Trends Identified
Greater access to innovative technologies
57% of KPMG member firm advisors answered that this trend has a large positive impact for the user organizations.
2015
Top trends and predictions for 2015 and beyond
KPMG
Expanding emerging market opportunities for selling goods/services
36% of KPMG member firm advisors answered that this trend has a large positive impact for the user organizations.
2015
Top trends and predictions for 2015 and beyond
KPMG
Improving consumer/customer demand
36% of KPMG member firm advisors answered that this trend has a large positive impact for the user organizations.
2015
Top trends and predictions for 2015 and beyond
KPMG
The ability to tap into skilled global talent pools
34% of KPMG member firm advisors answered that this trend has a large positive impact for the user organizations.
2015
Top trends and predictions for 2015 and beyond
KPMG
Improving/rebounding global economic conditions
34% of KPMG member firm advisors answered that this trend has a large positive impact for the user organizations.
2015
Top trends and predictions for 2015 and beyond
KPMG
Business friendly governments & administrations
23% of KPMG member firm advisors answered that this trend has a large positive impact for the user organizations.
2015
Top trends and predictions for 2015 and beyond
KPMG
Continued trade liberalism; globalization
23% of KPMG member firm advisors answered that this trend has a large positive impact for the user organizations.
2015
Top trends and predictions for 2015 and beyond
KPMG
Improved access to capital at competitive rates
21% of KPMG member firm advisors answered that this trend has a large positive impact for the user organizations.
2015
Top trends and predictions for 2015 and beyond
KPMG
Adversarial Machine Learning Becomes Key for Security & Fraud Prevention
Machine learning has had its advantage in effectively delivering rapid prediction of trends and establishing robust risk management and inference. Much investment, time and focus by organisations has been dedicated to programming and training machine learning algorithms to fulfil these functions. However, if these algorithms become compromised, they will be prone to attacks from threats and viruses. The chaotic damage that permeating cyberattacks have inflicted in algorithms can dangerously result in the misclassification/alteration of information within them; in effect an organisation’s entire system’s security can be at stake. Cybercriminals constantly seek to successfully exploit weaknesses of learning algorithms of highly valued organisations. Fraudsters are responding to the enhanced detection capabilities for transaction fraud and account fraud offered by fraud detection and prevention service providers. In some instances, they are also using machine learning algorithms to uncover weaknesses in fraud detection systems, in a type of machine learning chess match. It is here that the choice of FDP vendor becomes important, in terms of how its machine learning solution is implemented. Is a static model used, or does the vendor employ an adversarial model that adapts to changing conditions? Fraudsters will have very little knowledge of the precise algorithms being used to detect fraud. As a result, time, effort and funding must be sourced to identify weaknesses, which may then be applied or replicated across other merchants assumed to be using similar algorithms. Sectors (healthcare, industrial, advertising) where protection of huge amounts and types of sensitive data (e.g. consumer/public data) is a high priority – will be the drivers here. Spend on Fraud Detection & Prevention software in the financial sector, ie for eCommerce transactions including ticketing, money transfer and payments, will reach $10 billion by 2022. These sectors recognise the value of determining and containing susceptibilities (e.g. to detect unauthorised access points, weak security infrastructures, etc.) in machine learning approaches within adversarial circumstances. They will prioritise increasing their understanding of these vulnerabilities in machine learning algorithms. They’ll also engage with machine learning specialists to design and implement effective action steps to address these vulnerabilities. Juniper Research believes, moving forward into 2019 and beyond, that adversarial machine learning will be required by numerous industries to: · Identify weaknesses of machine learning algorithms during the learning and identification process · Enforce the protection and integrity and validity of data in these systems · Action steps in response to specific threats · Assess the potential damage of these threats · Programme algorithms to enhance their resistance to viruses · Eliminate the presence of opaque ‘black boxes’ · Grant adequate time to algorithm developers to invest into ‘breaking’ the efforts of cybercriminals to infect data Related Research: Online Payment Fraud: Emerging Threats, Segment Analysis & Market Forecasts 2018-2023.
2019
Top Tech trends 2019
Juniper Research
AI to Enhance EDGE Computing Power in IoT Systems
Over the past 12 to 18 months, the concept of EDGE computing to power the Internet of Things (IoT) applications, devices and systems surface has become a significant trend in the digital industry. The following factors have led to the a combination of EDGE Computing Power and IoT: · IT and technology providers, given their limitations, have had to decide how much processing power and computing resources to allocate from the cloud layer to the EDGE layer, particularly for applications that are highly Central Processing Unit (CPU) concentrated. · These providers aim to work alongside many sectors, industries, organisations to attain insightful analysis of specific, filtered data and nformation, via the use of machine learning and EDGE computing within IoT. · The collation of these significant amounts and types of intelligent data via smart sensors, actuators, servers, etc, for further analysis by EDGE and Cloud Analytics engines – instills confidence in decision makers to make the right moves. The aviation industry, for example, realises the power of the EDGE element and IoT together. By measuring and monitoring an aircraft’s performance, significant reductions in fuel and operational expenses and customer churn can be achieved. In 2017, SAS and Cisco announced the birth of their IoT Analytics platform, developed to enable analytics on devices at the edge of the network. Many industries see value in capturing and analysing data on the go, or in motion, rather than analyse data that was stored. Meanwhile, Huawei and Google have made efforts, in the last couple of years, to establish specific products to strengthen and enhance their IoT computing EDGE capabilities. In summary, the combination of AI/machine learning and EDGE within IoT systems allows IT and technology providers to: · Increase the running efficiency of their IoT operations via the gathering of intelligence from local data, particularly in locations where cloud connectivity is inconsistent. · As highlighted earlier, deliver real-time, fast predictions for critical IoT applications through machine learning processes that gather and process this data from devices and sensors. · Increase security of all types of sensor collected data via EDGE, with all privacy, security and compliance risks of these data fully eliminated. Juniper forecasts that the total number of connected IoT sensors and devices will exceed 50 billion by 2022, up from an estimated 21 billion in 2018. This growth, equivalent to 140% over the next 4 years, will be driven by EDGE computing services; increasing both deployment scalability and security. Incorporating powerful AI at the EDGE will enable faster processing, and analysis of IoT applications, and deliver improved data filtering, automation and workload distribution. We expect cloud corporates to potentially lead this space strongly, given their experience and background in AI. Related Research: The Internet of Things: Consumer, Industrial & Public Services 2018-2023
2019
Top Tech trends 2019
Juniper Research