Frequent mine disasters cause a large number of casualties and property losses. Autonomous driving is a fundamental measure for solving this problem, and track detection is one of the key technologies for computer vision to achieve downhole automatic driving. The track detection result based on the traditional convolutional neural network (CNN) algorithm lacks the detailed and unique description of the object and relies too much on visual postprocessing technology. Therefore, this paper proposes a track detection algorithm based on a multiscale conditional generative adversarial network (CGAN). The generator is decomposed into global and local parts using a multigranularity structure in the generator network. A multiscale shared convolution structure is adopted in the discriminator network to further supervise training the generator. Finally, the Monte Carlo search technique is introduced to search the intermediate state of the generator, and the result is sent to the discriminator for comparison. Compared with the existing work, our model achieved 82.43\% pixel accuracy and an average intersection-over-union (IOU) of 0.6218, and the detection of the track reached 95.01\% accuracy in the downhole roadway scene test set.

In this study, we present an analysis of model-based ensemble learning for 3D point-cloud object classification and detection. An ensemble of multiple model instances is known to outperform a single model instance, but there is little study of the topic of ensemble learning for 3D point clouds. First, an ensemble of multiple model instances trained on the same part of the $\textit{ModelNet40}$ dataset was tested for seven deep learning, point cloud-based classification algorithms: $\textit{PointNet}$, $\textit{PointNet++}$, $\textit{SO-Net}$, $\textit{KCNet}$, $\textit{DeepSets}$, $\textit{DGCNN}$, and $\textit{PointCNN}$. Second, the ensemble of different architectures was tested. Results of our experiments show that the tested ensemble learning methods improve over state-of-the-art on the $\textit{ModelNet40}$ dataset, from $92.65\%$ to $93.64\%$ for the ensemble of single architecture instances, $94.03\%$ for two different architectures, and $94.15\%$ for five different architectures. We show that the ensemble of two models with different architectures can be as effective as the ensemble of 10 models with the same architecture. Third, a study on classic bagging i.e. with different subsets used for training multiple model instances) was tested and sources of ensemble accuracy growth were investigated for best-performing architecture, i.e. $\textit{SO-Net}$. We also investigate the ensemble learning of $\textit{Frustum PointNet}$ approach in the task of 3D object detection, increasing the average precision of 3D box detection on the $\textit{KITTI}$ dataset from $63.1\%$ to $66.5\%$ using only three model instances. We measure the inference time of all 3D classification architectures on a $\textit{Nvidia Jetson TX2}$, a common embedded computer for mobile robots, to allude to the use of these models in real-life applications.

This paper presents a taxonomy of explainability in Human-Agent Systems. We consider fundamental questions about the Why, Who, What, When and How of explainability. First, we define explainability, and its relationship to the related terms of interpretability, transparency, explicitness, and faithfulness. These definitions allow us to answer why explainability is needed in the system, whom it is geared to and what explanations can be generated to meet this need. We then consider when the user should be presented with this information. Last, we consider how objective and subjective measures can be used to evaluate the entire system. This last question is the most encompassing as it will need to evaluate all other issues regarding explainability.

In this paper, we address the question answering challenge with the SQuAD 2.0 dataset. We design a model architecture which leverages BERT's capability of context-aware word embeddings and BiDAF's context interactive exploration mechanism. By integrating these two state-of-the-art architectures, our system tries to extract the contextual word representation at word and character levels, for better comprehension of both question and context and their correlations. We also propose our original joint posterior probability predictor module and its associated loss functions. Our best model so far obtains F1 score of 75.842% and EM score of 72.24% on the test PCE leaderboad.

Technology is a transformative force, and there is currently no technology more transformative than AI. Because AI encompasses so many different fields -- from robotics to deep learning to self-driving cars -- it can be hard to quantify the exact impact it has had on the economy and on individual industries. Whether you're heavily involved within the industry or an outside observer, by examining the different use cases and successes (or failures) that AI has been a part of, you can gain a better understanding of how to implement AI within your own organization, and how to generate the most value from it. Given the lucrative opportunities available, it is unsurprising that both established players in the health care industry (i.e., hospitals, pharmaceutical companies, insurers, etc.) and startups are looking to utilize AI to maximize their impact. A 2018 report revealed that health care-focused AI startups alone have raised $4.3 billion since 2013, more than startups in any other industry.

Humans can only interact with part of the surrounding environment due to biological restrictions. Therefore, we learn to reason the spatial relationships across a series of observations to piece together the surrounding environment. Inspired by such behavior and the fact that machines also have computational constraints, we propose \underline{CO}nditional \underline{CO}ordinate GAN (COCO-GAN) of which the generator generates images by parts based on their spatial coordinates as the condition. On the other hand, the discriminator learns to justify realism across multiple assembled patches by global coherence, local appearance, and edge-crossing continuity. Despite the full images are never generated during training, we show that COCO-GAN can produce \textbf{state-of-the-art-quality} full images during inference. We further demonstrate a variety of novel applications enabled by teaching the network to be aware of coordinates. First, we perform extrapolation to the learned coordinate manifold and generate off-the-boundary patches. Combining with the originally generated full image, COCO-GAN can produce images that are larger than training samples, which we called "beyond-boundary generation". We then showcase panorama generation within a cylindrical coordinate system that inherently preserves horizontally cyclic topology. On the computation side, COCO-GAN has a built-in divide-and-conquer paradigm that reduces memory requisition during training and inference, provides high-parallelism, and can generate parts of images on-demand.

In this paper, we study the problem of node representation learning with graph neural networks. We present a graph neural network class named recurrent graph neural network (RGNN), that address the shortcomings of prior methods. By using recurrent units to capture the long-term dependency across layers, our methods can successfully identify important information during recursive neighborhood expansion. In our experiments, we show that our model class achieves state-of-the-art results on three benchmarks: the Pubmed, Reddit, and PPI network datasets. Our in-depth analyses also demonstrate that incorporating recurrent units is a simple yet effective method to prevent noisy information in graphs, which enables a deeper graph neural network.

Current algorithms for deep learning probably cannot run in the brain because they rely on weight transport, where forward-path neurons transmit their synaptic weights to a feedback path, in a way that is likely impossible biologically. An algorithm called feedback alignment achieves deep learning without weight transport by using random feedback weights, but it performs poorly on hard visual-recognition tasks. Here we describe two mechanisms -- a neural circuit called a weight mirror and a version of an algorithm proposed by Kolen and Pollack in 1994 -- both of which let the feedback path learn appropriate synaptic weights quickly and accurately even in large networks, without weight transport or complex wiring. Tested on the ImageNet visual-recognition task, these mechanisms outperform both feedback alignment and the newer sign-symmetry method, and nearly match backprop, the standard algorithm of deep learning, which uses weight transport.

Speech separation has been very successful with deep learning techniques. Substantial effort has been reported based on approaches over spectrogram, which is well known as the standard time-and-frequency cross-domain representation for speech signals. It is highly correlated to the phonetic structure of speech, or "how the speech sounds" when perceived by human, but primarily frequency domain features carrying temporal behaviour. Very impressive work achieving speech separation over time domain was reported recently, probably because waveforms in time domain may describe the different realizations of speech in a more precise way than spectrogram. In this paper, we propose a framework properly integrating the above two directions, hoping to achieve both purposes. We construct a time-and-frequency feature map by concatenating the 1-dim convolution encoded feature map (for time domain) and the spectrogram (for frequency domain), which was then processed by an embedding network and clustering approaches very similar to those used in time and frequency domain prior works. In this way, the information in the time and frequency domains, as well as the interactions between them, can be jointly considered during embedding and clustering. Very encouraging results (state-of-the-art to our knowledge) were obtained with WSJ0-2mix dataset in preliminary experiments.

Deep learning algorithms have increasingly been shown to lack robustness to simple adversarial examples (AdvX). An equally troubling observation is that these adversarial examples transfer between different architectures trained on different datasets. We investigate the transferability of adversarial examples between models using the angle between the input-output Jacobians of different models. To demonstrate the relevance of this approach, we perform case studies that involve jointly training pairs of models. These case studies empirically justify the theoretical intuitions for why the angle between gradients is a fundamental quantity in AdvX transferability. Furthermore, we consider the asymmetry of AdvX transferability between two models of the same architecture and explain it in terms of differences in gradient norms between the models. Lastly, we provide a simple modification to existing training setups that reduces transferability of adversarial examples between pairs of models.