Intel is one of the most talked-about topics for the processor giant at a cost of billions of dollars to boost the artificial intelligence (AI) ecosystem Intel acquired through the acquisition and its Intel Capital investment in AI startups, A wide range of AI technology.
Several companies that have been acquired so far appear to be in favor of Intel's ambitions to expand the AI landscape from Altera (2015), Saffron (2015), Nervana (2016), Movidius (2016) and Mobileye (2017) Launched its AI portfolio with startups such as Mighty AI, Data Robot, Lumiata, CognitiveScale, Aeye Inc. and Element AI.
However, it is still not clear how Intel intends to integrate it, and since AI innovation is still in its early stages, it is also reasonable that Intel's AI strategy is obviously decentralized, and we may have to wait for some time to see a more coherent Vein development.
For example, Intel has announced it will ship the Nervana Neural Network Processor (NNP) by the end of this year, the previously known 'Lake Crest' Naveen Rao, former chief executive and co-founder of Nervana, described the NNP as a "deep learning-specific architecture." Naveen Rao is now Intel's vice president and general manager of AI products.
Intel also has other "weapons" in terms of AI chips, including the Xeon line, FPGAs from Altera, Mobileye for cars, and machine learning wafers for Movidius for the edge.
However, Intel has been silent about which aspects of its AI application or what areas of actual focus on AI.Although after all, is a vast and in-depth technical area.While Intel continues to expand the list of acquisitions, the biggest mystery is none other than Saffron.
Just two years after its acquisition of Saffron, Intel released a new product called Intel Saffron Anti-Money Laundering Advisor (AML) in October of this year that has drawn widespread publicity While AML is obviously implemented on the Xeon processor, it is not a hardware but a tool that helps investigators and analysts pinpoint financial crime.
EE Times has recently had the opportunity to visit Elizabeth Shriver-Procell, Saffron Technology's director of financial industry solutions, to gain insight into the AI technology behind Saffron products and the benefits she sees as Saffron becoming an Intel company.
Elizabeth Shriver-Procell, Head of Financial Industry Solutions, Saffron Technology On the other hand, we would like to know more about what a fighter like Shriver-Procell, a long-time financial crime defender, is mainly responsible for within the largest CPU company in the world.
Please talk about yourself. I have heard that you are a financial analyst who has worked in agencies such as the Ministry of Finance and many other companies.
Shriver-Procell: I am a lawyer and I work mainly in the fight against financial crime.I worked in international consulting firms and major financial institutions before leaving Bank of America and joining Saffron earlier this year. Wrong, I also worked as a project manager for analytics development at the U.S. Department of the Treasury.
Did you use Saffron products before joining Saffron?
Shriver-Procell: Saffron has been used by some of the organizations I've worked with (including consultants) and I've always been interested in this platform, so I firmly held it when opportunity came up.
So, what exactly does Saffron offer?
Shriver-Procell: Saffron always sells and advertises for the purpose of providing customized 'analytics platforms' for a wide range of applications, including supply chains, banks and insurance companies.
What has changed with Saffron's platform after the launch of the 'Money Laundering Advisor' tool?
Shriver-Procell: We are now rolling out more specific products for specific applications.
Associative memory AI another way
My guess is that Intel's acquisition of Saffron is mainly due to the acquisition of AI technology, not just financial crime (although it is worth it.) Please talk about Saffron designed AI expertise and applications, it and other AI technology is different?
Shriver-Procell: The AI technique used by Saffron, known as the Associative Memory method, is a differentiated AI-branching technique that comes from Deep Learning. Associative Memory AI is very good at observing large and Diversified information and discernment of signatures or patterns from widely divergent databases can consolidate structured and unstructured data from enterprise systems, e-mail, web and other sources.
Take the example of a bank client, Mary, who travels to London every other week and shop at Liberty Avenue stores. John, who lives in another country, also traveled to London almost at the same time as Mary, but his purpose seemed completely different. So what's the relationship between the two? What's in common? Can we see its IP address? Can we find any similarities in its login mode? Are there any signs that any wrongdoing is in progress?
Therefore, the point is that the associated memory AI can view and reorganize so many seemingly unrelated databases simultaneously?
Shriver-Procell: Not only that, it is also capable of performing extremely time-consuming tasks quickly, but while deep learning requires a lot of training, associative memory AI does not require any training, and the AI branch enables quick, one-time learning, No modeling required.
The press release mentions Saffron's white box AI, which can be explained in more detail?
Shriver-Procell: The so-called White Box AI, and we want to emphasize transparency, lets us explain how to come to some conclusion.Finally, financial institutions bought model-based fraud detection solutions from vendors, It's a 'black box' because users do not know how their software works in a black box, and regulators can not really give an explanation as the financial institutions ask them to conclude, because they can not see the black What's in the box and no way to tell if it's working.
In highly regulated industries, the ability of financial institutions to provide data transparency is crucial.
Sounds interesting and seems to be the exact opposite of deep learning AI Some security experts are concerned that automakers can not explain why AI made this decision when deploying deep learning AI in autonomous vehicles to implement driving decisions such as turning corners The lack of transparency has made it difficult for carmakers to verify the safety of autonomous vehicles.
Shriver-Procell: I think it's most important to recognize that AI has many different avenues. When Intel CEO talked about unlocking the AI's prospects and potential, he suggested that we should try something new and explore new learning paradigms .
Do you think different AI branch technologies will eventually converge at the same point?
Shriver-Procell: I think these branches are complementary, and as we see the convergence of converged applications, I think combining multiple types of AI to meet the needs of a variety of applications.
Please introduce more new products related to Saffron.
Shriver-Procell: As we said before, Saffron always sells products on a platform, and now we decide to start rolling out specific solutions that address different market challenges as we find specific requirements in a given market segment.
Saffron has always taken a very strong position in the financial markets based on its experience in combating financial crime, and through structured 360-degree views of both structured and unstructured data, we understand the patterns found across cross-material storage boundaries.
We also announced that Bank of New Zealand recently joined Intel Saffron's Early Import Program, which is designed to be an institution that is interested in delivering innovative financial services with the latest advancements in Lenovo's Memory AI.
What do you think of Saffron after joining Intel?
Shriver-Procell: The benefits of joining Intel are numerous and we are now discussing the serious issues facing large financial institutions that require the power and resources of giants such as Intel, as well as Intel's technical cooperation Partners with the full support in order to successfully build new features and applications on the Saffro platform and to enable them to be upgraded and expanded.With the rapid progress of AI, explore new things and new methods to achieve AI can not be ignored.
Compile: Susan Hong