Many of today's headlines are talking about how machine learning and artificial intelligence are transforming industry. The healthcare sector is one of the major industries affected by this trend, in large part because organizations are recognizing the need for improved data processing and analysis, according to research and healthcare publications such as Health Network, ITHealthcareNews, and Healthcare Data Management. Analysts project that by 2018, 30 percent of providers will run cognitive analytics on patient data. For instance, IDC has said that healthcare will access cognitive solutions for close to 50 percent of cancer patients, resulting in reduced costs and mortality rates. Unfortunately, healthcare has been slow to digitize its information on patients, and it has been a struggle to use the overwhelming amount of data it collects.

With a simple one-time Q&A session – and without the need to pre-define a single keyword, rule, topic or sentiment value – a junior employee in days can arrange to have Revuze deliver the most nuanced information available. The data is delivered, via a single screen, about consumers and their needs, across thousands of internal and external data sources, on any given product family. Revuze has created and is widely introducing its technology available for Product Experience Management, enabling brands to quickly understand product and customer satisfaction issues, and to automatically score and rank the brand's performance, both relative to its competitors and to the market as a whole. "Global brands have told us that this is a game changer for managing their brand health," said Revuze Co-Founder and CEO Ido Ramati. "They are now able to make the kind of key business decisions and utilize data from such varied sources in ways that before were not possible. Through our introduction of AI into this field, brands can now gain an immediate, granular understanding of what customers are saying about their brand, product or competitors, without having to hire teams of experts."

Artificial Intelligence (AI) or machine learning can be used to help improve the accuracy of the diagnosis in lung diseases, finds a study. Machine learning utilises algorithms that can learn from and perform predictive data analysis. The team developed an algorithm process in addition to the routine lung function parameters and clinical variables of smoking history, body mass index (BMI) and age. Based on the pattern of both the clinical and lung function data, the algorithm makes a suggestion for the most likely diagnosis. "We have demonstrated that AI can provide us with a more accurate diagnosis. The algorithm can simulate the complex reasoning that a clinician uses to give their diagnosis, but in a more standardised and objective way so it removes any bias," said Wim Janssens from the University of Leuven in Belgium.

My first job in market research was in 2001 as Director of Call Center Operations for a healthcare-focused MR firm. We did NCQA CAHPS studies using a rigid CATI sampling method, and we sold excess capacity as a field & tab provider for other research companies. We also were bidding on many government studies such as the CDC BRFSS, so I was receiving a crash course in sample theory and applying very hardcore probability sampling models to the studies we were conducting. I had come from a customer service call center operations management background, so I knew how useful IVR (Interactive Voice Response) could be and was exploring if there was a role for that technology in a research paradigm, and through that effort was recruited by my vendor, DialTek, to become their VP of Operations. DialTek did outbound (they still power the polling of Rasmussen Reports) and inbound (lots of satisfaction surveys when calling into a customer service line) IVR surveys, and during my time there we experimented with using it as a recruitment tool for qual, to build online panels, as a partial solution to streamline phone surveys, etc… In other words, I know the technology very well and understand the best uses cases for it.

Intel has announced the acquisition of Movidius, a chip manufacturer focusing on developing next-generation computer sensing and vision technology. San Mateo, California-based Movidius, which already counts Google and Lenovo as customers, develops sight capabilities for machines and PCs. The company's vision processing unit (VPU), the main shunt of the company, is a platform for on-device vision processing which works in tandem with Intel RealSense technology to give computer systems the capability to view 3D images, understand surroundings and objects, and then react accordingly. This technology can be found within drones, security cameras, artificial intelligence and virtual reality, and as these industries develop, the potential use for such inventions will also increase. The financial terms of the deal were not disclosed.

I've been thinking, mostly in the context of teaching, about how to specifically teach debugging of machine learning. Personally I find it very helpful to break things down in terms of the usual error terms: Bayes error (how much error is there in the best possible classifier), approximation error (how much do you pay for restricting to some hypothesis class), estimation error (how much do you pay because you only have finite samples), optimization error (how much do you pay because you didn't find a global optimum to your optimization problem). I've generally found that trying to isolate errors to one of these pieces, and then debugging that piece in particular (eg., pick a better optimizer versus pick a better hypothesis class) has been useful.

Artificial Intelligence (AI) or machine learning can be used to help improve the accuracy of the diagnosis in lung diseases, finds a study. Machine learning utilises algorithms that can learn from and perform predictive data analysis. The team developed an algorithm process in addition to the routine lung function parameters and clinical variables of smoking history, body mass index (BMI) and age. Based on the pattern of both the clinical and lung function data, the algorithm makes a suggestion for the most likely diagnosis. "We have demonstrated that AI can provide us with a more accurate diagnosis. The algorithm can simulate the complex reasoning that a clinician uses to give their diagnosis, but in a more standardised and objective way so it removes any bias," said Wim Janssens from the University of Leuven in Belgium.

Based on the disturbing number of successful data breaches over the past few years, it's pretty evident that organizations are being overwhelmed by the growing number of threats. However, a new breed of security solution has sprung up, offering to apply machine learning to enterprise security. These tools deliver the ability to analyze networks, learn about them, detect anomalies and protect enterprises from threats.

Of the many misconceptions about machine learning, the idea that they can't be prejudiced is likely the most harmful. As stated by Moritz Hardt in How big data is unfair, machine learning is not, by default, fair or just in any meaningful way. Even though many researchers and practitioners have noted this repeatedly in the past, the message is still lost. It's not uncommon to hear variations of "algorithms don't have in-built bias" even when there is an entire field of research dedicated to fighting that very issue. To make this clearer, prejudice in machine learning will haunt us for years to come.

Can you tell the difference between talking to a human and talking to a machine? Or, is it possible to create a machine which is able to converse like a human? In fact, what is it that even makes us human? Turing's Imitation Game, commonly known as the Turing Test, is fundamental to the science of artificial intelligence. Involving an interrogator conversing with hidden identities, both human and machine, the test strikes at the heart of any questions about the capacity of machines to behave as humans.