1 code implementation • 7 Jan 2020 • Ke Ding, Xuanji He, Guanglu Wan
Momentum Contrast (MoCo) is a recently proposed unsupervised representation learning framework, and has shown its effectiveness for learning good feature representation for downstream vision tasks.
no code implementations • 24 Aug 2021 • Yantao Gong, Cao Liu, Jiazhen Yuan, Fan Yang, Xunliang Cai, Guanglu Wan, Jiansong Chen, Ruiyao Niu, Houfeng Wang
To handle this problem, we propose a density-based dynamic curriculum learning model.
no code implementations • 23 Dec 2021 • Qicong Xie, Tao Li, Xinsheng Wang, Zhichao Wang, Lei Xie, Guoqiao Yu, Guanglu Wan
Moreover, the explicit prosody features used in the prosody predicting module can increase the diversity of synthetic speech by adjusting the value of prosody features.
no code implementations • 17 Mar 2022 • Yantao Gong, Cao Liu, Fan Yang, Xunliang Cai, Guanglu Wan, Jiansong Chen, Weipeng Zhang, Houfeng Wang
Experiments on the open datasets verify that our model outperforms the existing calibration methods and achieves a significant improvement on the calibration metric.
1 code implementation • 31 Mar 2022 • Keyu An, Huahuan Zheng, Zhijian Ou, Hongyu Xiang, Ke Ding, Guanglu Wan
The simulation module is jointly trained with the ASR model using a self-supervised loss; the ASR model is optimized with the usual ASR loss, e. g., CTC-CRF as used in our experiments.
no code implementations • 31 Mar 2022 • Huahuan Zheng, Keyu An, Zhijian Ou, Chen Huang, Ke Ding, Guanglu Wan
Based on the DR method, we propose a low-order density ratio method (LODR) by replacing the estimation with a low-order weak language model.
no code implementations • 13 May 2022 • Xiangyu Xi, Chenxu Lv, Yuncheng Hua, Wei Ye, Chaobo Sun, Shuaipeng Liu, Fan Yang, Guanglu Wan
Though widely used in industry, traditional task-oriented dialogue systems suffer from three bottlenecks: (i) difficult ontology construction (e. g., intents and slots); (ii) poor controllability and interpretability; (iii) annotation-hungry.
no code implementations • 7 Nov 2022 • Zhengkun Tian, Hongyu Xiang, Min Li, Feifei Lin, Ke Ding, Guanglu Wan
To reduce the peak latency, we propose a simple and novel method named peak-first regularization, which utilizes a frame-wise knowledge distillation function to force the probability distribution of the CTC model to shift left along the time axis instead of directly modifying the calculation process of CTC loss and gradients.
1 code implementation • 25 Nov 2022 • Xiangyu Xi, Jianwei Lv, Shuaipeng Liu, Wei Ye, Fan Yang, Guanglu Wan
As a pioneering exploration that expands event detection to the scenarios involving informal and heterogeneous texts, we propose a new large-scale Chinese event detection dataset based on user reviews, text conversations, and phone conversations in a leading e-commerce platform for food service.
no code implementations • 3 Apr 2023 • Yuncheng Hua, Xiangyu Xi, Zheng Jiang, Guanwei Zhang, Chaobo Sun, Guanglu Wan, Wei Ye
End-to-end generation-based approaches have been investigated and applied in task-oriented dialogue systems.
1 code implementation • 27 Jun 2023 • Yinyi Wei, Shuaipeng Liu, Hailei Yan, Wei Ye, Tong Mo, Guanglu Wan
Specifically, for an utterance, we generate its future context with pre-trained language models, potentially containing extra beneficial knowledge in a conversational form homogeneous with the historical ones.
no code implementations • 14 Sep 2023 • Lei Zhang, Zhengkun Tian, Xiang Chen, Jiaming Sun, Hongyu Xiang, Ke Ding, Guanglu Wan
To address this issue, we draw inspiration from the multifaceted capabilities of LLMs and Whisper, and focus on integrating multiple ASR text processing tasks related to speech recognition into the ASR model.
no code implementations • 18 Sep 2023 • Song Li, Yongbin You, Xuezhi Wang, Ke Ding, Guanglu Wan
To further expand the applications of multilingual artificial intelligence assistants and facilitate international communication, it is essential to enhance the performance of multilingual speech recognition, which is a crucial component of speech interaction.
1 code implementation • 1 Oct 2023 • Lucen Zhong, Hengtong Lu, Caixia Yuan, Xiaojie Wang, Jiashen Sun, Ke Zeng, Guanglu Wan
A global policy consistency task is designed to capture the multi-turn dialog policy sequential relation, and an act-based contrastive learning task is designed to capture similarities among samples with the same dialog policy.
1 code implementation • 14 Oct 2023 • Hang Shao, Bei Liu, Bo Xiao, Ke Zeng, Guanglu Wan, Yanmin Qian
Various Large Language Models~(LLMs) from the Generative Pretrained Transformer(GPT) family have achieved outstanding performances in a wide range of text generation tasks.
no code implementations • 28 Feb 2024 • Mengjie Ren, Boxi Cao, Hongyu Lin, Cao Liu, Xianpei Han, Ke Zeng, Guanglu Wan, Xunliang Cai, Le Sun
Instruction Fine-tuning~(IFT) is a critical phase in building large language models~(LLMs).
1 code implementation • COLING 2022 • Yinyi Wei, Shuaipeng Liu, Jianwei Lv, Xiangyu Xi, Hailei Yan, Wei Ye, Tong Mo, Fan Yang, Guanglu Wan
Many recent sentence-level event detection efforts focus on enriching sentence semantics, e. g., via multi-task or prompt-based learning.
no code implementations • SemEval (NAACL) 2022 • Cong Chen, Jiansong Chen, Cao Liu, Fan Yang, Guanglu Wan, Jinxiong Xia
Furthermore, we leverage two data augment strategies and auxiliary tasks to improve the performance on few-label data and zero-shot cross-lingual settings.