Profile cover photo
Profile photo
Ward Plunet
59,978 followers
59,978 followers
About
Posts

Post has attachment
New developments enabling blind people to see again

Enabling blind people to see again is the dream of many neuroscientists. We still have a long way to go to make this happen, but we have also made a lot of progress over the last twenty years, says Richard van Wezel of the Donders Institute for Brain, Cognition and Behaviour. He presented his research into the development of a 'prosthetic for blind people' on the occasion of World Sight Day (12 October), an annual event that focuses attention on blindness and vision loss. Van Wezel and his colleague Marcel van Gerven belong to the NESTOR consortium, consisting of participants from a range of disciplines including neurobiologists and engineers specialized in microelectronics and wireless apparatus. NESTOR, which received a grant last November from NWO Applied and Engineering Science AES, is working on the development of a prosthesis that uses micro-electrodes to stimulate the brains of blind people to evoke phosphenes. "These are phosphenes, comparable to the stars you see when you stand up too quickly. Blind people can also perceive them," Van Wezel explains. "We use electrodes to stimulate the brain in such a way that blind people can have a limited form of vision to see what is happening in the world around them." It is a potential solution for people who have become blind because their eyes or optical nerves are no longer functional. "For this group, stimulating the visual cortex is the only option for restoring vision."
Add a comment...

Post has attachment
Understand Deep Residual Networks — a simple, modular learning framework that has redefined state-of-the-art

Deep residual networks took the deep learning world by storm when Microsoft Research released Deep Residual Learning for Image Recognition. These networks led to 1st-place winning entries in all five main tracks of the ImageNet and COCO 2015 competitions, which covered image classification, object detection, and semantic segmentation. The robustness of ResNets has since been proven by various visual recognition tasks and by non-visual tasks involving speech and language. This post will summarize the three papers below, with simple and clean Keras implementations of the network architectures discussed. You will have a solid understanding of residual networks and their implementation by the end of this post.

Deep Residual Learning for Image Recognition — ResNet (Microsoft Research)
Wide Residual Networks (Université Paris-Est, École des Ponts ParisTech)
Aggregated Residual Transformations for Deep Neural Networks — ResNeXt (Facebook AI Research)
Add a comment...

Post has attachment
DeepMind’s Go-playing AI doesn’t need human help to beat us anymore

Google’s AI subsidiary DeepMind has unveiled the latest version of its Go-playing software, AlphaGo Zero. The new program is a significantly better player than the version that beat the game’s world champion earlier this year, but, more importantly, it’s also entirely self-taught. DeepMind says this means the company is one step closer to creating general purpose algorithms that can intelligently tackle some of the hardest problems in science, from designing new drugs to more accurately modeling the effects of climate change. The original AlphaGo demonstrated superhuman Go-playing ability, but needed the expertise of human players to get there. Namely, it used a dataset of more than 100,000 Go games as a starting point for its own knowledge. AlphaGo Zero, by comparison, has only been programmed with the basic rules of Go. Everything else it learned from scratch. As described in a paper published in Nature today, Zero developed its Go skills by competing against itself. It started with random moves on the board, but every time it won, Zero updated its own system, and played itself again. And again. Millions of times over. After three days of self-play, Zero was strong enough to defeat the version of itself that beat 18-time world champion Lee Se-dol, winning handily — 100 games to nil. After 40 days, it had a 90 percent win rate against the most advanced version of the original AlphaGo software. DeepMind says this makes it arguably the strongest Go player in history. “By not using human data — by not using human expertise in any fashion — we’ve actually removed the constraints of human knowledge,” said AlphaGo Zero’s lead programmer, David Silver, at a press conference. “It’s therefore able to create knowledge itself from first principles; from a blank slate [...] This enables it to be much more powerful than previous versions.”
Add a comment...

Post has attachment
Faster metabolism makeover—nurturing your gut bacteria

Here's how to take control of your cravings and lose weight for good by improving your gut health. You're not alone! Right now, you've got 100 trillion bacteria living in your digestive system. Most of us are familiar only with these tiny critters' embarrassing habit of releasing smelly gas at the wrong moments, but the truth is, your gut bugs are intimately involved with your weight. There's growing evidence that the right mix of bacteria in your intestines can help you make healthier food choices and stay slim, while the wrong mix encourages weight gain and a taste for junky processed foods. When researchers carefully checked the types of bacteria found from the digestive systems of 154 people, they found that those who were obese had the smallest variety of gut bacteria. A lab study with mice from the same team found that having more of a type of bacteria called Firmicutes may be related to weight gain. These bugs are great at sucking more kilojoules out of food—digesting complex sugars that other bacteria can't and converting them into simple sugars and fatty acids that get absorbed from your intestines into your bloodstream. In contrast, having more of a type of bacteria called Bacteroidetes has been associated with a slimmer physique. Gut bugs help control your weight in several ways, research suggests. Some send more kilojoules into your body, where they're likely to be stored as fat. But that's not all. Scientists have found that the bacteria Helicobacter pylori is involved in the regulation of certain hormones, including the hunger hormone ghrelin. While nobody wants an overabundance of H. pylori (it can cause painful stomach ulcers), the researchers note that the widespread use of antibiotics has reduced levels of H. pylori and could be making weight loss more difficult. In a 2011 study of 92 people published in the journal BMC Gastroenterology, found those who were prescribed antibiotics to knock out H. pylori (due to digestive-system problems) also saw ghrelin levels rise sixfold after the bacteria were completely eliminated. And in a recent lab study in mice, researchers found that a fatty acid called acetate, which is pumped out by gut bacteria, increased eating behaviours. The elevated release of acetate also increased production of ghrelin and of insulin, a key blood sugar control hormone that also promotes the storage of body fat....For example, when 21 people increased their daily intake of fibre by 21 grams, they had more Bacteroidetes and fewer Firmicutes in their systems after 3 weeks, according to a study. Bacteria, especially the good guys, love munching on the fibre found in abundance in clean foods like fruits, vegetables and whole grains. But loading up on processed junk food takes things in the other direction in a hurry.
Add a comment...

Post has attachment
Protecting Against AI’s Existential Threat

How to avoid the nightmare scenario of artificial intelligence? According to researchers from Elon Musk’s OpenAI, the trick is teaching machines to keep our interests in mind. On July 8, 2017, an AI system built by our research company, OpenAI, beat a semipro human player in solo matches of a battle arena video game called Dota 2. One month later, the same AI system beat a professional gamer ranked in the top 50. Three days after that it defeated the No. 1 solo Dota 2 player in the world. And it kept getting better: The Aug. 11 version of our AI beat the Aug. 10 version 60% of the time. Our AI learned to trick its opponents, predict what it couldn’t see and decide when to fight and when to flee. How do you create AI that doesn’t pose a threat to humanity? By teaching it to work with humans. Open AI collaborated with DeepMind, Google’s AI division, to design a training method that incorporates regular human feedback. The idea is to “humanize” AI systems by teaching them not only skills but also complex motivations and subtle goals that must be communicated precisely. A promising direction for AI safety, it ensures AI’s aims align with ours, no matter how hard those aims are to articulate.

Add a comment...

Post has attachment
Transfer Learning - the next leap in Machine Learning

In recent years, we have become increasingly good at training deep neural networks to learn a very accurate mapping from inputs to outputs, whether they are images, sentences, label predictions, etc. from large amounts of labeled data. What our models still frightfully lack is the ability to generalize to conditions that are different from the ones encountered during training. When is this necessary? Every time you apply your model not to a carefully constructed dataset but to the real world. The real world is messy and contains an infinite number of novel scenarios, many of which your model has not encountered during training and for which it is in turn ill-prepared to make predictions. The ability to transfer knowledge to new conditions is generally known as transfer learning and is what we will discuss in the rest of this post. Over the course of this blog post, I will first contrast transfer learning with machine learning's most pervasive and successful paradigm, supervised learning. I will then outline reasons why transfer learning warrants our attention. Subsequently, I will give a more technical definition and detail different transfer learning scenarios. I will then provide examples of applications of transfer learning before delving into practical methods that can be used to transfer knowledge. Finally, I will give an overview of related directions and provide an outlook into the future.
Add a comment...

Post has attachment
On-and-off fasting helps fight obesity

Up to sixteen weeks of intermittent fasting without otherwise having to count calories helps fight obesity and other metabolic disorders. Such fasting already shows benefits after only six weeks. This is according to a study by Kyoung-Han Kim and Yun Hye Kim in the journal Cell Research. Intermittent fasting in mice helped to kick-start the animals' metabolism and to burn fat by generating body heat. The research team was led by Hoon-Ki Sung of The Hospital for Sick Children in Ontario, Canada. Research has shown that our unhealthy eating habits and sedentary lifestyles are playing a major role in the development of lifestyle-related metabolic diseases such as diabetes, heart disease and obesity. For this reason, dietary interventions like intermittent fasting are gaining popularity to treat conditions such as obesity. The research team in this study wanted to better understand the reactions that interventions such as fasting trigger on a molecular level in the body. They exposed groups of mice to sixteen weeks of intermittent fasting. The recurring regimen saw the animals being fed for two days, followed by one day without anything to eat. Their calorie intake was not adjusted otherwise. Four months later the mice in the fasting group weighed less than those in the control group who continued to eat the same volume of food. The lower body weight of the mice in the fasting group was not the only effect. The fasting regime helped lower fat build-up in the white fat by increasing the brown-like fat (involved in burning energy and producing body heat) of mice on the high fat diet. Their glucose and insulin systems also remained more stable. In a further experiment, similar benefits were already seen after only six weeks of intermittent fasting.
Add a comment...

Post has attachment
Allen Institute neuroscientist Christof Koch on our future with AI
Add a comment...

Post has attachment
Microsoft joins the VR battle with Windows Mixed Reality today

Microsoft is launching its own answer to virtual reality today, taking on HTC and Oculus in the process. Windows Mixed Reality will be available in the Windows 10 Fall Creators Update, and headsets are now available to buy. Here’s everything you need to know about Windows Mixed Reality. While Microsoft has picked the “Mixed Reality” naming for its initial headsets, they’re only capable of virtual reality experiences right now. Microsoft’s range of headsets are similar to the Oculus Rift and HTC Vive, and many manufacturers are selling bundles that include touch controllers. The main difference between the Vive / Rift and Windows Mixed Reality is that the headsets do not require separate sensors.
Add a comment...

Post has attachment
Generalization in Deep Learning

By the group headed by Yoshua Bengio

This paper explains why deep learning can generalize well, despite large capacity and possible algorithmic instability, nonrobustness, and sharp minima, effectively addressing an open problem in the literature. Based on our theoretical insight, this paper also proposes a family of new regularization methods. Its simplest member was empirically shown to improve base models and achieve state-of-the-art performance on MNIST and CIFAR-10 benchmarks. Moreover, this paper presents both data-dependent and data-independent generalization guarantees with improved convergence rates. Our results suggest several new open areas of research.
Add a comment...
Wait while more posts are being loaded