Chinese artificial intelligence start-up SenseTime Group is planning an initial public offering (IPO) and aims to open a research and development (R&D) center in the United States as early as next year, its founder told Reuters in an interview. The Hong Kong and Beijing-based deep learning company founded by Tang Xiaoou, a professor at the Chinese University of Hong Kong, is a leader among Chinese AI start-ups that are enjoying fast growth thanks to demand from the government and private sector for their facial recognition technology.[nL4N1N72PS]

Ahead of Amazon's big AWS division Re:invent conference next week, the company has announced two developments in the area of artificial intelligence. AWS is opening a machine learning lab, ML Solutions Lab, to pair Amazon machine learning experts with customers looking to build solutions using the AI tech. And it's releasing new features within Amazon Rekognition, Amazon's deep learning-based image recognition platform: real-time face recognition and the ability to recognize text in images. The new lab and the enhancements to its image recognition platform underscore the push that Amazon and AWS are giving to AI at the company, both internally and as a potential area to grow its B2B business in this area. They come about a month after AWS announced it would be collaborating with Microsoft on Gluon, a deep learning interface designed for developers to build and run machine learning models for their apps and other services.

Article by Sunggy Koo, Samsung Electronics vice president of smart appliance AI, where are we now and where are we going? Since it was first envisioned in the 1950s, AI has made a palpable impact on our lives, giving us practical speech recognition, more effective web search and self-driving cars, among other innovations. Earlier this month, Google's AlphaGo AI program made news by mastering the ancient Chinese board game Go in just three days without any human assistance. This major advance comes just two decades after Deep Blue crushed chess grandmaster Garry Kasparov, illustrating that AI has not only come a long way in a short time but is on track to creating unthinkable opportunities across all industries that will add new value to our lives. The recent explosion in AI is enabled by a number of factors including a wider availability of GPUs, virtually infinite amounts of data, and more advanced machine and deep learning algorithms.

During training, the student network starts off in a random state. It is fed random white noise as an input and is tasked with producing a continuous audio waveform as output. The generated waveform is then fed to the trained WaveNet model, which scores each sample, giving the student a signal to understand how far away it is from the teacher network's desired output. Over time, the student network can be tuned - via backpropagation - to learn what sounds it should produce. Put another way, both the teacher and the student output a probability distribution for the value of each audio sample, and the goal of the training is to minimise the KL divergence between the teacher's distribution and the student's distribution.

We present a new algorithm that significantly improves the efficiency of exploration for deep Q-learning agents in dialogue systems. Our agents explore via Thompson sampling, drawing Monte Carlo samples from a Bayes-by-Backprop neural network. Our algorithm learns much faster than common exploration strategies such as $\epsilon$-greedy, Boltzmann, bootstrapping, and intrinsic-reward-based ones. Additionally, we show that spiking the replay buffer with experiences from just a few successful episodes can make Q-learning feasible when it might otherwise fail.

Critical periods are phases in the early development of humans and animals during which experience can affect the structure of neuronal networks irreversibly. In this work, we study the effects of visual stimulus deficits on the training of artificial neural networks (ANNs). Introducing well-characterized visual deficits, such as cataract-like blurring, in the early training phase of a standard deep neural network causes irreversible performance loss that closely mimics that reported in humans and animal models. Deficits that do not affect low-level image statistics, such as vertical flipping of the images, have no lasting effect on the ANN's performance and can be rapidly overcome with additional training, as observed in humans. In addition, deeper networks show a more prominent critical period. To better understand this phenomenon, we use techniques from information theory to study the strength of the network connections during training. Our analysis suggests that the first few epochs are critical for the allocation of resources across different layers, determined by the initial input data distribution. Once such information organization is established, the network resources do not re-distribute through additional training. These findings suggest that the initial rapid learning phase of training of ANNs, under-scrutinized compared to its asymptotic behavior, plays a key role in defining the final performance of networks.

Real-time crime forecasting is important. However, accurate prediction of when and where the next crime will happen is difficult. No known physical model provides a reasonable approximation to such a complex system. Historical crime data are sparse in both space and time and the signal of interests is weak. In this work, we first present a proper representation of crime data. We then adapt the spatial temporal residual network on the well represented data to predict the distribution of crime in Los Angeles at the scale of hours in neighborhood-sized parcels. These experiments as well as comparisons with several existing approaches to prediction demonstrate the superiority of the proposed model in terms of accuracy. Finally, we present a ternarization technique to address the resource consumption issue for its deployment in real world. This work is an extension of our short conference proceeding paper [Wang et al, Arxiv 1707.03340]. Keywords: Crime representation, Spatial-temporal deep learning, Real-time forecasting, Ternarization. 1 Introduction Forecasting crime at hourly or even finer temporal scales in micro-geographic regions is an important scientific and practical problem. Anticipating where and when crime is most likely to occur creates novel opportunities to prevent crime.

In this paper we describe the problem of painter classification, and propose a novel approach based on deep convolutional autoencoder neural networks. While previous approaches relied on image processing and manual feature extraction from paintings, our approach operates on the raw pixel level, without any preprocessing or manual feature extraction. We first train a deep convolutional autoencoder on a dataset of paintings, and subsequently use it to initialize a supervised convolutional neural network for the classification phase. The proposed approach substantially outperforms previous methods, improving the previous state-of-the-art for the 3-painter classification problem from 90.44% accuracy (previous state-of-the-art) to 96.52% accuracy, i.e., a 63% reduction in error rate.

This paper introduces the first deep neural network-based estimation metric for the jigsaw puzzle problem. Given two puzzle piece edges, the neural network predicts whether or not they should be adjacent in the correct assembly of the puzzle, using nothing but the pixels of each piece. The proposed metric exhibits an extremely high precision even though no manual feature extraction is performed. When incorporated into an existing puzzle solver, the solution's accuracy increases significantly, achieving thereby a new state-of-the-art standard.

This paper presents a novel deep learning based method for automatic malware signature generation and classification. The method uses a deep belief network (DBN), implemented with a deep stack of denoising autoencoders, generating an invariant compact representation of the malware behavior. While conventional signature and token based methods for malware detection do not detect a majority of new variants for existing malware, the results presented in this paper show that signatures generated by the DBN allow for an accurate classification of new malware variants. Using a dataset containing hundreds of variants for several major malware families, our method achieves 98.6% classification accuracy using the signatures generated by the DBN. The presented method is completely agnostic to the type of malware behavior that is logged (e.g., API calls and their parameters, registry entries, websites and ports accessed, etc.), and can use any raw input from a sandbox to successfully train the deep neural network which is used to generate malware signatures.