Robots hold promise in many scenarios involving outdoor use, such as search-and-rescue, wildlife management, and collecting data to improve environment, climate, and weather forecasting. However, autonomous navigation of outdoor trails remains a challenging problem. Recent work has sought to address this issue using deep learning. Although this approach has achieved state-of-the-art results, the deep learning paradigm may be limited due to a reliance on large amounts of annotated training data. Collecting and curating training datasets may not be feasible or practical in many situations, especially as trail conditions may change due to seasonal weather variations, storms, and natural erosion. In this paper, we explore an approach to address this issue through virtual-to-real-world transfer learning using a variety of deep learning models trained to classify the direction of a trail in an image. Our approach utilizes synthetic data gathered from virtual environments for model training, bypassing the need to collect a large amount of real images of the outdoors. We validate our approach in three main ways. First, we demonstrate that our models achieve classification accuracies upwards of 95% on our synthetic data set. Next, we utilize our classification models in the control system of a simulated robot to demonstrate feasibility. Finally, we evaluate our models on real-world trail data and demonstrate the potential of virtual-to-real-world transfer learning.

Abstract-- We applied predefined kernels also known as filters or masks developed for image processing to convolution neural network. Instead of letting neural networks find its own kernels, we used 41 different general-purpose kernels of blurring, edge detecting, sharpening, discrete cosine transformation, etc. for the first layer of the convolution neural networks. This architecture, thus named as general filter convolutional neural network (GFNN), can reduce training time by 30% with a better accuracy compared to the regular convolutional neural network (CNN). GFNN also can be trained to achieve 90% accuracy with only 500 samples. Furthermore, even though these kernels are not specialized for the MNIST dataset, we achieved 99.56% accuracy without ensemble nor any other special algorithms.

Application of neural networks to a vast variety of practical applications is transforming the way AI is applied in practice. Pre-trained neural network models available through APIs or capability to custom train pre-built neural network architectures with customer data has made the consumption of AI by developers much simpler and resulted in broad adoption of these complex AI models. While prebuilt network models exist for certain scenarios, to try and meet the constraints that are unique to each application, AI teams need to think about developing custom neural network architectures that can meet the tradeoff between accuracy and memory footprint to achieve the tight constraints of their unique use-cases. However, only a small proportion of data science teams have the skills and experience needed to create a neural network from scratch, and the demand far exceeds the supply. In this paper, we present NeuNetS : An automated Neural Network Synthesis engine for custom neural network design that is available as part of IBM's AI OpenScale's product. NeuNetS is available for both Text and Image domains and can build neural networks for specific tasks in a fraction of the time it takes today with human effort, and with accuracy similar to that of human-designed AI models.

Recently, deep learning has transformed many fields including medical imaging. Inspired by diversity of biological neurons, our group proposed quadratic neurons in which the inner product in current artificial neurons is replaced with a quadratic operation on inputs, thereby enhancing the capability of an individual neuron. Along this direction, we are motivated to evaluate the power of quadratic neurons in representative network architectures, towards quadratic neuron based deep learning. In this regard, our prior theoretical studies have shown important merits of quadratic neurons and networks. In this paper, we use quadratic neurons to construct an encoder-decoder structure, referred to as the quadratic auto-encoder, and apply it for low-dose CT de-noising. Then, we perform experiments on the Mayo low-dose CT dataset to demonstrate that the quadratic auto-encoder yields a better de-noising performance.

This paper proposes CodeX, an end-to-end framework that facilitates encoding, bitwidth customization, fine-tuning, and implementation of neural networks on FPGA platforms. CodeX incorporates nonlinear encoding to the computation flow of neural networks to save memory. The encoded features demand significantly lower storage compared to the raw full-precision activation values; therefore, the execution flow of CodeX hardware engine is completely performed within the FPGA using on-chip streaming buffers with no access to the off-chip DRAM. We further propose a fully-automated algorithm inspired by reinforcement learning which determines the customized encoding bitwidth across network layers. CodeX full-stack framework comprises of a compiler which takes a high-level Python description of an arbitrary neural network architecture. The compiler then instantiates the corresponding elements from CodeX Hardware library for FPGA implementation. Proof-of-concept evaluations on MNIST, SVHN, and CIFAR-10 datasets demonstrate an average of 4.65x throughput improvement compared to stand-alone weight encoding. We further compare CodeX with six existing full-precision DNN accelerators on ImageNet, showing an average of 3.6x and 2.54x improvement in throughput and performance-per-watt, respectively.

In order to better engage with customers, retailers rely on extensive customer and product databases which allows them to better understand customer behaviour and purchasing patterns. This has long been a challenging task as customer modelling is a multi-faceted, noisy and time-dependent problem. The most common way to tackle this problem is indirectly through task-specific supervised learning prediction problems, with relatively little literature on modelling a customer by directly simulating their future transactions. In this paper we propose a method for generating realistic sequences of baskets that a given customer is likely to purchase over a period of time. Customer embedding representations are learned using a Recurrent Neural Network (RNN) which takes into account the entire sequence of transaction data. Given the customer state at a specific point in time, a Generative Adversarial Network (GAN) is trained to generate a plausible basket of products for the following week. The newly generated basket is then fed back into the RNN to update the customer's state. The GAN is thus used in tandem with the RNN module in a pipeline alternating between basket generation and customer state updating steps. This allows for sampling over a distribution of a customer's future sequence of baskets, which then can be used to gain insight into how to service the customer more effectively. The methodology is empirically shown to produce baskets that appear similar to real baskets and enjoy many common properties, including frequencies of different product types, brands, and prices. Furthermore, the generated data is able to replicate most of the strongest sequential patterns that exist between product types in the real data.

Abstract: Machine learning competitions such as those organized by Kaggle or KDD represent a useful benchmark for data science research. In this work, we present our winning solution to the Game Data Mining competition hosted at the 2017 IEEE Conference on Computational Intelligence and Games (CIG 2017). The contest consisted of two tracks, and participants (more than 250, belonging to both industry and academia) were to predict which players would stop playing the game, as well as their remaining lifetime. The data were provided by a major worldwide video game company, NCSoft, and came from their successful massively multiplayer online game Blade and Soul. Here, we describe the long short-term memory approach and conditional inference survival ensemble model that made us win both tracks of the contest, as well as the validation procedure that we followed in order to prevent overfitting. In particular, choosing a survival method able to deal with censored data was crucial to accurately predict the moment in which each player would leave the game, as censoring is inherent in churn. The selected models proved to be robust against evolving conditions--since there was a change in the business model of the game (from subscription-based to free-to-play) between the two sample datasets provided--and efficient in terms of time cost. Thanks to these features and also to their ability to scale to large datasets, our models could be readily implemented in real business settings. Keywords: churn; competition; video games; user behavior; behavioral data 1.

Domain adaptation has become a prominent problem setting in machine learning and related fields. This review asks the questions: when and how a classifier can learn from a source domain and generalize to a target domain. As for when, we review conditions that allow for cross-domain generalization error bounds. As for how, we present a categorization of approaches, divided into, what we refer to as, sample-based, feature-based and inference-based methods. Sample-based methods focus on weighting individual observations during training based on their importance to the target domain. Feature-based methods focus on mapping, projecting and representing features such that a source classifier performs well on the target domain and inference-based methods focus on alternative estimators, such as robust, minimax or Bayesian. Our categorization highlights recurring ideas and raises a number of questions important to further research.

Deep Reinforcement Learning (DRL) has become increasingly powerful in recent years, with notable achievements such as Deepmind's AlphaGo. It has been successfully deployed in commercial vehicles like Mobileye's path planning system. However, a vast majority of work on DRL is focused on toy examples in controlled synthetic car simulator environments such as TORCS and CARLA. In general, DRL is still at its infancy in terms of usability in real-world applications. Our goal in this paper is to encourage real-world deployment of DRL in various autonomous driving (AD) applications. We first provide an overview of the tasks in autonomous driving systems, reinforcement learning algorithms and applications of DRL to AD systems. We then discuss the challenges which must be addressed to enable further progress towards real-world deployment.

We propose an unsupervised variational model for disentangling video into independent factors, i.e. each factor's future can be predicted from its past without considering the others. We show that our approach often learns factors which are interpretable as objects in a scene.