Intelligence is increasingly used in our daily systems, said Eric Tsui, professor and associate director at The Hong Kong Polytechnic University’s Knowledge Management and Innovation Research Centre. “More and more machine intelligence is embedded into the things that we interact with on a daily basis.”
Tsui argues that despite businesses starting to explore uses of this growing resource through big data and analytics, the emphasis should be shifted to “approximations” rather than “exactness” when integrating findings with business strategy.
Increasingly, he explained, “there are more and more phenomena that cannot be explained in the big data era, but people have to make rapid and correct decisions.” Tsui advised that businesses should realise that the speed of decisions can often be more important than finding out reasons.
For Tsui it is critical that newer smarter algorithms are created to support this faster decision-making amidst such large oceans of data – “With big data, exactness is difficult to define, so new algorithms need to be created.”
Referring to his recent paper, titled Cloud Computing and MOOCs for Supporting Knowledge Work, Innovation and Learning, Tsui highlighted that existing algorithms and enterprise applications are designed for handling smaller data, and current ETL (Extract, Transform, Load) tools are not sufficiently fast enough to deal with the velocity of big data.
He added that big data will place increasing demand on storage technologies to “datify” the data. “Each piece of data needs to be defined, tagged, and indexed for analysis. This places an extremely high demand on storage capacities and CPUs to meet “datafication” requirements,” said Tsui. “Some of the smart big data technologies gradually need to be embedded in storage technologies, most likely as hardware and firmware,” he suggested.
Speaking on the advancement of the Internet of Things (IoT) – placing sensors on physical products – Tsui hopes that it will offer additional dimensions of data for analysis to support verification and predictions. “Data collected can be related to a vast range of factors including temporality, location, status, intensity, and sequential activities. There is no doubt that this enrichment of data, if utilised correctly, will lead to higher accuracy and open up more applications for big data and analytical tools,” he said.
However, Tsui continued to stress that it will still take years for businesses and universities to groom the required amount of quality data scientists and to develop the advanced intelligent algorithms needed to realise the full potential of IoT and big data.
In the meantime, he suggests that retail, consumer goods and digital entertainment will find the biggest opportunities to optimise strategy based on big data and analytics: “There has already been a significant amount of work in these sectors around improving business intelligence. Data in these fields is accumulated in real time and in huge volumes; it is readily accessible and can be analysed.”
These areas are also revenue generation industries, he added, and are generally more willing to spend big and invest in attracting the right people and acquire the necessary tools to boost their revenue streams.