The interview concerns Dr. Maria Costanzo, Solution Engineers & Innovation Director of Oracle Italia, who illustrated the potential of Big Data, Oracle’s innovative solutions and some common points that this field presents with the Space Economy. Oracle Corporation is a software company operating in the field of software, data management, cloud solutions and database optimisation.
Dr. Maria Costanzo, can you give us an initial overview of Big Data and the role of Oracle Italy?
I consider extremely interesting the fact that big data can put together a large amount of information, which, although apparently unrelated to the beginning, once united in a single environment reveals repetitive patterns that were not identifiable before. This new approach paves the way for new discoveries, not only scientific ones, but others that cover the most disparate sectors as well: the quantity of data provides solutions which would be impossible to reach by observing the single data belonging to a specific device.
Therefore, the most important value of big data and the acquisition of information within analytical platforms is being able to identify hidden information deriving from the crossing of heterogeneous data.
Another key element is represented by the introduction and use of artificial intelligence technologies: in this regard, Oracle can provide significant support because in our Cloud platforms we have made available technologies that allow extremely advanced processing (GPU) functional to the elaboration and the training of complex artificial intelligence algorithms (i.e. neural networks, deep learning). One of the most interesting aspects in this regard, is given by the ability to process intelligence from this information without limiting itself to their simple observation: when we talk about augmented analytics in fact, we refer to the ability to be able to increase the value of the data, extracting new information deriving from their processing.
In terms of data management solutions, the company is moving towards the forefront of data management. In fact, the solutions promoted by Oracle in this field represent absolute excellence. The extra step Oracle took was to use artificial intelligence to make sure that the use of this excellence was available to everyone. Specifically, if a system self-manages and is also able to protect itself against cyber-attacks, surely the system reduces the level of risk and simplifies the level of management. This autonomy allows technicians to dedicate their activities to more relevant tasks, leaving the database to work alone on its management.
In this perspective, I therefore see Oracle’s Autonomous Database as a simplifier and as an accelerator. The security guaranteed by our Autonomous Database can be fundamental for the protection of data collected by satellites for military use. The solutions in which Oracle is significantly investing are security and AI: the Autonomous DB can generally be defined as a simplifying tool that, through AI, exploits all our know-how over the last forty years on data management to ensure that the security levels are extremely high and the DB is able to “defend itself”.
In this historical period, we have to deal with a growing data volume, and this requires an extremely agile environment in developing solutions suitable for their processing. In this sense, a Cloud environment represents the most effective solution for information management: in this case, the solutions that Oracle makes available to customers are designed to allow horizontal scalability when they are used, so there are no limits to the amount of data that can be processed (even) in real time.
Once elaborated patterns of AI algorithms are identified, another strength is represented by the application of this knowledge in real time, during the progressive data collection. This makes it possible to make the data “actionable” as they are collected, and to have an immediate result with respect to the information that arrives. To achieve this, we need extremely high processing capabilities that Oracle makes available with its own technology.
Dr. Maria Costanzo, what is actionable data?
In the moment in which a datum is acquired and it is elaborated through an algorithm, a “meaning” of the data is obtained: to that point it is necessary to apply such algorithm to the new data that will be acquired. This process is useful for classifying information received in real time, because it eliminates the need to process it later. This method therefore makes it possible to increase the speed of data interpretation precisely because this speed takes advantage of the analytical technologies made available in the initial phase of processing the first data collected.
Dr. Maria Costanzo, what role does Oracle marketing cloud play in all this?
Part of our SaaS offer, it’s the most effective tool for the dissemination of scientific information.
Dr. Maria Costanzo, how does one connect the usefulness of these systems to everyday life? What are the methods to be adopted to make the citizen understand the usefulness of this data?
The data processing process is currently seen as a derivative that is very distant from the benefits that reach the end users. What the end user perceives is a sort of “simplification” of one’s life or work activity. At present there is not yet a full awareness of the fundamental nature of data processing. From a general point of view, I feel I can say that big data helps to have a better level of visibility on what is happening, and therefore allows us to make more efficient decisions. When we begin to examine the “nuances” through detailed and large-scale analyses, the differences and details can be better identified. These last two factors allow us to make more precise decisions. Data analysis can be defined as a huge magnifying glass on something that we normally see in a much lower quality. Thanks to this enlargement, the details are identified and we start to understand how the correlations that are the basis of various phenomena, unfold. In reality they foresee many other causes that up to that moment had not been taken in consideration.
Can we therefore say that the intersection of different data leads to the discovery of a tertium genus of data?
Certainly. In this regard, there was a case where one of our customers had developed algorithms to increase their business. Algorithms were already precise in their own right, but could not increase business: the problem did not lie in the algorithm (this was already well structured) but in the fact that the information on which the forecasts were based was not sufficient. Only when other data was collected from sources, apparently not linked to that specific business phenomenon, did the system began to change its points of reference. In fact, drawing on new data sets, the indicators that were able to influence the phenomenon turned out to be different from those we originally thought would be. We then realised how other elements, initially not considered, were able to influence the phenomenon.
This is the true value of a data driven action: the more data that is put together, even if they are not necessarily correlated with each other, the greater the probability of expanding the ability to view detail and knowledge of a phenomenon through correlations that they escape the human mind but not the technologies. This process is not uniform and limited, on the contrary it is advisable to enrich the data sets more and more (to understand the factors that contribute to the data analysis phenomenon).
So data could be defined as a kind of hidden truth machine?
Yes, data can reveal everything about different phenomena. Personally, I think the open and free data policy is particularly stimulating: the greater the interaction between data belonging to different bodies and structures, the greater is the level of information that results. Proceeding in this direction, we will probably discover aspects that until now have been unimaginable. Nowadays data has become a real value, it is becoming an essential element, to the point of being able to be listed on the stock exchange as a private asset of a company. But it is necessary to go beyond this type of perspective: the data must be something totally available, because it is able to increase information and therefore promote knowledge.
Dr. Maria Costanzo, in this sense, can we hypothesise a social function of the data?
Certainly. But we must always remember that data is so powerful that it can also result in negative uses, and at that point an ethical problem arises. The real problem lies precisely in a possible incorrect use that can be made of the data. For example, I refer to influences on political decisions through the manipulation of the masses of information. There may also be cases in which the structure of the data sets on which the algorithms are based lead them to discriminate: some features of the social context could in fact lead to algorithms for exclusions and / or discrimination (i.e. failure to grant loans to people of a certain race or a certain sex). In conclusion, extreme automation is a crucial factor, but we must then process the information in the correct way to avoid it being amplified, through AI; which is an anomaly that instead is present as a social phenomenon.
Dr. Maria Costanzo, what are the policies that Oracle is adopting with regards to this correct and ethical data management?
The entire information management aspect also includes a data preparation phase, so that it can then be used for analytical or scientific purposes, such as data security management and governance. The use of AI for the Oracle Autonomous Database is based on data concerning the knowledge of the management of a database for which the decisions taken are taken based on the type of use that is made of the Database: this in an automation of automatic processes able to save a Database administrator’s time and work.
The Oracle Autonomous Database hinges on the concept of automated optimisation. What are the features of this concept?
The Autonomous DB makes the machines perform repetitive repair, updating and maintenance tasks, reserving the most valuable activities for humans, especially on the subject of safety. In eighty-five per cent of the attacks, in fact, the patches useful for guaranteeing invulnerability are already available, but the problem is that they have not yet been installed for issues related to the human component (delays, priorities, forgetfulness, etc.). Machine learning instead, in addition to providing a high level of vigilance, is able to guarantee that the software is always up to date; in this sense, thanks to automation, the factor of human error is eliminated.
An analogous situation occurs if the system deems to have an unsuitable functioning: when a request for information is made to the Database, it takes a certain path whose speed varies according to the degree of optimisation. The system is able to understand the types of research carried out and is able to learn from these. Through learning it will suggest the best paths to ensure that the answer arrives as quickly as possible, and all this is done in total autonomy. The system therefore learns through the use made by the user and optimises itself independently to guarantee the best performance based on the use of each specific user. In this way the system increasingly eliminates human intervention.
Dr. Maria Costanzo, what will Oracle’s future challenges for data management be?
The fundamental tasks of the company are those of AI and security. These two factors are crucial because we want to be able to provide our users with safe solutions. Considering that the data has enormous value and is placed in cloud platforms, we must guarantee our customers absolute security. We are working on different fields, including prevention of cyber-attacks.
Moreover, when it comes to Cloud solutions, the problem is no longer confined to the home but becomes global. The storage of information on the Cloud can arouse fear, especially if this information is sensitive. In this case the customer wants absolute guarantees that the information residing in distant places and managed by third parties is protected and any distorted uses of the same are avoided. In this perspective, all companies are beginning asked to make extremely strict security claims, and it is right that this is the case. Oracle is working hard to ensure the highest security standards. It is essential to adopt a conscientious data management, as the ethical component in this field is becoming increasingly important.
Carlo Belbusti holds a Master’s Degree in Law from Roma Tre University. He also attended a Postgraduate course in space law and policies at the Italian Society for the International Organization.