The face-recognition technology behind smartphones, self-driving cars and diagnostic imaging in healthcare has made massive strides of late. These examples all use solutions that make sense of objects in front of them, hence the term "computer vision" -- these computers are able to make sense of what they "see." During a recent Data Lab meetup at CompassRed in downtown Wilmington, Delaware, Chandra Kambhamettu, professor and director of the Video/Image Modeling and Synthesis Lab in the Department of Computer and Information Sciences at the University of Delaware, and Dave Wallin, manager of innovations at The Archer Group, offered a high-level explanation of how image technology works along with the deep learning technology that powers it. Much of the innovation in image recognition relies on deep learning technology, an advanced type of machine learning and artificial intelligence. Typical machine learning takes in data, pushes it through algorithms and then makes a prediction, making it appear that the computer is "thinking" and coming to its own conclusions.

The scale of Internet-connected systems has increased considerably, and these systems are being exposed to cyberattacks more than ever. The complexity and dynamics of cyberattacks require protecting mechanisms to be responsive, adaptive, and large-scale. Machine learning, or more specifically DRL, methods have been proposed widely to address these issues. By incorporating deep learning into traditional RL, DRL is highly capable of solving complex, dynamic, and especially high-dimensional cyber defense problems. This paper presents a survey of DRL approaches developed for cyber security.

Tesla has quietly acquired an artificial intelligence company to build out its Autopilot autonomous driving system. Tesla this week closed the acquisition of DeepScale, a company that uses sophisticated "deep neural networks" and other aspects of artificial intelligence to help a vehicle's in-car autonomous driving technology more effectively "see" what's around it. The news was first reported by CNBC, who cited sources claiming Tesla had acquired DeepScale. Tesla did not respond to Fortune's request for comment, and has so far remained quiet on the acquisition's details, but DeepScale CEO Forrest Iandola updated his LinkedIn account on Tuesday, saying that he has joined Tesla's Autopilot team to work on deep learning and autonomous driving. Tesla makes small acquisitions from time to time, but DeepScale appears to be its most significant acquisition since February, when the company announced that it would acquire Maxwell Technologies to improve its battery technology.

I recently caught up with Ashutosh Garg, CEO at Eightfold.ai to discuss how he and his team have deployed AI and machine learning to help with the needs of the talent management industry. With 6000 research citations, 50 patents, 35 peer-reviewed research publications, and the outstanding Ph.D. thesis award from UIUC for his Ph.D. thesis in Machine Learning, it's fair to say that Ashutosh is one of the world's experts in machine learning. After his time managing Search and Personalization efforts at both Google and IBM Research, Ashutosh founded Bloomreach, a leading vendor for Digital Experience Platforms. Now, he is applying his experience to the problem he is most truly passionate about --helping the world's talent find their most meaningful and fulfilling work. Can you give us a sense for what form of AI/machine learning is being used in your product?

Björn Brinne added: "The report is correct that there are a number of urgent challenges that need to be addressed. Many deep learning projects to date have been focused on small pockets of research, which presents issues in relation to repeatability, auditability and scalability which are needed to make a global impact. Also, lack of skills, cost and complexity remain as barriers. "For the NHS, this is a major challenge as budgets and talent are already limited. "There's also the data issue – deploying deep learning models in the health sector requires retraining them when new data comes in, a complex and often costly task. "Additionally, to begin with, the quantity of data available will be limited and the quality of it inconsistent, which could lead to inaccuracies. There are also obvious challenges in the sensitivity of the data that is needed and requirements for consent." "In order to overcome these challenges, deep learning needs to move away from being used as a research tool, and instead become operationalised to make outputs more robust and usable. This will make deep learning accessible for a wider group of users in the medical industry, so that data pools become greater and more varied over time, improving model performance and, by extension, the quality and effectiveness of patient care."

Practicing data science and working with real-world data and business problems is rather different than, for instance, building data science projects in Python using toy datasets. While being a part of a data science team in an enterprise, one should expect many challenges, including messy data, lack of data, unclear goals, difficult communication with business managers who want quick results, model performance in production being very different from testing performance, etc. Therefore, to become a successful data scientist with a portfolio of outstanding projects, it is not enough to be good at coding and building machine learning models. One should further be able to approach a project strategically and consider many different factors, not only from the viewpoint of a data scientist but also from a business perspective. However, what if you are actually not interested in succeeding in data science? In that case, read carefully through the tips provided below.

In Part 1 of this series, we explored the merits of modern identity documents, such as the e-Passport (or biometric passport), that include cryptographic features to aid machine verification of the presented document. In Part 2, we discussed how computer vision and deep learning can be used to improve the efficacy of machine-verifying legacy documents (i.e. In both posts, the focus was on machine-based techniques that can strongly evaluate the authenticity of a document presented as proof of identification, in an effort to prevent false documents from being accepted as valid.

One of AI's most powerful capabilities is being able to identify items or patterns and then apply that understanding to future activities. To enable an AI solution to do this, you need to feed it a large amount of data describing past events and teach it the attributes of those events. This requires the labor-intensive process of establishing a large body of high-quality data, then typically labeling items as positive, negative, or neutral. Teaching computers to recognize cats, as above, involved labeling lots of images of cats as positive.

Amongst a wave of buzzy deep-learning startups in the early 2010s, Tractable has managed to grow steadily by addressing a very specific problem with the technology: helping insurance companies to automatically assess vehicle damage. Founded by Alexandre Dalyac and Razvan Ranca in 2014 after graduating from the company builder programme at Entrepreneur First (EF), Adrien Cohen joined the company soon after as chief business officer, and set out to discover where to apply its computer vision expertise. "If you take a step back, at the time there was a wave of deep learning startups," Cohen told Techworld last week, speaking from a conference room at Tractable's London office, which is an entire floor of a WeWork near Old Street station in east London. "There was a big number of companies with great technology, but no problem to solve. Most of these companies, they ended up being acquired."

For decades Ai researchers weren't getting the results they wanted, and eventually many of the leading Ai scientists wondered if there was something wrong with their techniques to Ai software. The actual issue was not the software, but the hardware capabilities they were using to design and run the software. The hardware wasn't capable of handling what the Ai scientists wanted to do. This is obvious as now -- thanks to the exponential advancement of computing technologies including hardware virtualization and cloud computing -- we have hardware that is capable of supporting the kinds of software that can perform amazing amounts of processing tasks in parallel. A lot of this story and some very well-written explanations about the technology are available on Nvidia's company blog.