The field of question answering (QA) has seen rapid growth in new tasks and modeling approaches in recent years. Large scale datasets and focus on challenging linguistic phenomena have driven development in neural models, some of which have achieved parity with human performance in limited cases. However, an examination of state-of-the-art model output reveals that a gap remains in reasoning ability compared to a human, and performance tends to degrade when models are exposed to less-constrained tasks. We are interested in more clearly defining the strengths and limitations of leading models across diverse QA challenges, intending to help future researchers with identifying pathways to generalizable performance. We conduct extensive qualitative and quantitative analyses on the results of four models across four datasets and relate common errors to model capabilities. We also illustrate limitations in the datasets we examine and discuss a way forward for achieving generalizable models and datasets that broadly test QA capabilities.
Autonomous railway inspection with unmanned aerial vehicles (UAVs) has huge advantages over traditional inspection methods. As a prerequisite for UAV‐based autonomous following of railway lines, it is quite essential to develop intelligent railway track detection algorithms. However, there are no existing algorithms currently that can efficiently adapt to the demand for the various forms and changing inclination angles of railway tracks in the UAV aerial images. To address the challenge, this paper proposes a novel anchor‐adaptive railway track detection network (ARTNet), which constructs a dual‐branch architecture based on projection length discrimination to realize full‐angle railway track detection for the UAV aerial images taken from arbitrary viewing angles. Considering the potential capacity imbalance of the two branches that can be caused by the uneven distribution of railway tracks in the dataset, a balanced transpose co‐training strategy is proposed to train the two branches coordinately. Moreover, an extra customized transposed consistency loss is designed to guide the training of the network without increasing any computational complexity. A set of experiments have been conducted to verify the feasibility and superiority of the ARTNet. It is demonstrated that our approach can effectively realize full‐angle railway track detection and outperform other popular algorithms greatly in terms of both detection accuracy and reasoning efficiency. ARTNet can achieve a mean F1 of 76.12 and run at a speed of 50 more frames per second.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.