In this gridworld, the agent must navigate a'warehouse' to reach the green goal tile via one of two routes. It can head straight down the narrow corridor, where it has to pass a pink tile that interrupts the agent 50% of the time, meaning it will be stuck until the end of the episode. Or it can step on the purple button, which disables the pink tile and prevents any possibility of interruption but at the cost of a longer path. In this scenario, we always want agents to pass the pink tile, risking interruption, rather than learn to use the purple button. Our irreversible side effects environment tests whether an agent will change its behaviour to avoid inadvertent and irreversible consequences. For example, if a robot is asked to put a vase of flowers on a table, we want it to do so without breaking the vase or spilling the water.

This is one in a series of periodic guest columns by industry thought leaders. Autonomous technology is set to transform the motor industry, but there are no clear paths for manufacturers. As the autonomous mobility industry takes shape, artificial intelligence could play a much bigger role. The Society of Automotive Engineers uses a classification system of 5 levels of vehicle autonomy based on the amount of necessary driver intervention. Level 1, for example, requires a driver to be in control at all times but allows for automated acceleration and braking.

In this blog post we demonstrate how to run distributed offline inference on large datasets using Apache MXNet (incubating) and Apache Spark on Amazon EMR. We explain how offline inference is useful, why it is challenging, and how you can leverage MXNet and Spark on Amazon EMR to overcome these challenges. After a deep learning model has been trained, it's put to work by running inference on new data. Inference can be executed in real-time for tasks that require immediate feedback, such as fraud detection. This is typically known as online inference.

The problem descriptions are taken straightaway from the assignments. In this assignment a linear classifier will be implemented and it will be trained using stochastic gradient descent with numpy. To make things more intuitive, let's solve a 2D classification problem with synthetic data. As we can notice the data above isn't linearly separable. Hence we should add features(or use non-linear model). Note that decision line between two classes have form of circle, since that we can add quadratic features to make the problem linearly separable.

This paper proposes a principled information theoretic analysis of classification for deep neural network structures, e.g. convolutional neural networks (CNN). The output of convolutional filters is modeled as a random variable Y conditioned on the object class C and network filter bank F. The conditional entropy (CENT) H(Y |C,F) is shown in theory and experiments to be a highly compact and class-informative code, that can be computed from the filter outputs throughout an existing CNN and used to obtain higher classification results than the original CNN itself. Experiments demonstrate the effectiveness of CENT feature analysis in two separate CNN classification contexts. 1) In the classification of neurodegeneration due to Alzheimer's disease (AD) and natural aging from 3D magnetic resonance image (MRI) volumes, 3 CENT features result in an AUC=94.6% for whole-brain AD classification, the highest reported accuracy on the public OASIS dataset used and 12% higher than the softmax output of the original CNN trained for the task. 2) In the context of visual object classification from 2D photographs, transfer learning based on a small set of CENT features identified throughout an existing CNN leads to AUC values comparable to the 1000-feature softmax output of the original network when classifying previously unseen object categories. The general information theoretical analysis explains various recent CNN design successes, e.g. densely connected CNN architectures, and provides insights for future research directions in deep learning.

This paper proposes a real-time embedded fall detection system using a DVS(Dynamic Vision Sensor)(Berner et al. [2014]) that has never been used for traditional fall detection, a dataset for fall detection using that, and a DVS-TN(DVS-Temporal Network). The first contribution is building a DVS Falls Dataset, which made our network to recognize a much greater variety of falls than the existing datasets that existed before and solved privacy issues using the DVS. Secondly, we introduce the DVS-TN: optimized deep learning network to detect falls using DVS. Finally, we implemented a fall detection system which can run on low-computing H/W with real-time, and tested on DVS Falls Dataset that takes into account various falls situations. Our approach achieved 95.5% on the F1-score and operates at 31.25 FPS on NVIDIA Jetson TX1 board.

Long Short-Term Memory (LSTM) is one of the most powerful sequence models. Despite the strong performance, however, it lacks the nice interpretability as in state space models. In this paper, we present a way to combine the best of both worlds by introducing State Space LSTM (SSL) models that generalizes the earlier work \cite{zaheer2017latent} of combining topic models with LSTM. However, unlike \cite{zaheer2017latent}, we do not make any factorization assumptions in our inference algorithm. We present an efficient sampler based on sequential Monte Carlo (SMC) method that draws from the joint posterior directly. Experimental results confirms the superiority and stability of this SMC inference algorithm on a variety of domains.

This paper develops a novel methodology for using symbolic knowledge in deep learning. From first principles, we derive a semantic loss function that bridges between neural output vectors and logical constraints. This loss function captures how close the neural network is to satisfying the constraints on its output. An experimental evaluation shows that our semantic loss function effectively guides the learner to achieve (near-)state-of-the-art results on semi-supervised multi-class classification. Moreover, it significantly increases the ability of the neural network to predict structured objects, such as rankings and paths. These discrete concepts are tremendously difficult to learn, and benefit from a tight integration of deep learning and symbolic reasoning methods.

We propose stochastic, non-parametric activation functions that are fully learnable and individual to each neuron. Complexity and the risk of overfitting are controlled by placing a Gaussian process prior over these functions. The result is the Gaussian process neuron, a probabilistic unit that can be used as the basic building block for probabilistic graphical models that resemble the structure of neural networks. The proposed model can intrinsically handle uncertainties in its inputs and self-estimate the confidence of its predictions. Using variational Bayesian inference and the central limit theorem, a fully deterministic loss function is derived, allowing it to be trained as efficiently as a conventional neural network using mini-batch gradient descent. The posterior distribution of activation functions is inferred from the training data alongside the weights of the network. The proposed model favorably compares to deep Gaussian processes, both in model complexity and efficiency of inference. It can be directly applied to recurrent or convolutional network structures, allowing its use in audio and image processing tasks. As an preliminary empirical evaluation we present experiments on regression and classification tasks, in which our model achieves performance comparable to or better than a Dropout regularized neural network with a fixed activation function. Experiments are ongoing and results will be added as they become available.

We propose a framework for general probabilistic multi-step time series regression. Specifically, we exploit the expressiveness and temporal nature of Recurrent Neural Networks, the nonparametric nature of Quantile Regression and the efficiency of Direct Multi-Horizon Forecasting. A new training scheme for recurrent nets is designed to boost stability and performance. We show that the approach accommodates both temporal and static covariates, learning across multiple related series, shifting seasonality, future planned event spikes and cold-starts in real life large-scale forecasting. The performance of the framework is demonstrated in an application to predict the future demand of items sold on Amazon.com, and in a public probabilistic forecasting competition to predict electricity price and load.