no code implementations • CCL 2022 • Zhuo Li, Zhijuan Wang, Xiaobing Zhao
“机器音译是基于语音相似性自动将文本从一种语言转换为另一种语言的过程, 它是机器翻译的一个子任务, 侧重于语音信息的翻译。音译后可知道源单词在另一种语言中的发音, 使不熟悉源语言的人更容易理解该语言, 有益于消除语言和拼写障碍。机器音译在多语言文本处理、语料库对齐、信息抽取等自然语言应用中发挥着重要作用。本文阐述了目前机器音译任务中存在的挑战, 对主要的音译方法进行了剖析、分类和整理, 对音译数据集进行了罗列汇总, 并列出了常用的音译效果评价指标, 最后对该领域目前存在的问题进行了说明并对音译学的未来进行了展望。本文以期对进入该领域的新人提供快速的入门指南, 或供其他研究者参考。”
no code implementations • 26 Mar 2025 • Xuying Li, Zhuo Li, Yuji Kosuga, Victor Bian
In this work, we propose a Group Relative Policy Optimization (GRPO) framework with a multi-label reward regression model to achieve safe and aligned language generation.
no code implementations • 2 Mar 2025 • Xuying Li, Zhuo Li, Yuji Kosuga, Victor Bian
Large Language Models (LLMs) have demonstrated strong reasoning capabilities, but their safety under adversarial conditions remains a challenge.
no code implementations • 16 Feb 2025 • Yuhao Du, Zhuo Li, Pengyu Cheng, Zhihong Chen, Yuejiao Xie, Xiang Wan, Anningzhe Gao
More specifically, by directly minimizing the distribution gap between the learning LLM policy and the optimal solution of RLHF, we transform the alignment objective into a reward-driven re-weighted supervised fine-tuning (SFT) form, which only requires minor adjustment on the SFT loss to obtain noticeable improvement on training stability and effectiveness.
no code implementations • 19 Dec 2024 • Junjia Liu, Zhuo Li, Minghao Yu, Zhipeng Dong, Sylvain Calinon, Darwin Caldwell, Fei Chen
Humanoid robots are envisioned as embodied intelligent agents capable of performing a wide range of human-level loco-manipulation tasks, particularly in scenarios requiring strenuous and repetitive labor.
no code implementations • 5 Dec 2024 • Xuying Li, Zhuo Li, Yuji Kosuga, Yasuhiro Yoshida, Victor Bian
AI agents, powered by large language models (LLMs), have transformed human-computer interactions by enabling seamless, natural, and context-aware communication.
no code implementations • 22 Nov 2024 • Zhuo Li, Mingshuang Luo, Ruibing Hou, Xin Zhao, Hao liu, Hong Chang, Zimo Liu, Chen Li
Human motion generation plays a vital role in applications such as digital humans and humanoid robot control.
no code implementations • 2 Oct 2024 • Xuying Li, Zhuo Li, Yuji Kosuga, Yasuhiro Yoshida, Victor Bian
Large language models (LLMs) have demonstrated remarkable capabilities, but they also pose risks related to the generation of toxic or harmful content.
no code implementations • 3 Sep 2024 • Zhuo Li, Yuhao Du, Jinpeng Hu, Xiang Wan, Anningzhe Gao
To address these challenges, we introduce a self-instructed in-context learning framework that empowers LLMs to deliver more effective responses by generating reliable derived prompts to construct informative contextual environments.
1 code implementation • 30 Aug 2024 • Mouxiang Chen, Lefei Shen, Zhuo Li, Xiaoyun Joy Wang, Jianling Sun, Chenghao Liu
Surprisingly, without further adaptation in the time-series domain, the proposed VisionTS could achieve superior zero-shot forecasting performance compared to existing TSF foundation models.
no code implementations • 27 Aug 2024 • Yuhao Du, Zhuo Li, Pengyu Cheng, Xiang Wan, Anningzhe Gao
Given a particular harmful answer, Atoxia generates a corresponding user query and a misleading answer opening to examine the internal defects of a given LLM.
1 code implementation • 25 May 2024 • Mingshuang Luo, Ruibing Hou, Zhuo Li, Hong Chang, Zimo Liu, YaoWei Wang, Shiguang Shan
Third, M$^3$GPT learns to model the connections and synergies among various motion-relevant tasks.
no code implementations • 10 Apr 2024 • Zhuo Li, He Zhao, Zhen Li, Tongliang Liu, Dandan Guo, Xiang Wan
To solve the joint issue of long-tailed distribution and label noise, most previous works usually aim to design a noise detector to distinguish the noisy and clean samples.
no code implementations • 23 Feb 2024 • Zhisheng Lin, Han Fu, Chenghao Liu, Zhuo Li, Jianling Sun
However, current approaches typically either train adapters on individual tasks or distill shared knowledge from source tasks, failing to fully exploit task-specific knowledge and the correlation between source and target tasks.
no code implementations • 4 Feb 2024 • Qiheng Mao, Zemin Liu, Chenghao Liu, Zhuo Li, Jianling Sun
This collaboration harnesses the sophisticated linguistic capabilities of LLMs to improve the contextual understanding and adaptability of graph models, thereby broadening the scope and potential of GRL.
no code implementations • 25 Jan 2024 • Butian Xiong, Zhuo Li, Zhen Li
We introduce a novel large-scale scene reconstruction benchmark using the newly developed 3D representation approach, Gaussian Splatting, on our expansive U-Scene dataset.
2 code implementations • 23 Oct 2023 • Mouxiang Chen, Lefei Shen, Han Fu, Zhuo Li, Jianling Sun, Chenghao Liu
In this paper, we introduce a universal calibration methodology for the detection and adaptation of CDS with a trained model.
no code implementations • 29 Sep 2023 • Yuxiang Zhang, Zhuo Li, Jingze Lu, Wenchao Wang, Pengyuan Zhang
Based on these analyzes, an SSD method based on temporal consistency and distribution of speaker features is proposed.
1 code implementation • 27 Sep 2023 • Mouxiang Chen, Chenghao Liu, Zemin Liu, Zhuo Li, Jianling Sun
Unbiased Learning to Rank (ULTR) aims to train unbiased ranking models from biased click logs, by explicitly modeling a generation process for user behavior and fitting click data based on examination hypothesis.
no code implementations • 21 Sep 2023 • Yuxiang Zhang, Zhuo Li, Jingze Lu, Hua Hua, Wenchao Wang, Pengyuan Zhang
First, the reasons for the impact are explored, including the proportion of silence duration and the content of silence.
no code implementations • 19 Sep 2023 • Zhuo Li, Bo Li
Second, to generate the initial visible-NIR complementarity weight map, the difference maps of visible and NIR are filtered by the extend-DoG filter.
1 code implementation • journal 2023 • Jiajun Song, Zhuo Li, Weiqing Min, Shuqiang Jiang
Therefore, it is challenging to study the generalization of the model in food image retrieval.
no code implementations • 20 Aug 2023 • Yechen Zhang, Jian Sun, Gang Wang, Zhuo Li, Wei Chen
Discrete reinforcement learning (RL) algorithms have demonstrated exceptional performance in solving sequential decision tasks with discrete action spaces, such as Atari games.
no code implementations • 22 May 2023 • Zhuo Li, Jingze Lu, Zhenduo Zhao, Wenchao Wang, Pengyuan Zhang
Utilizing the large-scale unlabeled data from the target domain via pseudo-label clustering algorithms is an important approach for addressing domain adaptation problems in speaker verification tasks.
1 code implementation • 6 May 2023 • Kechi Zhang, Zhuo Li, Jia Li, Ge Li, Zhi Jin
Inspired by the process of human programming, we propose a generate-and-edit approach named Self-Edit that utilizes execution results of the generated code from LLMs to improve the code quality on the competitive programming task.
1 code implementation • 14 Mar 2023 • Kechi Zhang, Zhuo Li, Zhi Jin, Ge Li
Furthermore, we propose the Hierarchy Transformer (HiT), a simple but effective sequence model to incorporate the complete hierarchical embeddings of source code into a Transformer model.
1 code implementation • 15 Feb 2023 • Shihan Liu, Junlin Zha, Jian Sun, Zhuo Li, Gang Wang
This paper proposes an efficient, low-complexity and anchor-free object detector based on the state-of-the-art YOLO framework, which can be implemented in real time on edge computing platforms.
no code implementations • 27 Jan 2023 • Zhuo Li, Derui Zhu, Yujing Hu, Xiaofei Xie, Lei Ma, Yan Zheng, Yan Song, Yingfeng Chen, Jianjun Zhao
Generally, episodic control-based approaches are solutions that leverage highly-rewarded past experiences to improve sample efficiency of DRL algorithms.
1 code implementation • 16 Dec 2022 • Yujing Wang, Yaming Yang, Zhuo Li, Jiangang Bai, Mingliang Zhang, Xiangtai Li, Jing Yu, Ce Zhang, Gao Huang, Yunhai Tong
To the best of our knowledge, this is the first work that explicitly models the layer-wise evolution of attention maps.
no code implementations • 31 Oct 2022 • Jia Li, Zhuo Li, Huangzhao Zhang, Ge Li, Zhi Jin, Xing Hu, Xin Xia
The attackers aim to inject insidious backdoors into models by poisoning the training data with poison samples.
1 code implementation • 31 Oct 2022 • Jia Li, Ge Li, Zhuo Li, Zhi Jin, Xing Hu, Kechi Zhang, Zhiyi Fu
Pre-trained models are first pre-trained with pre-training tasks and fine-tuned with the code editing task.
no code implementations • 13 Oct 2022 • Yuxiang Zhang, Jingze Lu, Xingming Wang, Zhuo Li, Runqiu Xiao, Wenchao Wang, Ming Li, Pengyuan Zhang
The overfitting of the model to the training set leads to extreme values of the scores and low correlation of the score distributions, which makes score fusion difficult.
no code implementations • 5 Aug 2022 • Dandan Guo, Zhuo Li, Meixi Zheng, He Zhao, Mingyuan Zhou, Hongyuan Zha
Specifically, we view the training set as an imbalanced distribution over its samples, which is transported by OT to a balanced distribution obtained from the meta set.
no code implementations • 5 Jul 2022 • Zhihao Yuan, Xu Yan, Zhuo Li, Xuhao Li, Yao Guo, Shuguang Cui, Zhen Li
Recent progress in 3D scene understanding has explored visual grounding (3DVG) to localize a target object through a language description.
no code implementations • 1 Jul 2022 • Yuxiang Zhang, Zhuo Li, Wenchao Wang, Pengyuan Zhang
Based on the assumption that there is a correlation between anti-spoofing and speaker verification, a Total-Divide-Total integrated Spoofing-Aware Speaker Verification (SASV) system based on pre-trained automatic speaker verification (ASV) system and integrated scoring module is proposed and submitted to the SASV 2022 Challenge.
no code implementations • 19 May 2022 • Minghui Pang, Qiuming Zhu, Cheng-Xiang Wang, Zhipeng Lin, Junyu Liu, Chongyu Lv, Zhuo Li
Path probability prediction is essential to describe the dynamic birth and death of propagation paths, and build the accurate channel model for air-to-ground (A2G) communications.
1 code implementation • ACL 2022 • Jinpeng Hu, Zhuo Li, Zhihong Chen, Zhen Li, Xiang Wan, Tsung-Hui Chang
To address the limitation, we propose a unified framework for exploiting both extra knowledge and the original findings in an integrated way so that the critical information (i. e., key words and their relations) can be extracted in an appropriate way to facilitate impression generation.
no code implementations • 6 Sep 2021 • Minghui Pang, Qiuming Zhu, Fei Bai, Zhuo Li, Hanpeng Li, Kai Mao, Yue Tian
Based on the three-dimensional propagation characteristic under built-up scenarios, a height-dependent line-of-sight (LoS) probability model for air-to-ground (A2G) millimeter wave (mmWave) communications is proposed in this paper.
no code implementations • 11 Mar 2021 • Zhuo Li, Xu Zhou, Junruo Gao, Yifang Qin
Under the constraint of the best efficiency of migration in the whole and without migration conflict, selecting multiple sets of triples based on reinforcement learning, as the final migration of this round to attain the global optimal controller load balancing with minimum cost.
Networking and Internet Architecture
no code implementations • 11 Mar 2021 • Zhuo Li, Xu Zhou, Taixin Li, Yang Liu
With the mass deployment of computing-intensive applications and delay-sensitive applications on end devices, only adequate computing resources can meet differentiated services' delay requirements.
Edge-computing
Networking and Internet Architecture
2 code implementations • 9 Feb 2021 • Zhuo Li, Weiqing Min, Jiajun Song, Yaohui Zhu, Liping Kang, Xiaoming Wei, Xiaolin Wei, Shuqiang Jiang
Limited by the definition of AP, such methods consider both negative and positive instances ranking before each positive instance.
Ranked #3 on
Vehicle Re-Identification
on VehicleID Large
no code implementations • 18 Aug 2020 • Hai Pham, Amrith Setlur, Saket Dingliwal, Tzu-Hsiang Lin, Barnabas Poczos, Kang Huang, Zhuo Li, Jae Lim, Collin McCormack, Tam Vu
Despite the advent of deep learning in computer vision, the general handwriting recognition problem is far from solved.
no code implementations • 11 Feb 2020 • Han Fu, Yunyu Bai, Zhuo Li, Jun Shen, Jianling Sun
Paper documents are widely used as an irreplaceable channel of information in many fields, especially in financial industry, fostering a great amount of demand for systems which can convert document images into structured data representations.
1 code implementation • 19 Dec 2019 • Jian Peng, Bo Tang, Hao Jiang, Zhuo Li, Yinjie Lei, Tao Lin, Haifeng Li
It is due to two facts: first, as the model learns more tasks, the intersection of the low-error parameter subspace satisfying for these tasks becomes smaller or even does not exist; second, when the model learns a new task, the cumulative error keeps increasing as the model tries to protect the parameter configuration of previous tasks from interference.
no code implementations • 23 Nov 2019 • Huibing Wang, Yang Wang, Zhao Zhang, Xianping Fu, Zhuo Li, Mingliang Xu, Meng Wang
With the popularity of multimedia technology, information is always represented or transmitted from multiple views.
no code implementations • 20 Jan 2019 • Li Chen, Hailun Ding, Qi Li, Zhuo Li, Jian Peng, Haifeng Li
Understanding the internal representations of deep neural networks (DNNs) is crucal to explain their behavior.
1 code implementation • 4 Dec 2018 • Jian Peng, Jiang Hao, Zhuo Li, Enqiang Guo, Xiaohong Wan, Deng Min, Qing Zhu, Haifeng Li
In this paper, we proposed a Soft Parameters Pruning (SPP) strategy to reach the trade-off between short-term and long-term profit of a learning model by freeing those parameters less contributing to remember former task domain knowledge to learn future tasks, and preserving memories about previous tasks via those parameters effectively encoding knowledge about tasks at the same time.
no code implementations • 27 Nov 2018 • Li Chen, Hailun Ding, Qi Li, Zhuo Li, Jian Peng, Haifeng Li
Understanding the internal representations of deep neural networks (DNNs) is crucal to explain their behavior.