PARIS – A computer was better than human dermatologists at detecting skin cancer in a study that pitted human against machine in the quest for better, faster diagnostics, researchers said Tuesday. A team from Germany, the United States and France taught an artificial intelligence system to distinguish dangerous skin lesions from benign ones, showing it more than 100,000 images. The machine -- a deep learning convolutional neural network or CNN -- was then tested against 58 dermatologists from 17 countries, shown photos of malignant melanomas and benign moles. Just over half the dermatologists were at "expert" level with more than five years of experience, 19 percent had between two and five years' experience, and 29 percent were beginners with less than two years under their belt. "Most dermatologists were outperformed by the CNN," the research team wrote in a paper published in the journal Annals of Oncology.

"This would result in less unnecessary surgery." The computer was more accurate even when compared with the most experienced doctors, the study, published in the Annals of Oncology, found. "These findings show that deep learning convolutional neural networks are capable of out-performing dermatologists, including extensively trained experts, in the task of detecting melanomas," Prof Haenssle said. Researchers said such technology could be used to screen for skin cancer, meaning cases could be diagnosed far earlier. But they said the technology was likely to be used in conjunction with skilled doctors, rather than replacing them.

Some schools in China have incorporated paper-grading artificial intelligence into their classrooms, according to the South China Morning Post. One in every four schools, or about 60,000 institutions, are quietly testing a machine learning-powered system that can score students' work automatically, and even offer suggestions where appropriate. The AI, which can be accessed through various online portals, and which the report describes as similar to the system used by the Education Testing Service in the U.S., uses an evolving "knowledge base" to interpret the "general logic" and "meaning" of pupils' essays and to highlight stylistic, structural, and thematic areas that need improvement. It can read both English and Chinese, and it's reportedly perceptive enough to notice when paragraphs veer too far off topic. The 10-year-old grading software leverages deep learning algorithms to "compare notes" with human teachers' scores, suggestions, and comments.

Last quarter we reviewed artificial intelligence exit trends and saw strong growth. We now dig in one level deeper on our AI report and research platform to examine exits by category. We conclude that Deep Learning Applications and Computer Vision Platforms are at the forefront of AI exit activity. We'll illustrate these takeaways with some graphics that show AI exit activity by category. To help set the stage, the graphic below shows AI exit activity over time.

The artificial intelligence (AI) industry will be worth $1.2 trillion in 2018, with customer experience solutions creating the most business value. On Wednesday, Gartner released estimates on the projected value of AI over the course of this year. According to the research firm, the global enterprise value derived from AI will total $1.2 trillion this year, a 70 percent increase from 2017. AI-derived business value is projected to reach up to $3.9 trillion by 2022. "AI promises to be the most disruptive class of technologies during the next 10 years due to advances in computational power, volume, velocity and variety of data, as well as advances in deep neural networks (DNNs)," said John-David Lovelock, research vice president at Gartner.

Just released as part of Intel's Vision products lineup, Intel announces the new OpenVINO toolkit (Open Visual Inference and Neural Network Optimization, and formerly the Intel Computer Vision SDK) to help developers bring vision intelligence into their applications from edge to cloud. The toolkit is a free download that helps fast-track development of high-performance computer vision and deep learning inference solutions, and deliver fast and efficient deep learning workloads across multiple types of Intel platforms (CPU, CPU with integrated graphics, FPGA, and Movidius vision processing units (VPUs). Vision systems hold incredible promise to change the world and help us solve problems--whether they're making homes safer or discovering new medical cures--affording great opportunities for developers. The toolkit has a common API and is based on common development standards, such as OpenCL, OpenCV and OpenVX. Components to increase performance of Intel Processor Graphics for Linux*, including the Intel Media SDK open source version and OpenCL graphics drivers and runtimes.

When dermatologists received more clinical information and images at level II, their diagnostic performance improved. However, the CNN, which was still working solely from the dermoscopic images with no additional clinical information, continued to outperform the physicians' diagnostic abilities,

We develop a new learning-based approach to ill-posed inverse problems. Instead of directly learning the complex mapping from the measured data to the reconstruction, we learn an ensemble of simpler mappings from data to projections of the unknown model into random low-dimensional subspaces. We form the reconstruction by combining the estimated subspace projections. Structured subspaces of piecewise-constant images on random Delaunay triangulations allow us to address inverse problems with extremely sparse data and still get good reconstructions of the unknown geometry. This choice also makes our method robust against arbitrary data corruptions not seen during training. Further, it marginalizes the role of the training dataset which is essential for applications in geophysics where ground-truth datasets are exceptionally scarce.

The theory of situated cognition postulates that language is inseparable from its physical context--words, phrases, and sentences must be learned in the context of the objects or concepts to which they refer. Yet, statistical language models are trained on words alone. This makes it impossible for language models to connect to the real world--the world described in the sentences presented to the model. In this paper, we examine the generalization ability of neural language models trained with a visual context. A multimodal connectionist language architecture based on the Differential State Framework is proposed, which outperforms its equivalent trained on language alone, even when no visual context is available at test time. Superior performance for language models trained with a visual context is robust across different languages and models.

Even though probabilistic treatments of neural networks have a long history, they have not found widespread use in practice. Sampling approaches are often too slow already for simple networks. The size of the inputs and the depth of typical CNN architectures in computer vision only compound this problem. Uncertainty in neural networks has thus been largely ignored in practice, despite the fact that it may provide important information about the reliability of predictions and the inner workings of the network. In this paper, we introduce two lightweight approaches to making supervised learning with probabilistic deep networks practical: First, we suggest probabilistic output layers for classification and regression that require only minimal changes to existing networks. Second, we employ assumed density filtering and show that activation uncertainties can be propagated in a practical fashion through the entire network, again with minor changes. Both probabilistic networks retain the predictive power of the deterministic counterpart, but yield uncertainties that correlate well with the empirical error induced by their predictions. Moreover, the robustness to adversarial examples is significantly increased.