no code implementations • CCL 2021 • Xiaobing Zhao, Bo Jin, Yuan Sun
“神经机器翻译在低资源语言的翻译任务中存在翻译难度大、译文质量不佳的问题。本文针对低资源语言与汉语之间没有双语平行语料的情况, 采用正反向枢轴翻译的方法, 生成了三种低资源语言到汉语的平行句对, 采用词汇级的系统融合技术, 将Transformer模型和对偶学习模型翻译生成的目标语言译文进行融合, 然后通过混淆神经网络进行词汇选择, 生成了更为优质的目标语言译文。实验证明, 本文提出的多模型融合方法在爱沙尼亚语-汉语、拉脱维亚语-汉语、罗马尼亚语-汉语这三种低资源语言翻译任务中均优于独立模型的翻译效果, 进一步提升了低资源语言神经机器翻译的译文质量。”
no code implementations • COLING 2022 • Yuan Sun, Sisi Liu, Zhengcuo Dan, Xiaobing Zhao
Then, the types predicted by the question type classifier are fed into the question generator.
no code implementations • CCL 2021 • Yuan Sun, Jiaya Liang, Andong Chen, Xiaobing Zhao
“知识图谱表示学习是自然语言处理的一项关键技术, 现有的知识图谱表示研究主要集中在英语、汉语等语言, 而低资源语言的知识图谱表示学习研究还处于探索阶段, 例如藏语。本文基于前期构建的藏语知识图谱, 提出了一种联合胶囊神经网络(JCapsR)的藏语知识图谱表示学习模型。首先, 我们使用TransR模型生成藏语知识图谱的结构化信息表示。其次, 采用融合多头注意力和关系注意力的Transformer模型表示藏语实体的文本描述信息。最后, 采用JCapsR进一步提取三元组在知识图谱语义空间中的关系, 将实体文本描述信息和结构化信息融合, 得到藏语知识图谱的表示。实验结果表明, 相比基线系统, 联合胶囊神经网络JCapsR模型提高了藏语知识图谱表示学习的效果, 相关研究为其它低资源语言知识图谱表示学习的拓展优化提供了参考借鉴意义。”
no code implementations • CCL 2021 • Yuan Sun, Chaofan Chen, Sisi Liu, Xiaobing Zhao
“机器阅读理解旨在教会机器去理解一篇文章并且回答与之相关的问题。为了解决低资源语言上机器阅读理解模型性能低的问题, 本文提出了一种基于注意力机制的藏文机器阅读理解端到端网络模型Ti-Reader。首先, 为了编码更细粒度的藏文文本信息, 本文将音节和词相结合进行词表示, 然后采用词级注意力机制去关注文本中的关键词, 采用重读机制去捕捉文章和问题之间的语义信息, 采用自注意力机制去匹配问题与答案的隐变量本身, 为答案预测提供更多的线索。最后, 实验结果表明, Ti-Reader模型提升了藏文机器阅读理解的性能, 并且在英文数据集SQuAD上也有较好的表现。”
no code implementations • CCL 2021 • Yuan Sun, Sisi Liu, Chaofan Chen, Zhengcuo Dan, Xiaobing Zhao
“机器阅读理解是通过算法让机器根据给定的上下文回答问题, 从而测试机器理解自然语言的程度。其中, 数据集的构建是机器阅读理解的主要任务。目前, 相关算法模型在大多数流行的英语数据集上都取得了显著的成绩, 甚至超过了人类的表现。但对于低资源语言, 由于缺乏相应的数据集, 机器阅读理解研究还处于起步阶段。本文以藏语为例, 人工构建了藏语机器阅读理解数据集(TibetanQA), 其中包含20000个问题答案对和1513篇文章。本数据集的文章均来自云藏网, 涵盖了自然、文化和教育等12个领域的知识, 问题形式多样且具有一定的难度。另外, 该数据集在文章收集、问题构建、答案验证、回答多样性和推理能力等方面, 均采用严格的流程以确保数据的质量, 同时采用基于语言特征消融输入的验证方法说明了数据集的质量。最后, 本文初步探索了三种经典的英语阅读理解模型在TibetanQA数据集上的表现, 其结果难以媲美人类, 这表明在藏语机器阅读理解任务上还需要更进一步的探索。”
no code implementations • 22 Apr 2024 • Yonghao Dang, Jianqin Yin, Liyuan Liu, Yuan Sun, Yanzhu Hu, Pengxiang Ding
Multi-person pose estimation (MPPE) presents a formidable yet crucial challenge in computer vision.
no code implementations • 18 Apr 2024 • Pivithuru Thejan Amarasinghe, Diem Pham, Binh Tran, Su Nguyen, Yuan Sun, Damminda Alahakoon
This paper introduces a novel approach, evolutionary multi-objective optimisation for fairness-aware self-adjusting memory classifiers, designed to enhance fairness in machine learning algorithms applied to data stream classification.
no code implementations • 1 Feb 2024 • Su Nguyen, Dhananjay Thiruvady, Yuan Sun, Mengjie Zhang
In the proposed algorithm, evolved programs represent variable selectors to be used in the search process of constraint programming, and their fitness is determined by the quality of solutions obtained for training instances.
1 code implementation • 30 Dec 2023 • Wenjun Zhu, Yuan Sun, Jiani Liu, Yushi Cheng, Xiaoyu Ji, Wenyuan Xu
The proliferation of images captured from millions of cameras and the advancement of facial recognition (FR) technology have made the abuse of FR a severe privacy threat.
no code implementations • 14 Dec 2023 • Yuan Sun, Xuan Wang, Yunfan Zhang, Jie Zhang, Caigui Jiang, Yu Guo, Fei Wang
We present a method named iComMa to address the 6D camera pose estimation problem in computer vision.
1 code implementation • 12 Nov 2023 • Shengkun Zhu, Jinshan Zeng, Sheng Wang, Yuan Sun, Zhiyong Peng
Our experiments validate that FLAME, when trained on heterogeneous data, outperforms state-of-the-art methods in terms of model performance.
no code implementations • 2 Nov 2023 • Qingsen Yan, Tao Hu, Yuan Sun, Hao Tang, Yu Zhu, Wei Dong, Luc van Gool, Yanning Zhang
To address this challenge, we formulate the HDR deghosting problem as an image generation that leverages LDR features as the diffusion model's condition, consisting of the feature condition generator and the noise predictor.
1 code implementation • NeurIPS 2023 • Yang Qin, Yuan Sun, Dezhong Peng, Joey Tianyi Zhou, Xi Peng, Peng Hu
Recently, image-text matching has attracted more and more attention from academia and industry, which is fundamental to understanding the latent correspondence across visual and textual modalities.
no code implementations • 23 Sep 2023 • TingYu Zhao, Bo Peng, Yuan Sun, DaiPeng Yang, Zhenguang Zhang, Xi Wu
Recently, advancements in deep learning-based superpixel segmentation methods have brought about improvements in both the efficiency and the performance of segmentation.
no code implementations • 22 Sep 2023 • Pivithuru Thejan Amarasinghe, Su Nguyen, Yuan Sun, Damminda Alahakoon
However, developing an LLM for problem formulation is challenging, due to training data, token limitations, and lack of appropriate performance metrics.
no code implementations • 28 Jun 2023 • Yuan Sun, Nandana Pai, Viswa Vijeth Ramesh, Murtadha Aldeer, Jorge Ortiz
The standard approach is based on a CNN model, which our MLP model outperforms. GeXSe offers two types of explanations: sensor-based activation maps and visual domain explanations using short videos.
no code implementations • 21 Feb 2023 • Yuan Sun, Qiurong Song, Xinning Gui, Fenglong Ma, Ting Wang
Automated machine learning (AutoML) is envisioned to make ML techniques accessible to ordinary users.
no code implementations • 31 Dec 2022 • Xiaofa Liu, Jianqin Yin, Yuan Sun, Zhicheng Zhang, Jin Tang
Unlike most existing methods with offline feature generation, our method directly takes frames as input and further models motion evolution on two different temporal scales. Therefore, we solve the complexity problems of the two stages of modeling and the problem of insufficient temporal and spatial information of a single scale.
no code implementations • 4 Dec 2022 • JUNJIE DENG, Hanru Shi, Xinhe Yu, Wugedele Bao, Yuan Sun, Xiaobing Zhao
To solve the problem of scarcity of datasets on minority languages and verify the effectiveness of the MiLMo model, this paper constructs a minority multilingual text classification dataset named MiTC, and trains a word2vec model for each language.
no code implementations • 26 Nov 2022 • Yuan Sun, Winton Nathan-Roberts, Tien Dung Pham, Ellen Otte, Uwe Aickelin
In biomanufacturing, developing an accurate model to simulate the complex dynamics of bioprocesses is an important yet challenging task.
no code implementations • 26 Nov 2022 • Yuan Sun, Su Nguyen, Dhananjay Thiruvady, XiaoDong Li, Andreas T. Ernst, Uwe Aickelin
Finally, we demonstrate that hybridising the machine learning-based variable ordering methods with traditional domain-based methods is beneficial.
no code implementations • 31 Oct 2022 • Dhananjay Thiruvady, Su Nguyen, Yuan Sun, Fatemeh Shiri, Nayyar Zaidi, XiaoDong Li
While a number of optimisation methods have been proposed to tackle the deterministic problem, the uncertainty associated with resource availability, an inevitable challenge in mining operations, has received less attention.
1 code implementation • 14 Oct 2022 • Jieyi Bi, Yining Ma, Jiahai Wang, Zhiguang Cao, Jinbiao Chen, Yuan Sun, Yeow Meng Chee
Recent neural methods for vehicle routing problems always train and test the deep models on the same instance distribution (i. e., uniform).
no code implementations • 24 Aug 2022 • Wenbin Gan, Yuan Sun
To overcome these issues, in this paper we propose a prerequisite-driven Q-matrix refinement framework for learner knowledge assessment (PQRLKA) in online context.
no code implementations • 15 May 2022 • Yuan Sun, Sisi Liu, JUNJIE DENG, Xiaobing Zhao
Then, we train the Tibetan monolingual pre-trained language model named TiBERT on the data and vocabulary.
no code implementations • 7 Jan 2022 • Zecang Gu, Xiaoqi Sun, Yuan Sun, Fuquan Zhang
Unsupervised clustering algorithm can effectively reduce the dimension of high-dimensional unlabeled data, thus reducing the time and space complexity of data processing.
1 code implementation • 8 Dec 2021 • Yunzhuang Shen, Yuan Sun, XiaoDong Li, Andrew Eberhard, Andreas Ernst
In each iteration of CG, our MLPH leverages an ML model to predict the optimal solution of the pricing problem, which is then used to guide a sampling method to efficiently generate multiple high-quality columns.
1 code implementation • 2 Jul 2021 • Yunzhuang Shen, Yuan Sun, Andrew Eberhard, XiaoDong Li
This paper proposes a novel primal heuristic for Mixed Integer Programs, by employing machine learning techniques.
no code implementations • 18 Dec 2020 • Yuan Sun, Samuel Esler, Dhananjay Thiruvady, Andreas T. Ernst, XiaoDong Li, Kerri Morgan
We investigate an important research question for solving the car sequencing problem, that is, which characteristics make an instance hard to solve?
no code implementations • 10 Dec 2020 • Yi-Yu Lin, Jia-Rui Sun, Yuan Sun
We investigate the relations between bit thread, entanglement distillation and entanglement of purification in the holographic framework.
High Energy Physics - Theory
no code implementations • 2 Dec 2020 • Jun Zhang, Yuan Sun
Besides, we investigate the dependence of the greybody factor and the sparsity of Hawking radiation on the conformal parameters.
General Relativity and Quantum Cosmology
no code implementations • 13 Oct 2020 • Sheng Wang, Yuan Sun, Zhifeng Bao
This paper presents a thorough evaluation of the existing methods that accelerate Lloyd's algorithm for fast k-means clustering.
no code implementations • 29 Jul 2020 • Yuan Sun, Sheng Wang, Yunzhuang Shen, Xiao-Dong Li, Andreas T. Ernst, Michael Kirley
In the first phase of our ML-ACO algorithm, an ML model is trained using a set of small problem instances where the optimal solution is known.
1 code implementation • 12 May 2020 • Yuan Sun, Andreas Ernst, Xiao-Dong Li, Jake Weiner
In this paper, we examine the generalization capability of a machine learning model for problem reduction on the classic travelling salesman problems (TSP).
1 code implementation • ICLR 2019 • Wei Wang, Yuan Sun, Saman Halgamuge
To address this issue, we propose a repulsive loss function to actively learn the difference among the real data by simply rearranging the terms in MMD.
Ranked #19 on Image Generation on STL-10
no code implementations • Transportation Research Record, v 2672, n 45, p 106-114, 2018 • Jianqing Wu, Hao Xu, Yuan Sun, Jianying Zheng, and Rui Yue
Background filtering is the preprocessing step to obtain the HRMTD of different roadway users from roadside LiDAR data.
no code implementations • 11 Apr 2016 • Yuan Sun, Zhen Zhu
Person knowledge extraction is the foundation of the Tibetan knowledge graph construction, which provides support for Tibetan question answering system, information retrieval, information extraction and other researches, and promotes national unity and social stability.