"北约决策：大数据与人工智能时代"

人工智能

需积分: 5 87 浏览量更新于2024-03-23 收藏 2.81MB PDF 举报

身份认证购VIP最低享 7 折!

领优惠券(最高得80元）

资源详情

资源推荐

NATO DECISION-MAKING IN THE AGE OF

BIG DATA AND ARITFICAL INTELLIGENCE

OPPORTUNITIES AND CHALLENGES IN THE MANAGEMENT AND

GOVERNANCE OF BIG DATA

George Christou - University of Warwick

Abstract

The digitalization of ‘everything’ and subsequent proliferation of data that is available to organizations across

different sectors presents major opportunities, but also brings along greater complexity and challenges in

relation to the effective governance of Big Data Analytics (BDA). This chapter seeks to explore how

organizations have strategically adapted and innovated to manage their BDA approach. Moreover, it assesses

how they have navigated the challenges of Big Data analysis in order to harvest its potential power in gaining

insight and adding value and advantage, whether in the market place or in relation to the battlefield. It posits

that whilst responses and approaches are still very much formative, governance lessons can be gleaned from

available experiences in different sectors. This, in turn, can inform the evolution of good practice in the

security sector domain as a complement to internally developed responses that are also emerging in

organizations such as NATO to ensure BDA provides commanders, intelligence analysts and strategic

communicators with clear insight and data on which critical decisions can be taken.

Introduction

The evolution and availability of Big Data is dramatically redefining the opportunity landscapes across and

within many sectors – whether business, healthcare, pensions, energy or indeed, government, security and

defense. Big Data, in this sense, is understood as “a collection of large data sets that contain massive and

complex data” and “data that exceeds the processing capacity of conventional database systems” (Al-Badi,

Tarhini and Khan, 2018). In the new governance ecosystems that are emerging to manage Big Data,

commercial industry has been at the forefront of innovating to harvest and exploit the advantages and

benefits that Big Data can bring; not surprising, given that the Big Data Analytics (BDA) market is projected

to be worth USD 103 billion by 2023. Commercial industry then has led the way in in providing examples of

the organizational pathways, processes and strategies that are needed to ensure effective and efficient use

of the data revolution. As is noted in one report “Companies with good data analytics capabilities…are twice

as likely to be in the top quartile of performance within their industries” (PwC, 2019: 3). This, however, does

come with one caveat, as other leading industry reports indicate that ‘whilst it may come as no surprise that

data and analytics are reshaping industry competition…the persistently lacklustre response to this

phenomenon by most companies should raise some eyebrows” (McKinsey and Company, 2019: 2).

What the evolving literature points to on BDA is that no strategy will be the same, and that whether an

organization is in the business or military sector, strategic adaptability and flexibility are important factors in

shaping an approach that is unique to the requirements of its specific goals and objectives (for an overview

of key debates and issues see Morabito, 2015; Espinosa and Armour, 2016; Günther et al., 2017; Sivarajah et

al., 2017; Ladley, 2020). It is also evident that there are contrasting models being developed for the

NATO DECISION-MAKING IN THE AGE OF

BIG DATA AND ARITFICAL INTELLIGENCE

governance of BDA, which can be characterized broadly, within a centralized to decentralized spectrum. If

commercial industry gets its data analytics management right, there are gains in terms of getting ahead of

the competition and increased revenue; for security sector organizations, it can mean gaining a tactical

advantage in identifying and responding to threats, predicting the behavior of adversaries, optimising

logistics and protecting military networks from attack.

The purpose of this chapter is to show how BDA practices developed in the private and public sectors can

provide valuable lessons for those in the military and security sector. It is argued that governance lessons can

be gleaned from available experiences in different sectors and that this can inform the evolution of good

practice in the military and security sector domain; and this, of course, as a complement to internally

developed responses that are also emerging in organizations such as NATO to ensure BDA provides

commanders, intelligence analysts and strategic communicators with clear insight and data on which critical

decisions can be taken.

The chapter is structured in the falling way: the first section will discuss how organizations across different

sectors, in particular the commercial sector, have innovated to manage their BDA approach. The second

section will provide an overview of the challenges and pitfalls related to BDA. The third section will then

assess what sort of governance lessons can be adapted from the available experiences in other sectors that

can inform the evolution of good practice in the military and security sector domain. The final section will

provide concluding thoughts on challenges and opportunities going forward in relation to BDA in the military

and security sector.

Organizational Innovations and Revolutions in Big Data Analytics

To fully comprehend evolving strategies for governing Big Data, it is first necessary understand what it is –

what are its main features and what are the common challenges. Big Data is associated with massive data

sets “having larger, more varied, and complex structures that are difficult to store, analyse, and visualise

using traditional data processing technologies” (Yang et al., 2019: 2). To this end, when discussing Big Data,

it is not just the large volume of data that is important (see Table 1; Yan, 2018: 112). Big Data, according to

Mauro et al. (2016), “represents information assets characterised by…high volume, velocity, variety, and

veracity” with others having added further important pillars and characteristics such as the virtual, variability

and value (Priyadarshy, 2019; see also Van Puyvelde et al., 2017; Moorthy et al., 2015). What most general

definitions point to is the need for specific, innovative technologies and governance methods that will allow

the transformation of Big Data in to value.

NATO DECISION-MAKING IN THE AGE OF

BIG DATA AND ARITFICAL INTELLIGENCE

Table 1: Pillars of Big Data

Pillars of Big Data

Definition

Volume

The huge quantity of data that is generated and

needs to be stored and where new data storage

technologies are required accommodate the Big

Data.

Velocity

The speed at which the data needs to be generated

and processed in order to meet demands and

objectives. Particularly important is the ability to

process and action data in real-time.

Variety

It refers to the fact that data can come from a

number of formats and sources, with three typical

types: structured (sensor data, machine data,

financial models, etc.), semi-structured (XML

documents, emails, etc.) and unstructured (text,

videos, audio files, images, etc.).

Veracity

It refers to the truth in and trustworthiness of data.

That is, the extent to which any data can be used to

make critical decisions by any organization.

Virtual

It addresses the data duplication and lost in

transformation related problems whilst enabling

better data governance.

Variability and Value

They are inter-related with variability cutting across

the other pillars. Variability, in this sense, refers to

the different value of data within different phases

of the Big Data cycle.

What then, are the main ways in which organizations across different sectors have sought to create

governance frameworks that are fit for purpose when it comes to the management of BDA? It can be argued

that the public sector and private sector face common but also very different questions and challenges when

it comes to Big Data governance. Common, in the sense that data security, data integrity, data quality, data

accessibility, data availability and data consistency are generally important within any governance framework

for Big Data. Different, because commercial motives for pursuing a Big Data strategy are primarily driven by

competition, market advantage, revenue generation (and cost reduction), whereas public sector strategies

must consider broader questions beyond the benefits of commercialization, in particular in relation to public

interest and societal benefit (Kim & Cho, 2017: 384). Whilst Big Data frameworks are being developed by

various governments for a variety of purposes, from crime prevention and healthcare to counterterrorism

NATO DECISION-MAKING IN THE AGE OF

BIG DATA AND ARITFICAL INTELLIGENCE

and public administration,

commercial industry is most advanced in how to exploit data availability and

management. It is thus possible to argue that there is much to learn from the commercial sector in terms of

organizational innovation and indeed revolution when it comes to developing and implementing Big Data

management and governance strategies. How then, have businesses transformed themselves in order to be

able to effectively exploit the opportunities of the Big Data era?

There is a general consensus that any organization that seeks to exploit Big Data should have a clear strategy

and goals that helps them to understand, define and implement their Big Data objectives (McKinsey and

Company, 2019: 3). Similarly, others have argued that “for benefits to be sustainable over time and provide

strategic advantage to an organization, the BDA effort will require coordinated…and effective governance”

(Espinosa and Armour, 2016: 1117). To this end, any major organizational innovation in terms of Big Data

insight must be linked to a holistic consideration of the component parts of any business organization to

create an agile architecture and strong governance that will allow the full exploitation of Big Data in the

context of ever changing regulatory, technology, business and customer environments – as well as data

trends. An important starting point and innovation in this regard then, is to develop a data insights strategy

that is collaborative, to ensure buy-in from key stakeholders across the organization more easily. It is argued

that by moving away from siloed data-management approaches, stakeholders across an organization become

more accountable for how data is used, thereby reducing the risks associated with potential data misuse

(Deloitte, 2018: 1).

Beyond this general consensus there are a series of other inter-related issues on creating an effective Big

Data strategy that can be found in the existing literature. That is, innovations that move organizations beyond

traditional data management methods to governing Big Data through new organizational structures, new

skills and new roles.

First, in terms of organizational structure for data ownership, findings show that centralized, decentralized

and hybrid forms can work effectively as long as certain conditions are met; the most important being that

of establishing a governance framework that allows the various units to work in an integrated rather than

siloed manner. Here it is argued that whilst a siloed approach may have worked in the past to improve, for

example, information security, such an approach becomes more complex, costly and less effective, in a Big

Data environment where external and internal data sources grow, and where for such data to have value to

the organization as a whole, it must be of high quality and be in presented in a single, consistent format. Big

Data in this sense, must be normalized appropriately to be read by Machine Learning (ML) or Artificial

Intelligence (AI) algorithms (Deloitte, 2018: 5). What should also be borne in mind when considering the type

of structure for any data ownership is that any organization should have the ability to adapt according to its

needs (or level of maturity in BDA) and the changing data environment; to this end, companies may start

with a decentralized model but may transition to a centralized model over time or centralize in the first

instance and transition to a hybrid model. Additionally, it is argued that whilst data ownership should not be

According to the World Bank Group the “United Kingdom, the United States, Singapore and Korea are among many

that have adopted high-level Big Data strategies” – with France, New Zealand, Australia and Japan among others that

have developed Big Data strategies (2017: 1). Some examples of successful ways in which BDA have been used include

public security and public utilities. In terms of public security, Predpol, a predictive policing application has been used

in 50 US cities and which analyzes basic data, such as crime type, dates and location, to help law enforcement agencies

make better policing decisions; with Brazil, the UK and the Netherlands using similar approaches. Similarly, in terms of

public utilities, Shanghai’s municipal government has installed sensors throughout the city’s 3.300-kilometer water pipe

network. These sensors allow the authorities to identify precise locations of issues such as leaks, dramatically reducing

response times, infrastructure damage, maintenance costs and service disruption (World Bank Group, 2017: 4-6).

NATO DECISION-MAKING IN THE AGE OF

BIG DATA AND ARITFICAL INTELLIGENCE

centralized, data governance should, with top performing companies “often having data centralized within

business units” (McKinsey and Company, 2019: 3). Indeed, the evidence suggests that in locating a BDA unit,

organizations should ensure that it is cross-functional, integrated and accessible to ensure it can break

through inertia and enable transformation (McKinsey and Company, 2019: 3-4). Whether in the public or

private sector, locating a BDA unit with an information or business intelligence unit (e.g. a Centre of

Excellence - CoE) is a more productive organization-wide strategy than trying, for instance, to locate it in

specialized units such as IT, where staff might not possess the requisite ability to manage agile BDA functions.

Another important organizational innovation relates not just to the overall nature and location of BDA in an

organization but leadership and data culture (PwC, 2019: 13). There are several aspects to this, including new

roles, skills and education. For example, within the financial services sector there is evidence to suggest that

in order to maximize the strategic use of their data – including Big Data – Chief Data Officers are required to

provide “strategic guidance and execution support, and also to assure access to and the quality of critical

data”. It is argued that senior data-centric leadership positions are crucial to both manage and govern data,

but importantly in a Big Data environment, leverage the data using emerging technologies “that can generate

actionable analytical insights and tangible business benefits” (Deloitte, 2016: 2). Evidence also suggests, more

broadly, that another key ingredient for a successful BDA strategy is creating a data culture through

organizational leadership. This incorporates “a set of practices that brings together data talent, tools, and

decision-making so that data becomes the default support for company operations” (McKinsey and

Company, 2019: 5).

Key differentiators in high-performing companies in this regard are data leadership, the accessibility of data

tools to frontline employees, an organizational culture that supports rapid iteration and tolerates failure,

education (ensuring that employees trust Big Data, understand its value, its strengths and limitations), the

ability to attract and retain talent, and an underlying technology that can adequately support Big Data

exploitation and analytics. To this end, and in terms of talent, it seems that companies are in most need of

‘translators’ – or those that can bridge the gap between any BDA unit and other parts of the business. Indeed,

even though many data-management processes are now automated (from sourcing and cleaning data, to

ensuring data quality, reporting and visualization), management of such data is still reliant on human activity

– and therefore talent – not just in terms of translators that make sure sophisticated models are translated

into simple, visual tools for supporting front-line staff, but also data scientists, engineers and analysts that

support them. The companies that have managed to retain talented staff successfully are those that have

constructed clear opportunities and career paths that can ensure transitioning to other roles and learning

from those within the data analytics team with different roles. To fill any remaining gaps, such strategies

have often been accompanied by strategic partnerships with other organizations to access skills and talent

required (McKinsey and Company, 2019: 5-7; McKinsey and Company, 2018: 4-6).

Finally, in terms of success in a data-centric culture, technology infrastructure has been seen as critical for

achieving organizational objectives in data analytics; indeed, it is ranked second only behind BDA strategy as

the most difficult challenge for an organization. Here, the evidence suggests that traditional data

architectures that are characterized by technology that is inflexible and complex (and often more costly) with

data stored across a variety of dispersed technologies, are not fit for purpose when it comes to BDA. Surveys

show that more agile and robust data architecture “allows you to add more applications or new technological

enhancements as needed, and makes it easier to cohesively manage data” stored in a variety of locations

(Deloitte, 2018: 5). It also allows companies to support the rapid collection and sharing of data, access and

剩余99页未读，继续阅读

承让@

粉丝: 8
资源: 380

"北约决策：大数据与人工智能时代"

alb是哪个国家的拼音缩写

军事领域里面向系统架构设计的重要框架有哪些，介绍一下

什么是link11信号

动态RCS国内外研究现状

物流物流物流物流物流物流物流物流

DC-AC PWM逆变器simulink.rar

电子设计项目毕业设计论文综述单片机控制系统的抗干扰设计资料

采用V_F控制方法的感应电机闭环速度控制simulink.rar

RT-Thread Nano版本在STM32F103RB上的快速移植

linux操作系统基础命令.zip

springboot基于Hadoop的NBA球员大数据分析与可视化(1)(6).docx

RTL代码顶层集成工具

知识管理-PDF解析技术-文本与表格分离-构建本地知识库-科研和学习使用

Tomcat的安装及配置教程主要分为以下几个步骤： 一、准备工作 确认Java环境： 确保已经安装了Java Developm

基于ssm的校园活动管理平台设计与实现.docx

混合交直流和直流微电网测试系统仿真simulink.rar

gcc-9.1.0.tar.gz

MATLAB 和 Simulink 设计电池管理系统.rar

Simulink 模块集用于在 GTKWave 上显示来自 simulink 的仿真数据.rar

gcc-14.1.0.tar.gz

最新资源

Tomcat的安装及配置教程主要分为以下几个步骤：一、准备工作确认Java环境：确保已经安装了Java Developm