Daily Tech Digest - May 17, 2017

Artificial Intelligence in Financial Services: Opportunities and Challenges

Considering the future, while we realize that AI can prove to be invaluable to the growth of the fin-tech sector, the machines won’t take over just yet. While they may replace humans in some areas of functioning and play the role of personal assistants and digital laborers, there are challenges like bias, privacy, trust, trained staff and regulatory concerns that continue to be a hurdle to be dealt with. Augmented Intelligence, in which machines assist humans in their system functions, could be the more plausible answer. Another key area where AI will continue to play a pivotal role will be in big data. Sifting through and analyzing thousands of pages of data is a burden and a waste of human resource and more and more of these machines will be used to perform advanced analytics of patterns and trends.

Are CEOs Less Ethical Than in the Past?

Confidence and trust in large corporations and CEOs have been declining for decades. But the decline has accelerated since the financial crisis of 2007–08, the Great Recession, and the slow recovery that ensued. Corporations and executives received government bailouts, while seeming to suffer little in the aftermath. Although many companies paid large fines and settlements, few were charged criminally, even in instances where unethical and illegal activity was widespread and well documented. Media attention has also focused more and more on corporate tax avoidance and the offshoring of jobs, as well as record-high rates of executive compensation and rising income inequality in general. Those are the areas that, although not illegal, do not promote goodwill.

Automated Machine Learning — A Paradigm Shift That Accelerates Data Scientist Productivity

There is a growing community around creating tools that automate the tasks outlined above, as well as other tasks that are part of the machine learning workflow. The paradigm that encapsulates this idea is often referred to as automated machine learning, which I will abbreviate as “AML” for the rest of this post. There is no universally agreed upon scope of AML, however the folks who routinely organize the AML workshop at the annual ICML conference define a reasonable scope on their website, which includes automating all of the repetitive tasks defined above. The scope of AML is ambitious, however, is it really effective? The answer is it depends on how you use it. Our view is that it is difficult to perform wholesale replacement of a data scientist with an AML framework

The war over artificial intelligence will be won with visual data

This battle will be won by owning the connected camera. The majority of the data our brains analyze is visual, and therefore the majority of the data needed for artificial intelligence to have human (or better than human) skills, will rely on the ability for computers to translate high quality visual data. One of the business sectors that will be revolutionized by artificial intelligence is e-commerce. The Amazon’s Echo Look is a smart stake in the ground for Amazon. Adding a camera to their Echo validates a prediction of mine from last year called the Internet of Eyes which enables all inanimate objects to see. Inanimate objects with cameras enable companies to own the first step in gathering the data for computer vision and artificial intelligence algorithms to analyze.

What businesses are failing to see about AI

When machines take care of crunching data, conducting micro-analysis, and managing workflow, humans are free to focus on the bigger picture. Imagine a marketing team huddled around a table, plotting strategy. Right now, if they have a question, they might have to ask an analyst and wait hours or days for a response. In a few years, that team will be able to ask an AI chatbot and get an answer within seconds. That will allow them to brainstorm more productively. It’s still the humans’ job to come up with a brilliant marketing strategy — the robots just help them do it quicker. Or consider Kensho, a financial analytics AI system. According to a Harvard Business review, the program can answer 65 million possible question combinations — even off-the-wall ones like “Which cement stocks go up the most when a Category 3 hurricane hits Florida?”

Superintelligence: Myth or Pressing Reality?

All binary computers are literal at their core. Ultimately, they operate based solely on binary numbers, ones or zeros. Humans, endowed with true embedded neural networks with astounding connective complexity, tend to increasingly learn literate behavior. ... Humans are uniquely capable of ascribing specific amplifying context. This makes us capable of abstracting meaning to localized events and utterances that might otherwise be legitimately interpreted in myriad ways. These attributes of mind permit humans to form vast social networks, thus creating living societies and enduring cultures. This time-honored behavior tends to reinforce, grow, and replicate human intelligence. It underscores the essence of Type III AI. For example, the ability of a Type II AI self-driving car to adhere to navigation and physical rules of the road becomes secondary to the Type III AI that understands the reason for the trip. Here, theory of mind adjusts AI to meet a broader worldview.

How CISOs can answer difficult questions from CEOs

CEO: We hear all the time about bigger threats and greater urgency. How do we get security right for our organization – and for our customers?; CISO: Proactive engagement between key stakeholders before an incident occurs will ensure that the organization is able to respond quickly and effectively to modern cyber threats. In the end, what we mean when we say that it's really important to get security right is that we must lead by example. We can't do the same old thing. We can't tell customers that they need to get to the “next-gen paradigm shifts” if we aren't doing these things ourselves. We must think prevention first, reduce the attack surface within our own environments, augment with strong detect and disrupt capabilities, and we must continue to innovate in automating security into the business.

A guide on how to prevent ransomware

At its heart, ransomware is simply another form of a computer virus, albeit a very potent one. The methods it uses to infect a computer are the same ones other computer viruses employ. This article details several recommendations to help you in reducing the likelihood of future infection by ransomware, or indeed any other computer viruses or malware, against systems within your organisation. Note that each of these recommendations should be assessed for their applicability to your specific environment and you should conduct a thorough risk assessment to determine if the recommendations outlined in this document are suitable for your environment and are proportionate to the identified threat and risk. For ease of use the recommendations in this document have been divided into three categories and colour coded accordingly.

Machine learning in cybersecurity moves needle, doesn't negate threats

Only 6% of respondents said they're either not planning on or not interested in deploying these technologies. Forte said it's no surprise that the appetite for AI and machine learning in cybersecurity is strong. Tech vendors and their corporate clients are deploying these advanced technologies in a variety of functions within the enterprise, and starting to see returns on investment. He said early use cases show that these tools likewise have great potential in cyber defense, too. "There is a little bit of hype right now, but I think it's a promising hype," agreed Sebastian Hess, the immediate past CISO of Isabel Group, a Belgium-based financial firm. Hess listed the advantages that machine learning and AI platforms bring to cybersecurity.

Fog Orchestration for Internet of Things Services

Traditional Web-based service applications are deployed on servers within cloud data centers that are accessed by end devices such as tablets, smartphones, and desktop PCs. In contrast, IoT applications deployed within fog computing systems consist of the cloud, fog node, and “things,” as Figure 1 shows. In this context, a fog node is defined as equipment or middleware and is served as an agent that collects data from a set of sensors. This data is then transmitted to a centralized computing system that locally caches data and performs load balancing. Things include sensors and devices with built-in sensors. Similar to Web-based service applications, the cloud provisions centralized resource pools (compute, storage) to analyze collected data and automatically trigger decisions based on a predefined system logic.

Quote for the day:

"Always do right. This will gratify some people and astonish the rest." -- Mark Twain