CHENNAI: Google DeepMind's AlphaGo, an artificial intelligence programme developed using deep neural networks and machine learning techniques, hit global headlines last year when it beat South Korean Go grandmaster Lee Sedol to win the series 4-1. However, not many know that AlphaGo has consumed a whopping 30,000 watts of power to complete the task, while the human brain consumes around 20 watts! What gives the human brain such efficiency has so far proven elusive to replicate in computers. Not surprisingly, man's most defining organ is also the least understood. Although an adult human brain weighing 1.4 kg is made up of close to 100 billion neurons, scientists do not know how many different kinds of human neurons exist.

Artificial neural networks are a fascinating area of study, although they can be intimidating when just getting started. There are a lot of specialized terminology used when describing the data structures and algorithms used in the field. In this post you will get a crash course in the terminology and processes used in the field of multi-layer perceptron artificial neural networks. Crash Course In Neural Networks Photo by Joe Stump, some rights reserved. We are going to cover a lot of ground very quickly in this post.

Lantern Pharma Inc., is a privately held, global biotech company pioneering the field of precision oncology. Lantern's proprietary approach to drug development is driven by advanced genomics and machine learning-based artificial intelligence (AI), which when combined, are advancing a new wave of precision drug therapies that significantly reduce the time to market and overall risk associated with drug development. Lantern has reached an agreement to collaborate with India-based AI and data analytics company, Intuition Systems. Intuition will work closely with Lantern's existing computational team to bring additional AI, big data analysis, cloud support and infrastructure to support drug development and biomarker identification. The treatment of cancer represents a large market with many underserved areas where precision therapies will be highly valued.

New York state's top cops want to use machine-learning algorithms to detect drug dealers on social media networks like Instagram, a trend that "has become a severe problem in recent years," according to researchers from the University of Rochester and the New York Attorney General's office. Using social media to sell drugs began years ago and continues to this day. Newer networks like Tinder have become especially popular with drug dealers because they offer both sellers and customers a deal in close proximity. All of the networks rely on manual user reports to remove the illegal content in what has largely been a losing battle. The New York Attorney General's office co-authored new research on algorithms meant to examine millions of Instagram posts, spotlight drug dealers, and only then pass the suspects on to human officers for further investigation.

Nancy Brinker is a cancer advocate, a global consultant and founder of Susan G. Komen. Her opinions expressed in this article are her own. Elad Gil, Ph.D. is the chairman and co-founder of Color Genomics. The complexity of seeking a cure for cancer has vexed researchers for decades. While they've made remarkable progress, they are still waging a battle uphill as cancer remains one of the leading causes of death worldwide.

We present a general framework, the coupled compound Poisson factorization (CCPF), to capture the missing-data mechanism in extremely sparse data sets by coupling a hierarchical Poisson factorization with an arbitrary data-generating model. We derive a stochastic variational inference algorithm for the resulting model and, as examples of our framework, implement three different data-generating models---a mixture model, linear regression, and factor analysis---to robustly model non-random missing data in the context of clustering, prediction, and matrix factorization. In all three cases, we test our framework against models that ignore the missing-data mechanism on large scale studies with non-random missing data, and we show that explicitly modeling the missing-data mechanism substantially improves the quality of the results, as measured using data log likelihood on a held-out test set.

We study probability measures induced by set functions with constraints. Such measures arise in a variety of real-world settings, where prior knowledge, resource limitations, or other pragmatic considerations impose constraints. We consider the task of rapidly sampling from such constrained measures, and develop fast Markov chain samplers for them. Our first main result is for MCMC sampling from Strongly Rayleigh (SR) measures, for which we present sharp polynomial bounds on the mixing time. As a corollary, this result yields a fast mixing sampler for Determinantal Point Processes (DPPs), yielding (to our knowledge) the first provably fast MCMC sampler for DPPs since their inception over four decades ago. Beyond SR measures, we develop MCMC samplers for probabilistic models with hard constraints and identify sufficient conditions under which their chains mix rapidly. We illustrate our claims by empirically verifying the dependence of mixing times on the key factors governing our theoretical bounds.

Large labeled training sets are the critical building blocks of supervised learning methods and are key enablers of deep learning techniques. For some applications, creating labeled training sets is the most time-consuming and expensive part of applying machine learning. We therefore propose a paradigm for the programmatic creation of training sets called data programming in which users express weak supervision strategies or domain heuristics as labeling functions, which are programs that label subsets of the data, but that are noisy and may conflict. We show that by explicitly representing this training set labeling process as a generative model, we can "denoise" the generated training set, and establish theoretically that we can recover the parameters of these generative models in a handful of settings. We then show how to modify a discriminative loss function to make it noise-aware, and demonstrate our method over a range of discriminative models including logistic regression and LSTMs. Experimentally, on the 2014 TAC-KBP Slot Filling challenge, we show that data programming would have led to a new winning score, and also show that applying data programming to an LSTM model leads to a TAC-KBP score almost 6 F1 points over a state-of-the-art LSTM baseline (and into second place in the competition). Additionally, in initial user studies we observed that data programming may be an easier way for non-experts to create machine learning models when training data is limited or unavailable.

We propose a dual pathway, 11-layers deep, three-dimensional Convolutional Neural Network for the challenging task of brain lesion segmentation. The devised architecture is the result of an in-depth analysis of the limitations of current networks proposed for similar applications. To overcome the computational burden of processing 3D medical scans, we have devised an efficient and effective dense training scheme which joins the processing of adjacent image patches into one pass through the network while automatically adapting to the inherent class imbalance present in the data. Further, we analyze the development of deeper, thus more discriminative 3D CNNs. In order to incorporate both local and larger contextual information, we employ a dual pathway architecture that processes the input images at multiple scales simultaneously. For post-processing of the network's soft segmentation, we use a 3D fully connected Conditional Random Field which effectively removes false positives. Our pipeline is extensively evaluated on three challenging tasks of lesion segmentation in multi-channel MRI patient data with traumatic brain injuries, brain tumors, and ischemic stroke. We improve on the state-of-the-art for all three applications, with top ranking performance on the public benchmarks BRATS 2015 and ISLES 2015. Our method is computationally efficient, which allows its adoption in a variety of research and clinical settings. The source code of our implementation is made publicly available.

We pick the cell size to be bounded by (r)/( (n)), so that each grid cell will contain at most one sample, and thus the grid can be implemented as a simple n-dimensional array of integers: the default 1 indicates no sample, a non-negative integer gives the index of the sample located in a cell. Step 1. Select the initial sample, x0, randomly chosen uniformly from the domain.