Zobrazeno 1 - 10
of 16
pro vyhledávání: '"Ji, Haozhe"'
Autor:
Ji, Haozhe, Lu, Cheng, Niu, Yilin, Ke, Pei, Wang, Hongning, Zhu, Jun, Tang, Jie, Huang, Minlie
Publikováno v:
Forty-first International Conference on Machine Learning (ICML 2024)
The alignment of language models with human preferences is vital for their application in real-world tasks. The problem is formulated as optimizing the model's policy to maximize the expected reward that reflects human preferences with minimal deviat
Externí odkaz:
http://arxiv.org/abs/2402.00856
Publikováno v:
The Twelfth International Conference on Learning Representations (ICLR 2024)
Despite the remarkable advances in language modeling, current mainstream decoding methods still struggle to generate texts that align with human texts across different aspects. In particular, sampling-based methods produce less-repetitive texts which
Externí odkaz:
http://arxiv.org/abs/2310.01041
Publikováno v:
International Conference on Learning Representations (ICLR 2023)
The standard paradigm of neural language generation adopts maximum likelihood estimation (MLE) as the optimizing method. From a distributional view, MLE in fact minimizes the Kullback-Leibler divergence (KLD) between the distribution of the real data
Externí odkaz:
http://arxiv.org/abs/2302.13344
Despite the success of text-to-text pre-trained models in various natural language generation (NLG) tasks, the generation performance is largely restricted by the number of labeled data in downstream tasks, particularly in data-to-text generation tas
Externí odkaz:
http://arxiv.org/abs/2206.02712
Although Transformers with fully connected self-attentions are powerful to model long-term dependencies, they are struggling to scale to long texts with thousands of words in language modeling. One of the solutions is to equip the model with a recurr
Externí odkaz:
http://arxiv.org/abs/2204.07341
Autor:
Ji, Haozhe, Huang, Minlie
Despite the recent advances in applying pre-trained language models to generate high-quality texts, generating long passages that maintain long-range coherence is yet challenging for these models. In this paper, we propose DiscoDVT, a discourse-aware
Externí odkaz:
http://arxiv.org/abs/2110.05999
Autor:
Ke, Pei, Ji, Haozhe, Ran, Yu, Cui, Xin, Wang, Liwei, Song, Linfeng, Zhu, Xiaoyan, Huang, Minlie
Existing pre-trained models for knowledge-graph-to-text (KG-to-text) generation simply fine-tune text-to-text pre-trained models such as BART or T5 on KG-to-text datasets, which largely ignore the graph structure during encoding and lack elaborate pr
Externí odkaz:
http://arxiv.org/abs/2106.10502
Autor:
Zhang, Zhengyan, Han, Xu, Zhou, Hao, Ke, Pei, Gu, Yuxian, Ye, Deming, Qin, Yujia, Su, Yusheng, Ji, Haozhe, Guan, Jian, Qi, Fanchao, Wang, Xiaozhi, Zheng, Yanan, Zeng, Guoyang, Cao, Huanqi, Chen, Shengqi, Li, Daixuan, Sun, Zhenbo, Liu, Zhiyuan, Huang, Minlie, Han, Wentao, Tang, Jie, Li, Juanzi, Zhu, Xiaoyan, Sun, Maosong
Pre-trained Language Models (PLMs) have proven to be beneficial for various downstream NLP tasks. Recently, GPT-3, with 175 billion parameters and 570GB training data, drew a lot of attention due to the capacity of few-shot (even zero-shot) learning.
Externí odkaz:
http://arxiv.org/abs/2012.00413
Commonsense explanation generation aims to empower the machine's sense-making capability by generating plausible explanations to statements against commonsense. While this task is easy to human, the machine still struggles to generate reasonable and
Externí odkaz:
http://arxiv.org/abs/2009.11753
Despite the success of generative pre-trained language models on a series of text generation tasks, they still suffer in cases where reasoning over underlying commonsense knowledge is required during generation. Existing approaches that integrate com
Externí odkaz:
http://arxiv.org/abs/2009.11692