Listed below are some of the most frequently asked questions, of FAQS...
Unlike virtually all other data analysis solutions, the Uncertainty Engine does not rely on identification of breaches of pre-defined rules. Instead, it looks for patterns in datasets and identifies unusual, undefined and in some cases unauthorized trends. At this point, SI, in consultation with your staff, will determine whether a particular pattern is ‘good’ or ‘bad’ and instruct the Engine accordingly. Also, no two scenarios are the identical; rules-based solutions tend to get cluttered up with rules for eventualities that may never be seen again, leading to inefficiencies and waste of processing power. The Uncertainty Engine thrives on bigger volumes of data-its architecture is such that the more comprehensive the input, the greater the accuracy of output. It retains 100% of pattern knowledge which it applies to future data input, thus effectively ‘learning’ from past experience.
It's true, everyone 'talks their own book' as it were, and yes SI is small compared with some in the sector. However, many big players started off small in the technology sector, and we are very proud of the fact that in a recent trial, given by a major client to ourselves and a multi-national competitor for comparison purposes, SI's results proved to be of far more value to the client than the competitor's-this can be referenced directly with the client and can be arranged via email to Reference@sinusiridum.com
No. Although the engine works more accurately as more data is input, we feel that many so-called big data solutions focus too much on data integration, relying on algorithms, fuzzy logic, neural net and such. One of the real differentiators of the engine is that the software is geared to producing the highest quality of analysis rather than pointing up simple breaches of any number of pre-assigned rules.
The engine has been deployed across a number of sectors (pharmaceutical, chemical, retail, transport, public), but specifically within financial services has successfully identified fraudulent activity in a foreign exchange trading book for one client, for another, has identified poor-performing loans within a portfolio, and for a third, identified herding characteristics in high-volume trading.
In the fraud instance, as a result of SI screening, a tier 1 European Bank discovered trading book anomalies with potential losses numbered in millions of US Dollars, and several traders were either dismissed or severely disciplined. With regard to the loan portfolio, a tier 1 UK bank was able to make changes to its RWA calculations (currently a major concern for global regulators), release capital back into the business, and identify opportunity and future trends.
Not so. The Uncertainty Engine runs on its own hardware and software. All it requires is a data interface with your internal systems.
We understand that client confidentiality is of paramount importance. Although SI has its own secure processing centre, we are equally happy to locate the Engine within your own premises. Sinus Iridum’s secure hosting hardware is ISO27001 compliant, and PCC DSS certified-in other words of military grade.
No. As mentioned above, the Uncertainty Engine specifically searches for patterns in multiple datasets simultaneously, rather than seeing if any number of pre-defined rules have been broken. In addition, the Engine is structured into ‘Drive Clusters’, and we generally find that one Drive Cluster is sufficient-if more processing power is required, extra clusters can be added.
At the outset, we would determine our up-front costs dependent upon scale of the project, and these would be set out formally. These costs of course need to be judged against potential benefits. In recent years, we have seen enormous reputational damage done to several institutions, hard costs in terms of trading losses, and fines for unlawful activities which were simply not picked up by existing compliance and risk controls. Efficient use of capital and RoRWA are going to be major components of the incoming Basel 111 requirements, and will require significant resources to monitor.
It is fundamental to understand that the Uncertainty Engine is unlike anything else-we do not need you to give us a list of assumptions or pre-determined rules. Neither will we ask, as many others do, "What do you want us to find?". The engine will simply look for unusual or interesting patterns.The SI team will do an initial investigation, so a point of contact is all we require at the outset. After that, we embark on a proof of concept which usually takes three months to produce meaningful results, although patterns may emerge at any time. SI will maintain regular contact with you, and at the end of the trial, clients are invited to investigate the results and compare with actual outcomes or internal predictions.
The essence of the Uncertainty Engine is that it distils data into knowledge, and therefore into commercial advantage and opportunity. We recognize that our customers need a simple interpretation of all this realtime data processing, and we would always provide you with a bespoke dashboard, configured to your own requirements and with the capability to track rolling updates as your business evolves.
Language is not a problem as the engine absorbs data in binary.
Of course. Just pick up the phone or drop us an e-mail.