作者: 矛盾心理 時(shí)間: 2025-3-21 21:51
A Coarse Pipeline to Solve Hierarchical Multi-answer Questions with Conditions, span from the given context. However, in practice, the answer to a question may exist in multiple spans. Besides, multiple answers to a question may be of different granularity and hierarchically related to each other. In this paper, we propose a simple but effective pipeline to solve the hierarchi作者: 友好 時(shí)間: 2025-3-22 02:09
A Pipeline-Based Multimodal Military Event Argument Extraction Framework,The task is to match the textual information extracted from the text corpus with the visual information from images for event argument extraction. For this aim, we introduce a pipeline-based multimodal information extraction framework consisted of three models. The first one is a global pointer mode作者: 吵鬧 時(shí)間: 2025-3-22 07:26 作者: 悠然 時(shí)間: 2025-3-22 08:54 作者: 導(dǎo)師 時(shí)間: 2025-3-22 14:10
Cascaded Solution for Multi-domain Conditional Question Answering with Multiple-Span Answers,ution consists of Data Analysis and Processing, Condition-Answer Extraction, Post-extraction Processing, Condition-Answer Relation Classification, and Post-classification Processing. The rule-based post-extraction and Post-classification Processing modules consist of seven cascaded modules. Because 作者: 導(dǎo)師 時(shí)間: 2025-3-22 20:46 作者: 氣候 時(shí)間: 2025-3-22 22:41 作者: LASH 時(shí)間: 2025-3-23 01:49 作者: initiate 時(shí)間: 2025-3-23 08:01 作者: Sinus-Node 時(shí)間: 2025-3-23 10:17 作者: 嚴(yán)峻考驗(yàn) 時(shí)間: 2025-3-23 17:08
High-Quality Article Classification Based on Named Entities of Knowledge Graph and Multi-head Attend high-quality articles to users. This paper explores how to combine named entities of knowledge graph and multi-head attention mechanism with quality article identification, not only the contents of articles. For the article classification task of CCKS 2022, we proposed an incorporating named entit作者: flaggy 時(shí)間: 2025-3-23 20:01
Implementation and Optimization of Graph Computing Algorithms Based on Graph Database,e works can be generalized to graph computing and graph analysis tasks, such as shortest path search, hop-constrained reachability, PageRank, triangle counting, and closeness centrality computation. However, existing graph database query language (SPARQL, Gremlin, etc.) dose not implement these algo作者: 共同給與 時(shí)間: 2025-3-24 00:43 作者: 我就不公正 時(shí)間: 2025-3-24 04:20
Knowledge-Enhanced Classification: A Scheme for Identification of High-Quality Articles,s classification problem, from TF-IDF to word2vec, then to RNN and LSTM, and now to transformer-based models, such as Bert, have achieved great improvement in NLU tasks. However, for many specific problems, such as recognition of high-quality article, directly inputting the text content into the tra作者: Licentious 時(shí)間: 2025-3-24 07:18 作者: 慢慢沖刷 時(shí)間: 2025-3-24 11:01
,Learning to?Answer Complex Visual Questions from?Multi-View Analysis,y used in textbooks such as diagrams often contain complicated and abstract information (e.g. constructed graphs with logic and concepts). Therefore, Diagram Question answering (DQA) is a challenging but significant task, which is also helpful for machines to understand human cognitive behaviors and作者: 感情 時(shí)間: 2025-3-24 17:46
A Prompt-Based UIE Framework,y NLP tasks can be categorized as information extraction tasks, such as named entity extraction (NER), relation extraction (RE), event extraction (EE), etc. To dealing with different IE tasks of different situation, we propose a prompt-based universal information extraction framework which is friend作者: Etymology 時(shí)間: 2025-3-24 21:47
,Multi-modal Representation Learning with?Self-adaptive Threshold for?Commodity Verification,xt. By definition, identical commodities are those that have identical key attributes and are cognitively identical to consumers. There are two main challenges: 1) The extraction and fusion of multi-modal representation. 2) The ability to verify identical commodities by comparing the similarity betw作者: 親屬 時(shí)間: 2025-3-25 01:48 作者: lymphedema 時(shí)間: 2025-3-25 05:11 作者: OFF 時(shí)間: 2025-3-25 08:57 作者: 聯(lián)合 時(shí)間: 2025-3-25 13:17 作者: infelicitous 時(shí)間: 2025-3-25 16:21 作者: Prostatism 時(shí)間: 2025-3-25 20:33
https://doi.org/10.1007/978-3-8349-8168-4oss Entropy constraint Decoding”, which can effectively constrain the content generated by the text when performing multiple selection tasks. This method has obtained SOTA in the evaluation task of CCKS-2022, which fully proves the effectiveness of the method.作者: GLUT 時(shí)間: 2025-3-26 02:53 作者: 無(wú)法破譯 時(shí)間: 2025-3-26 08:14
Cascaded Solution for Multi-domain Conditional Question Answering with Multiple-Span Answers,is used for relation extraction of conditions, coarse-grained answers, and fine-grained answers, and the constraint extraction method is based on rules. The proposed solution obtains an F1 value of 0.74487 on the test set (ranking 3rd), and its effectiveness in multi-domain scenarios is verified.作者: Living-Will 時(shí)間: 2025-3-26 10:51
High-Quality Article Classification Based on Named Entities of Knowledge Graph and Multi-head Attennbalanced samples. Then we selected the top three results in terms of F1 scores for voting fusion. Finally, data augmentation was applied in accordance with certain rules in the fusion results. Our method achieves F1 score of 86.51% on the test set of the task and ranks the second place in the competition.作者: inventory 時(shí)間: 2025-3-26 12:50
Knowledge-Enhanced Classification: A Scheme for Identification of High-Quality Articles, achieved 83.6 F1-score in the official test set and ranked first among all teams in task 2 of CCKS-2022. This paper is divided into four parts: 1) The introduction of our task; 2) Main ideas of our model; 3) Other innovation strategies; 4) Experiments and result.作者: 的事物 時(shí)間: 2025-3-26 18:06 作者: 自由職業(yè)者 時(shí)間: 2025-3-26 22:04 作者: Curmudgeon 時(shí)間: 2025-3-27 01:42
Conference proceedings 2022k place in Qinhuangdao, China, in August 2022.?.The 25 full papers presented in this volume were carefully reviewed and selected from 42 submissions. CCKS technology evaluation track aims to provide researchers with platforms and resources for testing knowledge and semantic computing technologies, a作者: Multiple 時(shí)間: 2025-3-27 08:04 作者: invade 時(shí)間: 2025-3-27 09:33
Georg Dettmar (Generalsekret?r) which we propose a search-enhanced path mining and ranking method. The method divides the process of CKBQA into four parts: question classification, principal entity extraction, search-enhanced candidate path mining and candidate path ranking. Finally, both the preliminary and final evaluation results prove the effectiveness of our method.作者: amplitude 時(shí)間: 2025-3-27 14:10 作者: 裂縫 時(shí)間: 2025-3-27 21:28 作者: 詢問(wèn) 時(shí)間: 2025-3-28 00:25 作者: 粘連 時(shí)間: 2025-3-28 05:25
Diagram Question Answering with Joint Training and Bottom-Up and Top-Down Attention,y regions of interest to questions and use a same model to jointly train multiple choice questions and true false questions. Our approach on test dataset of official CCKS2022 textbook diagram question answering session achieves the accuracy of 58.09%.作者: Optic-Disk 時(shí)間: 2025-3-28 08:58
High Quality Article Recognition Based on Ernie and Knowledge Mapping,he basis of a deeper semantic understanding of the article. We mainly use Ernie pre training model and Article title - knowledge map - the modeling method of part of the article content, the use of pseudo label data sets and other ways to get the predicted value, and finally achieved the result of 0.82 in the B list.作者: Hdl348 時(shí)間: 2025-3-28 10:26
1865-0929 ologies, algorithms and systems, promote the technical development in the field of domestic knowledge, and the integration of academic achievements and industrial needs..978-981-19-8299-6978-981-19-8300-9Series ISSN 1865-0929 Series E-ISSN 1865-0937 作者: 自愛(ài) 時(shí)間: 2025-3-28 16:58 作者: 鐵塔等 時(shí)間: 2025-3-28 20:47 作者: 剛開(kāi)始 時(shí)間: 2025-3-29 00:42 作者: 空中 時(shí)間: 2025-3-29 06:26 作者: Rheumatologist 時(shí)間: 2025-3-29 09:44 作者: granite 時(shí)間: 2025-3-29 13:00
Compound Property Prediction Based on Multiple Different Molecular Features and Ensemble Learning,ations generated based on different ways were concatenated, and they were input into the ensemble model for prediction. Finally, the score of 0.8985 was obtained in the test dataset, and won the first place.作者: Vldl379 時(shí)間: 2025-3-29 18:42 作者: 送秋波 時(shí)間: 2025-3-29 21:47 作者: 防銹 時(shí)間: 2025-3-30 02:27 作者: surrogate 時(shí)間: 2025-3-30 07:44
https://doi.org/10.1007/978-3-658-30515-4 present a three-stage approach leveraging translation model to this benchmark. Our approach outperforms in the benchmark, which reaches 0.9320 as the precision score ranking the first place on the leaderboard.作者: 網(wǎng)絡(luò)添麻煩 時(shí)間: 2025-3-30 10:20 作者: isotope 時(shí)間: 2025-3-30 14:58
,A Translation Model-Based Question Answering Approach over?Cross-Lingual Knowledge Graphs, present a three-stage approach leveraging translation model to this benchmark. Our approach outperforms in the benchmark, which reaches 0.9320 as the precision score ranking the first place on the leaderboard.作者: NUL 時(shí)間: 2025-3-30 19:02 作者: 美色花錢 時(shí)間: 2025-3-31 00:24
https://doi.org/10.1007/978-3-662-33888-9otential of machine learning techniques, especially deep learning. This paper presents our proposed solution for CCKS-2022 task 8, a chemical domain knowledge-aware framework for multi-view molecular property prediction. As a generative self-supervised approach to molecular graph representation lear作者: AWRY 時(shí)間: 2025-3-31 02:53 作者: landmark 時(shí)間: 2025-3-31 06:55 作者: 不可磨滅 時(shí)間: 2025-3-31 13:01
Georg Dettmar (Generalsekret?r)g. In order to enhance the application of cross-lingual knowledge, the 16th China Conference on Knowledge Graph and Semantic Computing (CCKS2022) has released a cross-lingual knowledge base question answering (CKBQA) task. This paper presents the submission of our team (HW-TSC) to the CKBQA task, in作者: reject 時(shí)間: 2025-3-31 13:21
https://doi.org/10.1007/978-3-658-30515-4topic like QA over cross-lingual knowledge graphs. CCKS2022 holds a benchmark competition on QA over cross-lingual knowledge graphs. In this paper, we present a three-stage approach leveraging translation model to this benchmark. Our approach outperforms in the benchmark, which reaches 0.9320 as the作者: 古代 時(shí)間: 2025-3-31 18:47
https://doi.org/10.1007/978-3-658-30515-4ution consists of Data Analysis and Processing, Condition-Answer Extraction, Post-extraction Processing, Condition-Answer Relation Classification, and Post-classification Processing. The rule-based post-extraction and Post-classification Processing modules consist of seven cascaded modules. Because 作者: 熱情贊揚(yáng) 時(shí)間: 2025-3-31 21:54
Nachtragsmanagement in der Baupraxisproperties prediction..For this task, we proposed to generate vector representations of chemical molecules by using molecular descriptors and pharmacophore fingerprints, and using large-scale chemical molecular data for unsupervised training to generate vector representations of chemical molecules. 作者: 集聚成團(tuán) 時(shí)間: 2025-4-1 03:57 作者: 螢火蟲(chóng) 時(shí)間: 2025-4-1 07:00
Ulrich Elwert,Alexander Flassakg processes such as diagram element detection. However, due to low resource constraints, achieving efficient extraction of diagram elements is challenging. In addition, vision tasks rely on image feature extraction, and most feature extraction today is based on real scenario images on ImageNet. To s作者: Venules 時(shí)間: 2025-4-1 12:18 作者: 單色 時(shí)間: 2025-4-1 18:01
Die Narkose und ihre allgemeine Theoriey high-quality articles and distribute them to users has important research significance and practical application value. The task of this competition is to introduce external Knowledge mapping, combined with the internal knowledge logic of the article, realizes high-quality article recognition on t
