"Innovation" is such an exciting concept, it's a shame the mainstream conversation about it has gotten so boring. My Twitter feed is an endless scroll of promises about the next-best, game-changing technology that's going to be "über of" whatever industry. The buzzwords alone are enough to make my eyes glaze over, which is a shame because the startups and founders emerging today, no longer beholden to the Silicon Valley zip code, have fresh approaches to solving the world's pressing problems. So with the global proliferation of thought-provoking solutions in fields like artificial intelligence, blockchain, sustainability, and diversity, how do we overcome innovation fatigue and remember why we love startups and their technology in the first place? Here's my take: innovation is only as exciting as its application.

Are you an experienced Machine Learning / Computer Vision Engineer looking for a new opportunity? Vave Health is a startup based in the heart of Silicon Valley. Our mission is to provide the world with connected and personal tools that will help deliver better care, improve patient experience, and drive healthcare efficiency. Our next generation wireless connected device enables faster diagnosis and treatment at the point of care resulting in better patient outcomes. Vave Health is at the forefront of medical imaging and digital health.

Artificial intelligence (AI) marketing companies bagged $2.5bn of investment last year as marketers turned to the new technology to help analyse huge troves of data. Last year's investment surge has continued into 2019, with $1bn invested in the second quarter alone, according to figures compiled by tech investment firm GP Bullhound. The report shows that marketing AI remains a nascent sector, with private placements outnumbering merger and acquisition transactions. However, the steady rise highlights how marketers and increasingly looking to technology to help sort and analyse growing amounts of user data. "Artificial intelligence heralds the beginning of a new marketing era, driven by the need to connect vast amounts of disparate data, uncover patterns and make predictions, which only AI can accomplish," said Oliver Schweitzer, executive director at GP Bullhound.

Artificial Intelligence and Machine Learning solutions offer large possibilities to optimize and automate processes, save costs and make less human error possible for many industries. Food and Beverage is not an exception, where it can be beneficially applied in restaurants, bar and cafe businesses as well as in food manufacturing. These two segments have common use cases where AI in the food industry can be applied, as well as different ones, which is linked to different problems that must be solved. Knowing what goods to manufacture in large amounts or what dishes are the best choice to include in your restaurant menu is the key to increase earnings. Often customers' and market demands are changing very fast and so it is even more important to be one step ahead to take measures in time.

Adam Neilson, Chief Technology Officer at Wefarm discusses the ways in which machine learning can transform the African agricultural industry. Ever since Fritz Lang's Metropolis was first shown in the cinemas of 1927, the film industry has been forecasting how technology of the future would transform humanity. Fast forward to current day and we may not have flying cars or replica people mining in off planet worlds, but we do have something that I believe in the long run will be far more important to the future survival of our species. Over the last few years, machine learning (ML) has steadily rolled across the "hype cycle" from the "peak of inflated expectations" to officially entering the mainstream, and is now beginning to quietly revolutionise every aspect of our lives. For us consumers, it's now so deeply embedded within so many of the everyday products and services that we interact with it's almost invisible.

This model was open sourced back in June 2019 as an implementation of the paper Transfer Learning from Speaker Verification to Multispeaker Text-To-Speech Synthesis. This service is being offered by Resemble.ai. With this product, one can clone any voice and create dynamic, iterable, and unique voice content. Users input a short voice sample and the model -- trained only during playback time -- can immediately deliver text-to-speech utterances in the style of the sampled voice. Bengaluru's Deepsync offers an Augmented Intelligence that learns the way you speak.

For the second year in a row, Canada has refused visas to dozens of researchers - most of them from Africa - who were hoping to attend an artificial intelligence (AI) conference in Vancouver. The hassles have caused at least one other AI conference to choose a different country for their next event. The Neural Information Processing Systems conference (NeurIPS), which brings together thousands of experts and researchers from all over the world, will be held in Vancouver next month. Last week, NeurIPS began hearing that several attendees had had their visas denied. It was the second year in a row the conference has had visa troubles.

Intelligent robots, intelligent virtual assistants, intelligent cars intelligently driving themselves, intelligent search systems learning and already knowing our browsing habits, interests, knowing what we are going to do online and even in real life. Siri and Alexa, Tesla, Amazon and Google, artificially intelligent algorithms that are everywhere, able to do many things instead of us. In the future, AI is going to change everything. As for now, there are lots of discussions about 4 main AI trends that are prone to shape the AI mechanized future of mankind. Here they are: deep learning, facial recognition, cloud, privacy and policy.

A Configuration-Space Decomposition Scheme for Learning-based Collision Checking Yiheng Han 1, Wang Zhao 1, Jia Pan 2, Zipeng Y e 1, Ran Yi 1 and Y ong-Jin Liu 1† Abstract -- Motion planning for robots of high degrees-of- freedom (DOFs) is an important problem in robotics with sampling-based methods in configuration space C as one popular solution. Recently, machine learning methods have been introduced into sampling-based motion planning methods, which train a classifier to distinguish collision free subspace from in-collision subspace in C . In this paper, we propose a novel configuration space decomposition method and show two nice properties resulted from this decomposition. Using these two properties, we build a composite classifier that works compatibly with previous machine learning methods by using them as the elementary classifiers. Experimental results are presented, showing that our composite classifier outperforms state-of-the-art single-classifier methods by a large margin. A real application of motion planning in a multi-robot system in plant phenotyping using three UR5 robotic arms is also presented. I. INTRODUCTION Motion planning plays an important role in robotics, which finds a collision-free path to move a robot from a source to a target position.

We propose Neural Forest Learning (NFL), a novel deep learning based random-forest-like method. In contrast to previous forest methods, NFL enjoys the benefits of end-to-end, data-driven representation learning, as well as pervasive support from deep learning software and hardware platforms, hence achieving faster inference speed and higher accuracy than previous forest methods. Furthermore, NFL learns non-linear feature representations in CNNs more efficiently than previous higher-order pooling methods, producing good results with negligible increase in parameters, floating point operations (FLOPs) and real running time. We achieve superior performance on 7 machine learning datasets when compared to random forests and GBDTs. On the fine-grained benchmarks CUB-200-2011, FGVC-aircraft and Stanford Cars, we achieve over 5.7%, 6.9% and 7.8% gains for VGG-16, respectively. Moreover, NFL can converge in much fewer epochs, further accelerating network training. On the large-scale ImageNet ILSVRC-12 validation set, integration of NFL into ResNet-18 achieves top-1/top-5 errors of 28.32%/9.77%, which outperforms ResNet-18 by 1.92%/1.15% with negligible extra cost and the improvement is consistent under various architectures.