Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence最新文献

A Semi-Parametric Method for Text-to-Image Synthesis from Prior Knowledge 基于先验知识的文本到图像合成半参数方法

Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence Pub Date : 2022-12-23 DOI: 10.1145/3579654.3579717

Jiadong Liang

{"title":"A Semi-Parametric Method for Text-to-Image Synthesis from Prior Knowledge","authors":"Jiadong Liang","doi":"10.1145/3579654.3579717","DOIUrl":"https://doi.org/10.1145/3579654.3579717","url":null,"abstract":"Text-to-image synthesis adopts only text descriptions as input to generate consistent images which should have high visual quality and be semantically aligned with the input text. Compared to images, the textual semantics is ambiguous and sparse, which makes it challenging to map features directly and accurately from text space to image space. To address this issue, the intuitive method is to construct an intermediate space connecting text and image. Using layout as a bridge between text and image not only mitigates the difficulty of the task, but also constrains the spatial distribution of objects in the generated images, which is crucial to the quality of synthesized images. In this paper, we build a two-stage framework for text-to-image synthesis, i.e., Layout Searching by Text Matching, and Layout-to-Image Synthesis with Fine-Grained Textual Semantic Injection. Specifically, we build the prior layout knowledge from the training dataset and propose a semi-parametric layout searching strategy to retrieve the layout that matches the input sentence by measuring the semantic distance between different textual descriptions. In the stage of layout-to-image synthesis, we construct the Textual and Spatial Alignment Generative Adversarial Networks (TSAGANs) that are designed to guarantee the fine-grained alignment of the generated images with the input text and layout obtained in the first stage. Extensive experiments conducted on the COCO-stuff dataset manifest that our method can obtain more reasonable layouts and improve the performance of synthesized images significantly.","PeriodicalId":146783,"journal":{"name":"Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114676135","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A new decoding method of the grammatical evolution 一种新的语法演变解码方法

Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence Pub Date : 2022-12-23 DOI: 10.1145/3579654.3579757

Zhengheng Yan, Pei He, Wei-Zhong Huang, Su Liu

{"title":"A new decoding method of the grammatical evolution","authors":"Zhengheng Yan, Pei He, Wei-Zhong Huang, Su Liu","doi":"10.1145/3579654.3579757","DOIUrl":"https://doi.org/10.1145/3579654.3579757","url":null,"abstract":"GE (grammatical Evolution) is an evolutionary algorithm that evolves programs in an arbitrary language using a variable-length binary string. The binary genome determines which production rules from the grammar definition of the Backs-Naur form(BNF) are used for the program in the genotype-phenotype mapping process. However, the matter relating to the completeness of individual phenotype throughout population formation isn't well self-addressed, so limiting both the convergence speed and the accuracy of the evaluations to some extent. In this paper, we propose an improved GE algorithm that aims to guarantee the completeness of individual phenotypes during initialization as well as subsequent evolution. A comparison of the improved algorithm (IGE) with the classical GE algorithm (CGE) and NGE(integer-coded grammatical evolution) conducted on the symbolic regression problems shows that the improved algorithm (IGE) not only reduces the search space and improves the accuracy of the algorithm, but also speeds up the convergence of the algorithm in constructing both. Definition 1. “Incomplete”: An individual is called an incomplete individual if its corresponding sentential form contains a non-terminal symbol, and its corresponding mapping process is called incomplete mapping. Definition 2. \"Recursive\" production rule: Refers to the BNF in which the same non-terminal symbol appears on the left and right sides of the production rule. Definition 3. Combined production rule: For a production rule with a non-terminal symbol in the right part, if it can be followed by only one type of production rule, it is said that the production rule and its subsequent followable production rule are combined production rules. Definition 4. Non-combined production rule: Production rule except for combined production rule","PeriodicalId":146783,"journal":{"name":"Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127357771","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A probabilistic power flow prediction and situation awareness method based on long and short term memory network 一种基于长短期记忆网络的概率潮流预测与态势感知方法

Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence Pub Date : 2022-12-23 DOI: 10.1145/3579654.3579746

Xu Lin, Xinlei Cai, Jinzhou Zhu, Yanlin Cui, Xinglang Xie

{"title":"A probabilistic power flow prediction and situation awareness method based on long and short term memory network","authors":"Xu Lin, Xinlei Cai, Jinzhou Zhu, Yanlin Cui, Xinglang Xie","doi":"10.1145/3579654.3579746","DOIUrl":"https://doi.org/10.1145/3579654.3579746","url":null,"abstract":"With the massive integration of renewable energy into the power system, the randomness and volatility of power generation in the power system are increasing day by day. These characteristics have a great impact on the direction and size of power flow in the power grid. This paper presents a probabilistic power flow prediction and situation awareness method based on long and short term memory network. This paper first introduces the probability model based on wind power generation, photovoltaic power generation, demand side load, electric vehicle charging, generator set; Secondly, based on the NATAF transformation method, several probability models are transformed by de-correlation standard normal distribution, and a probability scheduling model with minimum cost of multi-party coordination and complementarity is established. Then, a probabilistic power flow solution based on long short-term memory network is proposed for the probabilistic scheduling model. Finally, an actual power grid is taken as an example to verify and compare the proposed algorithms, and the results prove the effectiveness of the proposed methods.","PeriodicalId":146783,"journal":{"name":"Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122121746","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Based on Navmesh to implement AI intelligent pathfinding in three-dimensional maps in UE4 基于Navmesh在UE4中实现三维地图的AI智能寻径

Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence Pub Date : 2022-12-23 DOI: 10.1145/3579654.3579752

Xinyuan Zhang, Xinyou Zhang

{"title":"Based on Navmesh to implement AI intelligent pathfinding in three-dimensional maps in UE4","authors":"Xinyuan Zhang, Xinyou Zhang","doi":"10.1145/3579654.3579752","DOIUrl":"https://doi.org/10.1145/3579654.3579752","url":null,"abstract":"Abstract: The development of science and technology has led to the development of the game industry. As the mainstream development engine of the current three-dimensional game, UE4 is popular with game developers for its powerful rendering technology and model processing technology. There will be a large number of non-player characters (NPC, non-player character) in the game, and these AI-controlled characters are an important way for developers to interact with users. Among them, pathfinding is the most important part of the game's AI character, which gives the AI character vitality, so that the AI character can interact with the player, chase, and automatically complete the movement to the target point. Pathfinding is the core function of an AI character, and there are many current pathfinding methods, such as Dijstra algorithm, best-first search algorithm, A-star algorithm. Most of these pathfinding algorithms are used in 2D static maps, but there is no detailed description of 3D pathfinding by pathfinding algorithms in UE4. This paper first introduces and compares several map pathfinding models, then experimentally compares different pathfinding algorithms, and then uses blueprint programming to implement AI role pathfinding in the three-dimensional map in UE4 based on the optimal performance algorithm. Finally, the overall experimental results are summarized and sorted out, and the direction for future development is pointed out.","PeriodicalId":146783,"journal":{"name":"Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121174262","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

DGSG: A Efficient Goal Directed Sequence Generator for Pedestrian Trajectory Prediction 一种高效的行人轨迹预测目标定向序列生成器

Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence Pub Date : 2022-12-23 DOI: 10.1145/3579654.3579695

Dingye Yang, Xiaolin Zhai, Zhengxi Hu, Jingtai Liu

引用次数: 0

Contrast-Teacher-Student(CTS): A novel Semi-supervised Learning Approach for Sleep Staging 对比师生(CTS):一种新颖的半监督学习睡眠分期方法

Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence Pub Date : 2022-12-23 DOI: 10.1145/3579654.3579726

Yang Feng, Xian-Lung Tang, Xingjun Wang, Liang Zhao

引用次数: 0

Trust Region Method Using K-FAC in Multi-Agent Reinforcement Learning 基于K-FAC的多智能体强化学习信任域方法

Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence Pub Date : 2022-12-23 DOI: 10.1145/3579654.3579702

Jiali Yu, Fengge Wu, Junsuo Zhao

引用次数: 0

Hybrid Feature Measurement based on Linear and Nonlinear Nonnegative Matrix Factorization 基于线性和非线性非负矩阵分解的混合特征测量

Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence Pub Date : 2022-12-23 DOI: 10.1145/3579654.3579672

Sicong Ye, Yang Zhao, J. Pei

{"title":"Hybrid Feature Measurement based on Linear and Nonlinear Nonnegative Matrix Factorization","authors":"Sicong Ye, Yang Zhao, J. Pei","doi":"10.1145/3579654.3579672","DOIUrl":"https://doi.org/10.1145/3579654.3579672","url":null,"abstract":"The nonnegative matrix factorization algorithm is an effective data dimensionality reduction method. The principle is to convert the image into a nonnegative linear combination of low dimensional basis images. Nonnegative matrix factorization can be divided into linear algorithm and nonlinear algorithm. Because of different decomposition theory, linear NMF algorithms mainly extract first-order features of data, while nonlinear NMF algorithms mainly extract high-order features. Most of the current studies only focus on one of the models without combining the two together, which leads to the lack of data features. Therefore it is necessary to integrate the two types of algorithms for research. The paper proposes hybrid feature measurement based on linear and nonlinear nonnegative matrix factorization. The algorithm utilizes the idea of feature fusion. The basis image features of the two algorithms are mixed in the model. Finally a feature similarity measurement is obtained as the measure method. The proposed algorithm has good performance on the public datasets and effectively improves the recognition.","PeriodicalId":146783,"journal":{"name":"Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133013767","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Domain Adaptation Based on Multi-Kernel Learning 基于多核学习的领域自适应

Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence Pub Date : 2022-12-23 DOI: 10.1145/3579654.3579735

Liyan Han, Weixin Ling

{"title":"Domain Adaptation Based on Multi-Kernel Learning","authors":"Liyan Han, Weixin Ling","doi":"10.1145/3579654.3579735","DOIUrl":"https://doi.org/10.1145/3579654.3579735","url":null,"abstract":"Abstract: Domain adaptation is used to solve the inconsistency in distributions of training samples (source domain) and test samples (target domain) and improve the accuracies of traditional learning machines. Domain adaptation methods attempt to map the two domains to a latent space where the distributions of them are aligned. The model trained in the source domain then can be effectively generalized to the target domain. However, the linear mapping adopted by the existing distribution matching methods has a limited ability to represent the complex transformation between source domain and target domain. In order to overcome this defect, we put forward Domain Adaptation based on Multi-Kernel learning (DAMK) method, which uses a nonlinear mapping. In order to satisfy the different requirements in the nonlinearity of the feature mapping of different datasets, DAMK uses the weighted sum of multiple mappings and optimizes the weighted coefficients. Because of the difficulty in obtaining nonlinear mapping directly, we adopt multi-kernel function instead of explicit expression to express the nonlinear mapping function. Experiments conducted on object recognition datasets and face recognition datasets show that DAMK is more effective that the existing linear mapping methods.","PeriodicalId":146783,"journal":{"name":"Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence","volume":"73 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130438957","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Text abstraction based on user intent and deep analysis 基于用户意图和深度分析的文本抽象

Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence Pub Date : 2022-12-23 DOI: 10.1145/3579654.3579759

Lijuan Liu

引用次数: 0