Machine learning fairness is a young subfield of machine learning that has been growing in popularity over the last few years in response to the rapid integration of machine learning into social realms. Computer scientists, unlike doctors, are not necessarily trained to consider the ethical implications of their actions. It is only relatively recently (one could argue since the advent of social media) that the designs or inventions of computer scientists were able to take on an ethical dimension. This is demonstrated in the fact that most computer science journals do not require ethical statements or considerations for submitted manuscripts. If you take an image database full of millions of images of real people, this can without a doubt have ethical implications. By virtue of physical distance and the size of the dataset, computer scientists are so far removed from the data subjects that the implications on any one individual may be perceived as negligible and thus disregarded. In contrast, if a sociologist or psychologist performs a test on a small group of individuals, an entire ethical review board is set up to review and approve the experiment to ensure it does not transgress across any ethical boundaries.
Operators of paint shops, warehouses, and call centers have reached the same conclusion. Rather than replace humans, they employ machines alongside people, to make them more efficient. The reasons stem not just from sentimentality but because many everyday tasks are too complex for existing technology to handle alone. With that in mind, the dermatology researchers tested three ways doctors could get help from an image analysis algorithm that outperformed humans at diagnosing skin lesions. They trained the system with thousands of images of seven types of skin lesion labeled by dermatologists, including malignant melanomas and benign moles. One design for putting that algorithm’s power into a doctor’s hands showed a list of diagnoses ranked by probability when the doctor examined a new image of a skin lesion. Another displayed only a probability that the lesion was malignant, closer to the vision of a system that might replace a doctor. A third retrieved previously diagnosed images that the algorithm judged to be similar, to provide the doctor some reference points.
Intelligent behaviour has long been considered a uniquely human attribute. But when computer science and IT networks started evolving, artificial intelligence and people who stood by them were on the spotlight. AI in today’s world is both developing and under control. Without a transformation here, AI will never fully deliver the problems and dilemmas of business only with data and algorithms. Wise leaders do not only create and capture vital economic values, rather build a more sustainable and legitimate organisation. Leaders in AI sectors have eyes to see AI decisions and ears to hear employees perspective. A futuristic AI leader plans to work not just for now but also for the years ahead. A company’s development in AI involves automating business processes using robotic technologies, gaining insight through data analysis and enhancement, cost-effective predictions based on algorithms and engagement with employees through natural language processing chatbots, intelligent agents and machine learning. Without a far-sighted leader, bringing all this to reality will be merely impossible.
In the context of a large-scale shipping operation, for instance, there may be thousands of containers filled with millions of packages or assets. Using a system that can track every asset with full certainty, any concerns can be eliminated about whether the items are where they are supposed to be, or if anything is missing. As blockchain expands, so too will the data it records, which in turn increases trust. By ensuring via this secured digital ledger that an asset has moved from a warehouse to a lorry on a Thursday afternoon, more data can then be added. For example, it can show that the asset moved from a specific shelf in a warehouse on a specific street and was moved by a specific truck operated by a specific driver. Securing the location data with full trust provides assurance that things are happening correctly and means that financial transactions can be made with more confidence. Layering mapping capabilities and rich location data to a blockchain record also enables fraud detection. Without blockchain, it cannot be certain that the delivery updates provided are in fact accurate. Blockchain makes transactions transparent and decentralised, enabling the possibility to automatically verify their accuracy by matching the real location of an item with the location report from a logistics company.
At Ignite, Microsoft provided its answer on how Azure Arc brings cloud control on premises. The cornerstone of Azure Arc is Azure Resource Manager, the nerve center that is used for creating, updating, and deleting resources in your Azure account. That encompasses allocating compute and storage to specific workloads and then monitoring performance, policy compliance, updates and patches, security status, and so on. You can also fire up and access Azure Resource Manager through several paths ranging from the Azure Portal to APIs or command line interface (CLI). It provides a single pane of glass for indicating when specific servers are out of compliance; specific VMs are insecure; or certificates or specific patches are out of date – and it can then show recommended remedial actions for IT and development teams to take. While it requires at least some connection to the Azure Public Cloud, it can run offline when the network drops. Microsoft has built a lot of flexibility as to the environments that Azure Arc governs. It can be used for controlling bare metal environments as well as virtual machines running on any private or public cloud, SQL Server, or Kubernetes (K8s) clusters.
Cybersecurity-related incentives are misaligned and often perverse. If you had a real chance to become a millionaire or even a billionaire by ignoring security and a much smaller chance if you slowly baked in security, which path would you choose? We also fail to account for, and sometimes flat out ignore, the unintended consequences and harmful effects of the innovative technology and ideas we create. Who would have thought that a 2003 social media app, built in a dorm room, would later help topple governments and make the creator one of the richest people in the world? Cybersecurity companies and individual experts face the difficult challenge of balancing personal gain versus the greater good. If you develop a new offensive tool or discover a new vulnerability, should you keep it secret or make a name for yourself through disclosure? Concerns over liability and competitive advantage inhibit the sharing of best practices and threat information that could benefit the larger business ecosystem. Data has become the coin of the realm in the modern age. Data collection is central to many business models, from mature multi-national companies to new start-ups. Have a data blind spot?
Differential privacy is a technique for sharing knowledge or analytics about a dataset by drawing the patterns of groups within the dataset and at the same time reserving sensitive information concerning individuals in the dataset. The concept behind differential privacy is that if the effect of producing an arbitrary single change in the database is small enough, the query result cannot be utilised to infer much about any single person, and hence provides privacy. Another way to explain differential privacy is that it is a constraint on the algorithms applied to distribute aggregate information on a statistical database, which restricts the exposure of individual information of database entries. Fundamentally, differential privacy works by adding enough random noise to data so that there are mathematical guarantees of individuals’ protection from reidentification. This helps in generating the results of data analysis which are the same whether or not a particular individual is included in the data. Facebook has utilised the technique to protect sensitive data it made available to researchers analysing the effect of sharing misinformation on elections. Uber employs differential privacy to detect statistical trends in its user base without exposing personal information.
Originally, hardware was only capable of performing a fixed set of operations on input vertices. An application was only able to set different transformation matrices (such as world, camera, projection, etc.) and instruct the hardware how to transform input vertices with these matrices. This was very limiting in what an application could do with vertices, so to generalize the stage, vertex shaders were introduced. Vertex shaders were a huge improvement over fixed-function vertex transform stage because now developers were free to implement any vertex process algorithm. There was however a big limitation - a vertex shader takes exactly one vertex as input and produces exactly one vertex as output. Implementing more complex algorithms that would require processing entire primitives or generate them entirely on the GPU was not possible. This is where geometry shaders were introduced, which was an optional stage after the vertex shaders. Geometry shader takes the whole primitive as an input and may output zero, one or more primitives.
Today's huge influx of data is resulting in multiple inefficiencies, according to Mike Sprunger, senior manager of cloud and network security at Insight Enterprises, a global technology solution provider. He cited the example of an employee who generates a spreadsheet and shares it with half a dozen co-workers, who then send the spreadsheet to half a dozen others. The 1 MB file morphs into 36 MBs, and when that information is backed up, data volumes double again. As cloud and flash technologies lowered storage pricing dramatically, many companies simply added more storage capacity as data demands grew. While companies stored more, they purged less. Furthermore, industry and government rules and guidelines for maintaining data have been evolving. So it can be unclear how to meet regulatory requirements, Sprunger noted, and decide what data can go and what must be kept. Compounding the challenge, communication between IT and business units is often mediocre or nonexistent. So neither group understands the business requirements or the technical possibilities of deleting outdated data, he added.
Feiman says solutions like RASP and WAF have emerged from "desperation" to protect application data but are insufficient. The market needs a technology that is focused on detection rather than prevention. Indeed, in an effort to address the problems with RASP, he and his team at WhiteHat are in the process of beta testing an application security technology that performs app testing without instrumentation. As far as existing RASP technologies go, it's unlikely they'll stick around in their current form. Rather than an independent technology, Feiman believes RASP will ultimately get absorbed into application runtime platforms like the Amazon AWS and Microsoft Azure cloud platforms. This could happen through a combination of acquisitions and companies like AWS building their own lightweight RASP capabilities into their technologies. "The idea will stay, the market hardly will," says Feiman. On that, Sqreen's Aviat disagrees, saying RASP is "indeed a standalone technology." "I expect RASP to become a crucial element of any application security strategy, just like WAF or SCA is today – in fact, RASP is already referenced by NIST as critical to lowering your application security risk," he said.
Quote for the day:
"The leader has to be practical and a realist, yet must talk the language of the visionary and the idealist." -- Eric Hoffer