in this paper we propose fine-tuning the models with a novel training objective based on the recentl

Externí odkaz: http://arxiv.org/abs/2106.02208

Zobrazit plný text záznamu