Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence最新文献_第4页

Local Perception based Auxiliary Neural Network for Object Tracking 基于局部感知的辅助神经网络目标跟踪

Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence Pub Date : 2021-03-05 DOI: 10.1145/3461353.3461392

Qing Lei, Dongyun Xu, Yun Gao

{"title":"Local Perception based Auxiliary Neural Network for Object Tracking","authors":"Qing Lei, Dongyun Xu, Yun Gao","doi":"10.1145/3461353.3461392","DOIUrl":"https://doi.org/10.1145/3461353.3461392","url":null,"abstract":"In recent years, tracking algorithms based on deep learning frameworks in the field of object tracking have attracted much attention. The focus of their research is to improve the performance of algorithms by designing deep network models and constructing training data sets. However, this type of algorithm usually uses an end-to-end method to output the predicted target position, and the predicted result always has a certain deviation from the real target position. This paper proposes a object tracking algorithm based on local perception-assisted neural network. The algorithm expands the training dataset. The expanded dataset adds training samples with partial target content and enhances the training label. Train an auxiliary neural network model with local perception capabilities, when the predicted result of the tracking algorithm deviates from the real position, the auxiliary neural network model is used to offset the tracking result to promote the accuracy of the tracking result. Experiments on the VOT2018 and VOT2016 benchmark datasets show that compared with the 6 mainstream algorithms in the object tracking field, the algorithm in this paper has better tracking performance under the three indicators of EAO, accuracy, robustness.","PeriodicalId":114871,"journal":{"name":"Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130543122","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

HMA: An Efficient Training Method for NLP Models HMA: NLP模型的有效训练方法

Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence Pub Date : 2021-03-05 DOI: 10.1145/3461353.3461384

Yuetong Yang, Zhiquan Lai, Lei Cai, Dongsheng Li

引用次数: 0

Reversible Data Hiding in Encrypted Images Based on Bit-plane Rearrangement and Huffman Coding 基于位平面重排和霍夫曼编码的加密图像可逆数据隐藏

Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence Pub Date : 2021-03-05 DOI: 10.1145/3461353.3461372

Mengqi Liu, Tiegang Gao

引用次数: 1

KABI: Class-Incremental Learning via knowledge Amalgamation and Batch Identification KABI:通过知识合并和批识别的类增量学习

Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence Pub Date : 2021-03-05 DOI: 10.1145/3461353.3461367

Caixia Li, Wenhua Xu, Xizhu Si, Ping Song

{"title":"KABI: Class-Incremental Learning via knowledge Amalgamation and Batch Identification","authors":"Caixia Li, Wenhua Xu, Xizhu Si, Ping Song","doi":"10.1145/3461353.3461367","DOIUrl":"https://doi.org/10.1145/3461353.3461367","url":null,"abstract":"In class-incremental learning setting, classes are typically presented batch by batch over time. Incremental learning often suffers from catastrophic forgetting: the performance on previous classes abruptly degrades when adapting a model to new classes. We find that incremental models trained using knowledge distillation are skilled at discriminating classes within a batch, whereas they have confusion among classes in different batches. We propose a class-incremental learning approach with knowledge amalgamation and batch identification (KABI), which can effectively alleviate catastrophic forgetting. The idea is to first train an expert model for new classes at current state, and then train an amalgamation model by amalgamating knowledge from the expert model and the amalgamation model of previous state to discriminate different classes within a batch, and particularly train a batch identifier to discriminate different batches. We conduct extensive experiments on three datasets: MNIST, CIFAR-100, ILSVRC 2012, and show that KABI outperforms the second-best approach by 1.29%, 14.26% and 21.51% respectively. Surprisingly, classification accuracies of our approach are even sometimes higher than the oracle results which is obtained by training a model using all training samples from all classes.","PeriodicalId":114871,"journal":{"name":"Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122501898","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: 3

A Reinforcement Learning-based Path Planning Method for Complex Thin-walled Structures in 3D Printing 基于强化学习的3D打印复杂薄壁结构路径规划方法

Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence Pub Date : 2021-03-05 DOI: 10.1145/3461353.3461382

Jingyi Ge, Yi Wang, Jiayi Li, Huiwen Bai, Linsheng Liu, Shengfa Wang, Xinwei Xue, Fengqi Li

引用次数: 0

On the Innovation of Foreign Language Teaching in the Era of Artificial Intelligence 论人工智能时代外语教学的创新

Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence Pub Date : 2021-03-05 DOI: 10.1145/3461353.3461355

Jinxiang Xue

引用次数: 1

Visible-Thermal Pedestrian Detection via Unsupervised Transfer Learning 基于无监督迁移学习的可见热行人检测

Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence Pub Date : 2021-03-05 DOI: 10.1145/3461353.3461369

Chengjin Lyu, Patrick Heyer-Wollenberg, A. Munir, L. Platisa, C. Micheloni, B. Goossens, Wilfried Philips

{"title":"Visible-Thermal Pedestrian Detection via Unsupervised Transfer Learning","authors":"Chengjin Lyu, Patrick Heyer-Wollenberg, A. Munir, L. Platisa, C. Micheloni, B. Goossens, Wilfried Philips","doi":"10.1145/3461353.3461369","DOIUrl":"https://doi.org/10.1145/3461353.3461369","url":null,"abstract":"Recently, pedestrian detection using visible-thermal pairs plays a key role in around-the-clock applications, such as public surveillance and autonomous driving. However, the performance of a well-trained pedestrian detector may drop significantly when it is applied to a new scenario. Normally, to achieve a good performance on the new scenario, manual annotation of the dataset is necessary, while it is costly and unscalable. In this work, an unsupervised transfer learning framework is proposed for visible-thermal pedestrian detection tasks. Given well-trained detectors from a source dataset, the proposed framework utilizes an iterative process to generate and fuse training labels automatically, with the help of two auxiliary single-modality detectors (visible and thermal). To achieve label fusion, the knowledge of daytime and nighttime is adopted to assign priorities to labels according to their illumination, which improves the quality of generated training labels. After each iteration, the existing detectors are updated using new training labels. Experimental results demonstrate that the proposed method obtains state-of-the-art performance without any manual training labels on the target dataset.","PeriodicalId":114871,"journal":{"name":"Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127257431","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: 2

Soft-Gated Self-Supervision Network for Action Reasoning: Soft-Gated Self-Supervision Network with Attention Mechanism and Joint Multi-Task Training Strategy for Action Reasoning 行动推理软门自监督网络:具有注意机制和联合多任务训练策略的行动推理软门自监督网络

Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence Pub Date : 2021-03-05 DOI: 10.1145/3461353.3461377

Shengli Wang, Lifang Wang, Liang Gu, Shichang He, Huijuan Hao, Meiling Yao

引用次数: 0