Autonomous driving in multi-agent and dynamic traffic scenarios is challenging, where the behaviors of other road agents are uncertain and hard to model explicitly, and the ego-vehicle should apply complicated negotiation skills with them to achieve both safe and efficient driving in various settings, such as giving way, merging and taking turns. Traditional planning methods are largely rule-based and scale poorly in these complex dynamic scenarios, often leading to reactive or even overly conservative behaviors. Therefore, they require tedious human efforts to maintain workability. Recently, deep learning-based methods have shown promising results with better generalization capability but less hand engineering effort. However, they are either implemented with supervised imitation learning (IL) that suffers from the dataset bias and distribution mismatch problems, or trained with deep reinforcement learning (DRL) but focus on one specific traffic scenario. In this work, we propose DQ-GAT to achieve scalable and proactive autonomous driving, where graph attention-based networks are used to implicitly model interactions, and asynchronous deep Q-learning is employed to train the network end-to-end in an unsupervised manner. Extensive experiments through a high-fidelity driving simulation show that our method can better trade-off safety and efficiency in both seen and unseen scenarios, achieving higher goal success rates than the baselines (at most 4.7$\times$) with comparable task completion time. Demonstration videos are available at https://caipeide.github.io/dq-gat/.

Rapid progress in natural language processing has led to its utilization in a variety of industrial and enterprise settings, including in its use for information extraction, specifically named entity recognition and relation extraction, from documents such as engineering manuals and field maintenance reports. While named entity recognition is a well-studied problem, existing state-of-the-art approaches require large labelled datasets which are hard to acquire for sensitive data such as maintenance records. Further, industrial domain experts tend to distrust results from black box machine learning models, especially when the extracted information is used in downstream predictive maintenance analytics. We overcome these challenges by developing three approaches built on the foundation of domain expert knowledge captured in dictionaries and ontologies. We develop a syntactic and semantic rules-based approach and an approach leveraging a pre-trained language model, fine-tuned for a question-answering task on top of our base dictionary lookup to extract entities of interest from maintenance records. We also develop a preliminary ontology to represent and capture the semantics of maintenance records. Our evaluations on a real-world aviation maintenance records dataset show promising results and help identify challenges specific to named entity recognition in the context of noisy industrial data.

Medical images, especially volumetric images, are of high resolution and often exceed the capacity of standard desktop GPUs. As a result, most deep learning-based medical image analysis tasks require the input images to be downsampled, often substantially, before these can be fed to a neural network. However, downsampling can lead to a loss of image quality, which is undesirable especially in reconstruction tasks, where the fine geometric details need to be preserved. In this paper, we propose that high-resolution images can be reconstructed in a coarse-to-fine fashion, where a deep learning algorithm is only responsible for generating a coarse representation of the image, which consumes moderate GPU memory. For producing the high-resolution outcome, we propose two novel methods: learned voxel rearrangement of the coarse output and hierarchical image synthesis. Compared to the coarse output, the high-resolution counterpart allows for smooth surface triangulation, which can be 3D-printed in the highest possible quality. Experiments of this paper are carried out on the dataset of AutoImplant 2021 (https://autoimplant2021.grand-challenge.org/), a MICCAI challenge on cranial implant design. The dataset contains high-resolution skulls that can be viewed as 2D manifolds embedded in a 3D space. Codes associated with this study can be accessed at https://github.com/Jianningli/voxel_rearrangement.

Development in the field of Single Board Computers (SBC) have been increasing for several years. They provide a good balance between computing performance and power consumption which is usually required for mobile platforms, like application in vehicles for Advanced Driver Assistance Systems (ADAS) and Autonomous Driving (AD). However, there is an ever-increasing need of more powerful and efficient SBCs which can run power intensive Deep Neural Networks (DNNs) in real-time and can also satisfy necessary functional safety requirements such as Automotive Safety Integrity Level (ASIL). ProAI is being developed by ZF mainly to run powerful and efficient applications such as multitask DNNs and on top of that it also has the required safety certification for AD. In this work, we compare and discuss state of the art SBC on the basis of power intensive multitask DNN architecture called Multitask-CenterNet with respect to performance measures such as, FPS and power efficiency. As an automotive supercomputer, ProAI delivers an excellent combination of performance and efficiency, managing nearly twice the number of FPS per watt than a modern workstation laptop and almost four times compared to the Jetson Nano. Furthermore, it was also shown that there is still power in reserve for further and more complex tasks on the ProAI, based on the CPU and GPU utilization during the benchmark.

Video question answering has recently received a lot of attention from multimodal video researchers. Most video question answering datasets are usually in the form of multiple-choice. But, the model for the multiple-choice task does not infer the answer. Rather it compares the answer candidates for picking the correct answer. Furthermore, it makes it difficult to extend to other tasks. In this paper, we challenge the existing multiple-choice video question answering by changing it to open-ended video question answering. To tackle open-ended question answering, we use the pretrained GPT2 model. The model is fine-tuned with video inputs and subtitles. An ablation study is performed by changing the existing DramaQA dataset to an open-ended question answering, and it shows that performance can be improved using video metadata.

Multimodal abstractive summarization with sentence output is to generate a textual summary given a multimodal triad -- sentence, image and audio, which has been proven to improve users satisfaction and convenient our life. Existing approaches mainly focus on the enhancement of multimodal fusion, while ignoring the unalignment among multiple inputs and the emphasis of different segments in feature, which has resulted in the superfluity of multimodal interaction. To alleviate these problems, we propose a Multimodal Hierarchical Selective Transformer (mhsf) model that considers reciprocal relationships among modalities (by low-level cross-modal interaction module) and respective characteristics within single fusion feature (by high-level selective routing module). In details, it firstly aligns the inputs from different sources and then adopts a divide and conquer strategy to highlight or de-emphasize multimodal fusion representation, which can be seen as a sparsely feed-forward model - different groups of parameters will be activated facing different segments in feature. We evaluate the generalism of proposed mhsf model with the pre-trained+fine-tuning and fresh training strategies. And Further experimental results on MSMO demonstrate that our model outperforms SOTA baselines in terms of ROUGE, relevance scores and human evaluation.

Uncertainties in Deep Neural Network (DNN)-based perception and vehicle's motion pose challenges to the development of safe autonomous driving vehicles. In this paper, we propose a safe motion planning framework featuring the quantification and propagation of DNN-based perception uncertainties and motion uncertainties. Contributions of this work are twofold: (1) A Bayesian Deep Neural network model which detects 3D objects and quantitatively captures the associated aleatoric and epistemic uncertainties of DNNs; (2) An uncertainty-aware motion planning algorithm (PU-RRT) that accounts for uncertainties in object detection and ego-vehicle's motion. The proposed approaches are validated via simulated complex scenarios built in CARLA. Experimental results show that the proposed motion planning scheme can cope with uncertainties of DNN-based perception and vehicle motion, and improve the operational safety of autonomous vehicles while still achieving desirable efficiency.

Background: Cervical cancer seriously affects the health of the female reproductive system. Optical coherence tomography (OCT) emerges as a non-invasive, high-resolution imaging technology for cervical disease detection. However, OCT image annotation is knowledge-intensive and time-consuming, which impedes the training process of deep-learning-based classification models. Objective: This study aims to develop a computer-aided diagnosis (CADx) approach to classifying in-vivo cervical OCT images based on self-supervised learning. Methods: Besides high-level semantic features extracted by a convolutional neural network (CNN), the proposed CADx approach leverages unlabeled cervical OCT images' texture features learned by contrastive texture learning. We conducted ten-fold cross-validation on the OCT image dataset from a multi-center clinical study on 733 patients from China. Results: In a binary classification task for detecting high-risk diseases, including high-grade squamous intraepithelial lesion (HSIL) and cervical cancer, our method achieved an area-under-the-curve (AUC) value of 0.9798 Plus or Minus 0.0157 with a sensitivity of 91.17 Plus or Minus 4.99% and a specificity of 93.96 Plus or Minus 4.72% for OCT image patches; also, it outperformed two out of four medical experts on the test set. Furthermore, our method achieved a 91.53% sensitivity and 97.37% specificity on an external validation dataset containing 287 3D OCT volumes from 118 Chinese patients in a new hospital using a cross-shaped threshold voting strategy. Conclusion: The proposed contrastive-learning-based CADx method outperformed the end-to-end CNN models and provided better interpretability based on texture features, which holds great potential to be used in the clinical protocol of "see-and-treat."

In the era of intelligent transportation, driver behavior profiling has become a beneficial technology as it provides knowledge regarding the driver's aggressiveness. Previous approaches achieved promising driver behavior profiling performance through establishing statistical heuristics rules or supervised learning-based models. Still, there exist limits that the practitioner should prepare a labeled dataset, and prior approaches could not classify aggressive behaviors which are not known a priori. In pursuit of improving the aforementioned drawbacks, we propose a novel approach to driver behavior profiling leveraging an unsupervised learning paradigm. First, we cast the driver behavior profiling problem as anomaly detection. Second, we established recurrent neural networks that predict the next feature vector given a sequence of feature vectors. We trained the model with normal driver data only. As a result, our model yields high regression error given a sequence of aggressive driver behavior and low error given at a sequence of normal driver behavior. We figured this difference of error between normal and aggressive driver behavior can be an adequate flag for driver behavior profiling and accomplished a precise performance in experiments. Lastly, we further analyzed the optimal level of sequence length for identifying each aggressive driver behavior. We expect the proposed approach to be a useful baseline for unsupervised driver behavior profiling and contribute to the efficient, intelligent transportation ecosystem.

Adversarial patch attack against image classification deep neural networks (DNNs), in which the attacker can inject arbitrary distortions within a bounded region of an image, is able to generate adversarial perturbations that are robust (i.e., remain adversarial in physical world) and universal (i.e., remain adversarial on any input). It is thus important to detect and mitigate such attack to ensure the security of DNNs. This work proposes Jujutsu, a technique to detect and mitigate robust and universal adversarial patch attack. Jujutsu leverages the universal property of the patch attack for detection. It uses explainable AI technique to identify suspicious features that are potentially malicious, and verify their maliciousness by transplanting the suspicious features to new images. An adversarial patch continues to exhibit the malicious behavior on the new images and thus can be detected based on prediction consistency. Jujutsu leverages the localized nature of the patch attack for mitigation, by randomly masking the suspicious features to "remove" adversarial perturbations. However, the network might fail to classify the images as some of the contents are removed (masked). Therefore, Jujutsu uses image inpainting for synthesizing alternative contents from the pixels that are masked, which can reconstruct the "clean" image for correct prediction. We evaluate Jujutsu on five DNNs on two datasets, and show that Jujutsu achieves superior performance and significantly outperforms existing techniques. Jujutsu can further defend against various variants of the basic attack, including 1) physical-world attack; 2) attacks that target diverse classes; 3) attacks that use patches in different shapes and 4) adaptive attacks.