随着人工智能和大数据等新兴专业的日益普及,相关专业的人才培养对高校提出了更高的要求,特别是工科专业,实验教学是人才培养环节不可或缺的一部分.由于新兴专业建设尚在起步阶段,实验教学方面流程还不够成熟,实验平台建设缓慢,无法满足新技术的更迭要求.针对以上问题,提出了一种基于Docker的人工智能实验平台建设方法,使用Docker容器技术制作人工智能实验课程镜像,镜像内封装实验内容和实验环境,用户根据课程需要拉取镜像并利用容器启动实验环境.结果表明该方法可以有效支撑人工智能实验教学工作,同时该方法简化了实验平台搭建步骤,优化了实验教学流程,有助于提升人工智能和大数据专业学生的实践能力,提高高校人才培养质量.

开展水利大数据分区管理模型应用研究,构建水利大数据分区管理模型,开发水利大数据分区管理系统,实现水利大数据的分区管理与深度挖掘应用.在水利大数据环境下,研究了如何将分散在不同系统、不同部门的水利系统数据进行高效整合,做到数据互联.基于大数据互联及存储平台,对如何进行全员数据统一分析、展示,实现跨业务系统的协同应用,实现水利信息化集成数据与互联网数据的互通与结合,进行了关键技术分析.

针对海量图片存储,已有若干个基于Hadoop的方案被设计出来。这些方案在系统层小文件合并、全局名字空间以及通用性方面存在不足。本文基于HBase提出了一种海量图片存储技术,成功解决了上述问题。本文将介绍基于HBase海量图片存储技术方案,分析其原理及优势,该方案在城市交通监控中得到应用验证。

Along with bioinfomatics softwares appear one after another,how to supply convenient computing services to bioinfomatics researcher is getting more and more important.In the background of bioinfomatics computing,a architecture layered as Portal,application services,key services,resource is presented in this paper.According to this 4-layered architecture,a new bioinfomatics service platform which is based on Web services is constructed in JAVA language and other tools such as ECLIPSE.And the author introduce the realization mechanism in detail.

In order to helping biological researchers to solve the problem hard to find,to learn and to in tool using completely,a new versatile web station,named biological study service platform(BSP),has been developed,which allows users to learn and use biological tools efficiently online.It mainly has three components,that is,software pool,software forum and web workstation.Software warehouse was designed to collect tools for users and also provide tool tryout;tool forum was designed for tool study and using experience intercourse and web workstation was founded for tool use and data manage.Also,it can be used as the teaching and learning platform for teachers and students in life science.It can be accessed from the follow linkage: 218.57.145.30:9999/biosp/.Besides,in order to assist researching work well,more new tools will be added,and more software will be developed continuously in this web station.

At first this paper descirbes the basic operation principle of Vector Space Model in information filtering.In view of the shortages of current information filtering technology as low precision and low speed,this paper studied the acquirement of BP ANN input data,the choice of network topology,the training process and the processing method of BP ANN output data by using the advantage of BP ANN such as stable performance and strong learning capability.And in the end,this paper proposes a new information filtering model based on BP ANN.The experiment results proved that this model filter is remarkable in efficiency.