检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Ou, Ruizhe Hu, Yuan Zhang, Fan Chen, Jiaxin Liu, Yu Pattern Recognition and Intelligent System Lab School of Artificial Intelligence Beijing University of Posts and Telecommunications Beijing100876 China Institute of Remote Sensing and Geographic Information Systems School of Earth and Space Sciences Peking University Beijing100871 China Peking University Ordos Research Institute of Energy Ordos017000 China

Multi-modal large language models (MLLMs) have achieved remarkable success in image- and region-level remote sensing (RS) image understanding tasks, such as image captioning, visual question answering, and visual grounding. However, existing RS MLLMs lack the pixel-level dialogue capability, which involves responding to user instructions with segmentation masks for specific instances. In this paper, we propose GeoPix, a RS MLLM that extends image understanding capabilities to the pixel level. This is achieved by equipping the MLLM with a mask predictor, which transforms visual features from the vision encoder into masks conditioned on the LLM’s segmentation token embeddings. To facilitate the segmentation of multi-scale objects in RS imagery, a class-wise learnable memory module is integrated into the mask predictor to capture and store class-wise geo-context at the instance level across the entire dataset. In addition, to address the absence of large-scale datasets for training pixel-level RS MLLMs, we construct the GeoPixInstruct dataset, comprising 65,463 images and 140,412 instances, with each instance annotated with text descriptions, bounding boxes, and masks. Furthermore, we develop a two-stage training strategy to balance the distinct requirements of text generation and masks prediction in multi-modal multi-task optimization. Extensive experiments verify the effectiveness and superiority of GeoPix in pixel-level segmentation tasks, while also maintaining competitive performance in image- and region-level benchmarks. The models, dataset, and code are publicly available at https://***/Norman-Ou/GeoPix. © 2025, CC BY.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

BUPT_PRIS at TREC 2012 Crowdsourcing Track1 21

BUPT_PRIS at TREC 2012 Crowdsourcing Track1

引用

21st Text REtrieval Conference, TREC 2012

作者： Zhang, Chuang Zeng, Minjie Sang, Xiaokang Zhang, Kailai Kang, Houfu Pattern Recognition and Intelligent System Lab Beijing University of Posts and Telecommunications China

In this paper, the strategies and methods used by the team BUPT-WILDCAT in the TREC 2012 Crowdsourcing Track1 will be mainly introduced. The Crowdsourcing solution is designed and carried out on the CrowdFlower Platform. Corwdsourcing tasks are released on the AMT. The relevance labels are gathered from workers of AMT and optimized by the inner algorithms of Crowdflower Platform. © 2012 21st Text REtrieval Conference, TREC 2012. All rights reserved.

关键词： Crowdsourcing

来源：评论

学校读者我要写书评

暂无评论

Sentiment information to vector, a more automatic approach for sentiment analysis

Sentiment information to vector, a more automatic approach f...

引用

作者： Du, Lin Xu, Wei Ran Liu, Ping Yang Pattern Recognition and Intelligent System Lab Beijing University of Posts and Telecommunications China

ISBN: (纸本)9783038351221

As sentence level sentiment analysis having been studied extensively, it has been proven that the syntactic structure of a sentence usually holds important information for sentiment analysis, especially for handling polarity reversal. However, the previous attempts of adopting such structural information mainly focus on making certain predefined rules which requires large linguistic expertise of the rule-maker,and the procedure itself is often manually labored and time consuming. To solve this problem, in this paper we propose a novel simple vector model to represent a sentence's syntactic structure and its prior sentiment information uniformly and rapidly. Experiment results show that our proposed approach performs well in COAE 2013 dataset, and could also be used for machine learning algorithms to extract more distinguish features automatically. © (2014) Trans Tech Publications, Switzerland.

关键词： Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

BUPT_PRIS at TREC 2012 Session Track 21

BUPT_PRIS at TREC 2012 Session Track

引用

21st Text REtrieval Conference, TREC 2012

作者： Zhang, Chuang Wang, Xiaotian Wen, Songlin Li, Runze Pattern Recognition and Intelligent System Lab Beijing University of Posts and Telecommunications China

In this paper, we introduce our experiments carried out at TREC 2012 session track. Based on the work of our group in TREC 2011 session track, we propose several methods to improve the retrieval performance by considering the user behavior information over the session, which includes use query expansion based on meta data, query expansion based on click order, optimization based on history ranked lists and so on. The results show that some methods can really improve the search performance and some methods need to be optimized. © 2012 21st Text REtrieval Conference, TREC 2012. All rights reserved.

关键词： Behavioral research

来源：评论

学校读者我要写书评

暂无评论

BUPT-WILDCAT at TREC 2011 session track

BUPT-WILDCAT at TREC 2011 session track

引用

20th Text REtrieval Conference, TREC 2011

作者： Liu, Tang Zhang, Chuang Gao, Yasi Xiao, Wenjun Huang, Hao Pattern Recognition and Intelligent System Lab Beijing University of Posts and Telecommunications China

This paper is an overview of the runs carried out at TREC 2011 Session track, which proposes several approaches to improve the retrieval performance over one session including the search model based on user behavior, VSM meta similarity model, optimization based on history ranked lists, optimization based on user's attention time and anchor log. The evaluation results show that our implementations are effective.

关键词： Behavioral research

来源：评论

学校读者我要写书评

暂无评论

BUPT-WILDCAT at TREC crowdsourcing track: Crowdsourcing for relevance evaluation

BUPT-WILDCAT at TREC crowdsourcing track: Crowdsourcing for ...

引用

20th Text REtrieval Conference, TREC 2011

作者： Xia, Tao Zhang, Chuang Li, Tai Xie, Jingjing Pattern Recognition and Intelligent System Lab Beijing University of Posts and Telecommunications China

In recent years, crowdsourcing has become an effective method in many fields, such as relecance evaluation. Based on our experiment carried out in Beijing University of Posts and Telecommunications for the TREC 2011 Crowdsourcing track, in this paper we introduce our strategies in recruiting workers, obtaining their relevance and rank juegements and quality control. Then we explain the improved EM algorithm and Gaussian model that we make use of to calculate the consensus of labels. The result shows that our stategies and algorithms are effective.

关键词： Crowdsourcing

来源：评论

学校读者我要写书评

暂无评论

ACTIVATION FORCE-BASED AIR POLLUTION TRACING 5

ACTIVATION FORCE-BASED AIR POLLUTION TRACING

引用

2016 5th IEEE International Conference on Network Infrastructure and Digital Content(IC-NIDC 2016)

作者： Huanyu Zhou Ni Zhang Di Huang Zhanyu Ma Weisong Hu Jun Guo Pattern Recognition and Intelligent System Lab. Beijing University of Posts and Telecommunications NEC Labs China

ISBN: (纸本)9781509012473

Atmospheric pollution is getting more and more attention because of the serious *** the pollutant source is one of the key steps to solve this *** paper introduces a new method to trace the pollution diffusion path according to the relevance of different areas where they were polluted(the relatively small areas regard as points or lattices in this paper).We present the single point tracing to show the specific traceability and the area tracing to show the pollution spread trend.

关键词： Air pollution Activation Force Tracing

来源：评论

学校读者我要写书评

暂无评论

The study of human behavior dynamics based on blogosphere

The study of human behavior dynamics based on blogosphere

引用

2010 International Conference on Web Information systems and Mining, WISM 2010

作者： Song, Yading Zhang, Chuang Wu, Ming Pattern Recognition and Intelligent System Lab Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9780769542249

Blog and microblog have become one of the most popular applications which individuals could be the message source. Therefore, interactivities between individuals have been largely enhanced in today's world. In terms of message delivery and human behavior, understanding the mechanism behind is of significance. Though human dynamics has become a popular subject, most of relevant studies so far used the Poisson process to approximately describe the human behavior. However, extensive evidences like email exchange, online game, and mobile communication showed a non-Poisson statistics with heavy tail in those applications. Meanwhile the study in human dynamics is still rare to some extent. Thus in this paper, the authors provided empirical evidences to support such a non-Poisson statistics being existed in the web2.0 applications blog and microblog, also showing the difference between these two applications. The fact that both human behavior in blog and microblog follow a heavy tailed distribution with α≈1.3 and α≈2 will be revealed. At last, the study of human behavior related to social network can also be enriched. © 2010 IEEE.

关键词： Blogs

来源：评论

学校读者我要写书评

暂无评论

Modeling microblogging communication based on human dynamics

Modeling microblogging communication based on human dynamics

引用

2011 8th International Conference on Fuzzy systems and Knowledge Discovery, FSKD 2011, Jointly with the 2011 7th International Conference on Natural Computation, ICNC'11

作者： Xie, Jianjun Zhang, Chuang Wu, Ming Pattern Recognition and Intelligent System Lab Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9781612841816

Microblog is a large-scale information sharing platform where intensive communications are taking place through interactive user behaviors. Previous studies have analyzed and modeled a series of traditional communications, such as letter, email and phone calls. The timing of human activities tends to be non-Poisson with bursts and heavy tails. Hence, several models have been proposed to explain human dynamics of bursts and heavy tails in various fields. However, as a newly developed product of Web2.0, microblog possesses inherent new characteristics, e.g. user-centric broadcast medium and asymmetric user relations. The communication in microblog is still poorly understood. Our work proposed an interest-driven model to simulate basic user communicating behaviors and processes. We came to the conclusion that, in microblogging communication, individual behaviors are bursts of rapidly occurring events separated by long periods of inactivity. Collective behaviors follow heavy-tailed Power Law distribution, whose origin is the descent of interests. Empirical statistics were also given to verify the model. © 2011 IEEE.

关键词： Behavioral research

来源：评论

学校读者我要写书评

暂无评论

An efficient algorithm of hot events detection in text streams

An efficient algorithm of hot events detection in text strea...

引用

International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery

作者： Bai, Junliang Guo, Jun Chen, Guang Xu, Weiran Du, Gang Pattern Recognition and Intelligent System Lab. Beijing University of Posts and Telecommunications China

ISBN: (纸本)9780769542355

Hot events detection in text streams has drawn increasing attention in recent sequential data mining works. Different from traditional TDT task which find all the real events' cluster, hot events detection only identify hot events concerned by public. This paper proposes a novel approach to identify those events based on burst terms, terms co-occurrence and generative probabilistic model. Experiments with huge text stream sets crawled from WWW suggest that our algorithm can work on-line and identify hot events effectively and efficiently. © 2010 IEEE.

关键词： Data mining

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：