標(biāo)題: Titlebook: Natural Language Processing and Chinese Computing; 13th National CCF Co Derek F. Wong,Zhongyu Wei,Muyun Yang Conference proceedings 2025 Th [打印本頁] 作者: 使委屈 時(shí)間: 2025-3-21 19:33
書目名稱Natural Language Processing and Chinese Computing影響因子(影響力)
書目名稱Natural Language Processing and Chinese Computing影響因子(影響力)學(xué)科排名
書目名稱Natural Language Processing and Chinese Computing網(wǎng)絡(luò)公開度
書目名稱Natural Language Processing and Chinese Computing網(wǎng)絡(luò)公開度學(xué)科排名
書目名稱Natural Language Processing and Chinese Computing被引頻次
書目名稱Natural Language Processing and Chinese Computing被引頻次學(xué)科排名
書目名稱Natural Language Processing and Chinese Computing年度引用
書目名稱Natural Language Processing and Chinese Computing年度引用學(xué)科排名
書目名稱Natural Language Processing and Chinese Computing讀者反饋
書目名稱Natural Language Processing and Chinese Computing讀者反饋學(xué)科排名
作者: 生命層 時(shí)間: 2025-3-21 23:34 作者: muffler 時(shí)間: 2025-3-22 00:50
LasQ: Largest Singular Components Fine-Tuning for?LLMs with?Quantizationge generation tasks. The experiments show that our method can significantly outperform existing methods with fewer training parameters. Compared with LoftQ and QLoRA methods, it has a 2%–15% improvement, and it can even achieve equivalent LoRA fine-tuning effects and full parameter fine-tuning effects.作者: Cpr951 時(shí)間: 2025-3-22 08:10
Sparse Mixture of?Experts Language Models Excel in?Knowledge Distillationillation using MoE without the necessity of continued pretraining. Experimental results indicate that our approach enhances the model’s capabilities compared to dense model distillation, achieving superior performance across a multitude of tasks. We will release our code at ..作者: 隱語 時(shí)間: 2025-3-22 11:37
Evaluation and?Analysis of?the?Chinese Semantic Dependency Understanding Ability of?Large Language Mderstanding of high-order semantic structure knowledge and semantic relation knowledge. Furthermore, our experiments reveal that while LLMs perform well on the in-domain (ID) test set via SFT, their generalization ability on out-of-domain (OOD) test set remains inadequate.作者: 枯萎將要 時(shí)間: 2025-3-22 15:10 作者: SUGAR 時(shí)間: 2025-3-22 20:15
0302-9743 ing and Chinese Computing, NLPCC 2024, held in Hangzhou, China, during November 2024..The 161 full papers and 33 evaluation workshop papers included in these proceedings were carefully reviewed and selected from 451 submissions. They deal with the following areas: Fundamentals of NLP; Information Ex作者: CALL 時(shí)間: 2025-3-22 21:13
Improving Causal Inference of?Large Language Models with?SCM Toolsools, and combine the inference results of the causal inference tools to generate the final causal question answers. The experimental results show that the method proposed in this paper outperforms the best existing methods.作者: 遭受 時(shí)間: 2025-3-23 04:53 作者: 壕溝 時(shí)間: 2025-3-23 06:02 作者: 誘惑 時(shí)間: 2025-3-23 10:40
Zhenyang Hua,Shuyue Xing,Huixing Jiang,Chen Wei,Xiaojie Wangabel was associated with public sector arrangements observed in countries as diverse as England (O’Toole & Jordan 1995), Sweden (Fortin 1996; Gustafsson and Rhodes 1989), Portugal, Japan (James 2000), The Netherlands (Ter Bogt 1999; Van der Knaap et al. 1997), Latvia (Pollitt et al. 2001; Pollitt 20作者: MILL 時(shí)間: 2025-3-23 15:16
Yufei Li,Chunshan Li,Zecheng Wang,Dianbo Sui,Jianen Yanstimulating computer-supported collaborative-learning inquiry experience. The origin of geometry was a turning point in the pre-history of informatics, literacy, and rational thought. Yet, this triumph of human intellect became ossified through historic layers of systematization, beginning with Eucl作者: 認(rèn)為 時(shí)間: 2025-3-23 21:25 作者: 撕裂皮肉 時(shí)間: 2025-3-24 01:42
Xiang Zhao,Beining Lin,You Songrt commission, even before the Tribunal was established by the Security Council in 1993. These interpreters later formed the core of the ICTY language services during the initial investigative phase. The international community may not have believed that the Tribunal would adjudicate, that there wou作者: Gudgeon 時(shí)間: 2025-3-24 06:01 作者: 泄露 時(shí)間: 2025-3-24 09:32 作者: bibliophile 時(shí)間: 2025-3-24 14:14 作者: 糾纏 時(shí)間: 2025-3-24 15:15
Zizhuo Shen,Wei Li,Yanqiu Shaoheir stories into English are nuanced and varied, as are the politics of literary expression interweaving them. In this chapter, I focus on Alia Mamdouh’s novel .[.] (1986/2000), the first novel by an Iraq woman writer to be published twice in English translation. The first version, .(1995) translat作者: EXALT 時(shí)間: 2025-3-24 23:03 作者: Transfusion 時(shí)間: 2025-3-24 23:57
Biao Ma,Fang Kongins with a concluding summary of the main findings in delivering the representational, the compositional and the interactive meanings in subtitling. Then, it moves on to the discussion of these findings by referring to relevant previous studies and the theoretical framework employed in this book. In作者: Ganglion 時(shí)間: 2025-3-25 05:08
Hang Jiang,Guoyong Cai,Sihui Liins with a concluding summary of the main findings in delivering the representational, the compositional and the interactive meanings in subtitling. Then, it moves on to the discussion of these findings by referring to relevant previous studies and the theoretical framework employed in this book. In作者: Aboveboard 時(shí)間: 2025-3-25 09:35
Yang Jiang,Jun Chen,Kai Han,Yi Liu,Siqi Ma,Yuqing Song,Zhe Liu practitioner lawyers and scholars. However this patchworking process has rarely taken place with attention to methodology, rather it has been arbitrary, depending upon which participants are most influential in a given forum, and what their preference is for specific domestic systems of liability. 作者: 涂掉 時(shí)間: 2025-3-25 13:23
Pengfei Wu,Jiahao Liu,Zhuocheng Gong,Qifan Wang,Jinpeng Li,Jingang Wang,Xunliang Cai,Dongyan Zhaotates as the predominant law-makers, and considers domestic case law to be merely a subsidiary source of law. This is a static view of the law, which does not take into account the actual processes by which certain branches of international law are formed. A more dynamic view is therefore proposed i作者: chisel 時(shí)間: 2025-3-25 16:22
Yuan Wutates as the predominant law-makers, and considers domestic case law to be merely a subsidiary source of law. This is a static view of the law, which does not take into account the actual processes by which certain branches of international law are formed. A more dynamic view is therefore proposed i作者: 搖擺 時(shí)間: 2025-3-25 23:17
Qiuyu Liang,Weihua Wang,Jie Yu,Feilong Baoair towards defendants. Two sets of tensions created by these demands are discussed in this chapter. The first is the tension between efficacy and symbolism; the desire for an efficient system of prosecutions requires a clarity of goals, yet there are numerous and competing goals asserted by those w作者: 不給啤 時(shí)間: 2025-3-26 03:35 作者: carotid-bruit 時(shí)間: 2025-3-26 04:38
Improving Causal Inference of?Large Language Models with?SCM Toolscent study showed the poor ability of LLMs to perform causal inference based on causal graphs and data. In this paper, we propose a method to enhance LLMs based on Structure Causal Model (SCM) tools. We constructed 10 causal inference tools based on SCM theory for solving 10 different types of causa作者: 伴隨而來 時(shí)間: 2025-3-26 09:38 作者: 前奏曲 時(shí)間: 2025-3-26 14:37 作者: Measured 時(shí)間: 2025-3-26 17:53
LasQ: Largest Singular Components Fine-Tuning for?LLMs with?Quantization of fine-tuning the entire model becomes extremely high. To address this challenge, we focus on applying quantization and LoRA fine-tuning together in pre-training scenarios and propose an efficient parameter fine-tuning (PEFT) method, the LasQ (Largest Singular Components Fine-tuning for LLMs with 作者: 吸引人的花招 時(shí)間: 2025-3-27 01:02 作者: CAJ 時(shí)間: 2025-3-27 02:07
What is the?Best Model? Application-Driven Evaluation for?Large Language Models and industry as they generalize foundation models to various practical tasks in a prompt manner. To assist users in selecting the best model in practical application scenarios, i.e., choosing the model that meets the application requirements while minimizing cost, we introduce A-Eval, an applicatio作者: syring 時(shí)間: 2025-3-27 09:15
Sparse Mixture of?Experts Language Models Excel in?Knowledge Distillationn distilling large language models have primarily focused on loss functions and training methodologies, with limited attention given to structural improvements of student models. This is largely due to the challenges posed by cross-architecture distillation and the substantial computational resource作者: 來這真柔軟 時(shí)間: 2025-3-27 10:43 作者: reperfusion 時(shí)間: 2025-3-27 14:47
Reparameterization-Based Parameter-Efficient Fine-Tuning Methods for Large Language Models: A Systemning objectives to achieve unprecedented performance. To fully exploit the potential of LLMs, fine-tuning LLMs on specific downstream tasks is essential. However, traditional full fine-tuning methods pose significant computational challenges, prompting the emergence of Parameter-Efficient Fine-Tunin作者: 自由職業(yè)者 時(shí)間: 2025-3-27 21:28 作者: white-matter 時(shí)間: 2025-3-27 22:21 作者: LAITY 時(shí)間: 2025-3-28 04:54 作者: 學(xué)術(shù)討論會(huì) 時(shí)間: 2025-3-28 06:50
FIRP: Faster LLM Inference via?Future Intermediate Representation Predictionnature of LLM decoding, which generates only a single token per forward propagation, fails to fully exploit the parallel computational power of GPUs, leading to considerable latency. To address this, we introduce a novel speculative decoding method named FIRP which generates multiple tokens instead 作者: colostrum 時(shí)間: 2025-3-28 11:48 作者: perpetual 時(shí)間: 2025-3-28 17:21
Effective Knowledge Graph Embedding with?Quaternion Convolutional Networkshallenges. Firstly, CNN-based models with simple architectures are unable to extract latent features. Secondly, these models enhance feature extraction by adding extra modules, which inevitably increases training cost with limited performance improvement. To address these challenges, we go beyond tr作者: AVID 時(shí)間: 2025-3-28 19:09 作者: Cumbersome 時(shí)間: 2025-3-29 02:45 作者: 填料 時(shí)間: 2025-3-29 06:26 作者: Atheroma 時(shí)間: 2025-3-29 11:19
Zhenyang Hua,Shuyue Xing,Huixing Jiang,Chen Wei,Xiaojie Wang contexts (see OECD 1997a:19). By the turn of century however the success story of agencies had changed. They were no longer regarded as solutions in a whole range of contexts but rather as a problem for accountability and coordination (OECD 2002; James 2004; Smullen 2004).作者: 共同確定為確 時(shí)間: 2025-3-29 13:11
Yufei Li,Chunshan Li,Zecheng Wang,Dianbo Sui,Jianen Yanmputers portend an interactive approach to exploring dynamic geometry as well as broadened prospects for collaboration. The proposed conception of geometry emph978-3-031-01072-9978-3-031-02200-5Series ISSN 1946-7680 Series E-ISSN 1946-7699 作者: enterprise 時(shí)間: 2025-3-29 15:38
Biao Ma,Fang Kongention in previous research. Next, this chapter considers the practical implications deriving from these key findings. Then the research limitations and the future research directions are presented. This chapter ends with a conclusion of the discussion.作者: TATE 時(shí)間: 2025-3-29 20:23 作者: hangdog 時(shí)間: 2025-3-30 02:40 作者: 極微小 時(shí)間: 2025-3-30 05:23 作者: 膽汁 時(shí)間: 2025-3-30 11:56
Pengfei Wu,Jiahao Liu,Zhuocheng Gong,Qifan Wang,Jinpeng Li,Jingang Wang,Xunliang Cai,Dongyan Zhao A comparative perspective on this patchworking can alleviate the current clash of legal cultures observable in the debates on liability, and can aid in making a selection from among all the possible translations from domestic criminal law.作者: 得罪 時(shí)間: 2025-3-30 16:04
Yuan Wu A comparative perspective on this patchworking can alleviate the current clash of legal cultures observable in the debates on liability, and can aid in making a selection from among all the possible translations from domestic criminal law.作者: colloquial 時(shí)間: 2025-3-30 19:47 作者: 強(qiáng)制令 時(shí)間: 2025-3-31 00:43 作者: harpsichord 時(shí)間: 2025-3-31 03:51 作者: Kinetic 時(shí)間: 2025-3-31 05:23 作者: BAN 時(shí)間: 2025-3-31 11:21 作者: 清晰 時(shí)間: 2025-3-31 14:58 作者: condone 時(shí)間: 2025-3-31 20:26 作者: 彎腰 時(shí)間: 2025-3-31 22:41 作者: ensemble 時(shí)間: 2025-4-1 03:35 作者: falsehood 時(shí)間: 2025-4-1 06:55
SACL: Sequential Augmentation with?Curriculum Learning in?Dataset Level and resources. To validate our method, we conduct experiments on the MET-MEME dataset for a multimodal emotion classification task and deeply analyze the impact of curriculum learning on the training results. The results prove the effectiveness of our method.作者: 芳香一點(diǎn) 時(shí)間: 2025-4-1 11:39
Classifiers Guided Controllable Text Generation for?Discrete Diffusion Language Modelsained controllable tasks demonstrate the effectiveness of the proposed algorithms, achieving success rates up to 99% on certain tasks. Furthermore, the inherent multi-modality problem in discrete diffusion models is mitigated by framing multi-modality suppression as a controllable generation task. T作者: atrophy 時(shí)間: 2025-4-1 15:03
AugMixSpeech: A Data Augmentation Method and?Consistency Regularization for?Mandarin Automatic Speec intrinsic features of the data. Extensive experiments on the validation and test of Aishell-1 achieve recognition accuracy of 4.23% and 4.79%, which outperforms existing approaches and demonstrates its effectiveness in automatic speech recognition.
