Research Highlights (Required)To create your highlights, please type the highlights against each \item command.It should be short collection of bullet points that convey the core findings of the article. It should include 3 to 5 bullet points (maximum 85 characters, including spaces, per bullet point.)• We survey deep learning based HAR in sensor modality, deep model, and application.• We comprehensively discuss the insights of deep learning models for HAR tasks.• We extensively investigate why deep learning can improve the performance of HAR.• We also summarize the public HAR datasets frequently used for research purpose.• We present some grand challenges and feasible solutions for deep learning based HAR. ABSTRACTSensor-based activity recognition seeks the profound high-level knowledge about human activities from multitudes of low-level sensor readings. Conventional pattern recognition approaches have made tremendous progress in the past years. However, those methods often heavily rely on heuristic hand-crafted feature extraction, which could hinder their generalization performance. Additionally, existing methods are undermined for unsupervised and incremental learning tasks. Recently, the recent advancement of deep learning makes it possible to perform automatic high-level feature extraction thus achieves promising performance in many areas. Since then, deep learning based methods have been widely adopted for the sensor-based activity recognition tasks. This paper surveys the recent advance of deep learning based sensor-based activity recognition. We summarize existing literature from three aspects: sensor modality, deep model, and application. We also present detailed insights on existing work and propose grand challenges for future research.

Cambrian sedimentary rocks in the southern part of the South China Craton were derived from a source that lay to the south or southeast, beyond the current limits of the craton and which is no longer preserved nearby. U‐Pb ages and Hf isotope data on detrital zircons from the Cambrian sequence define two distinctive age peaks at 1120 Ma and 960 Ma, with εHf(t) values for each group identical to the coeval detrital zircons from Western Australia and the Tethyan Himalaya zone, respectively. The circa 1120 Ma detrital zircons were most likely derived from the Wilkes‐Albany‐Fraser belt between southwest Australia and Antarctica, whereas the circa 960 Ma detrital zircons could have been sourced from the Rayner‐Eastern Ghats belt between India and Antarctica. Derivation of detritus from these sources suggests that south China was located at the nexus between India, Antarctica, and Australia, along the northern margin of East Gondwana during the Cambrian.

In activity recognition, it is often expensive and time-consuming to acquire sufficient activity labels. To solve this problem, transfer learning leverages the labeled samples from the source domain to annotate the target domain which has few or none labels. Existing approaches typically consider learning a global domain shift while ignoring the intra-affinity between classes, which will hinder the performance of the algorithms. In this paper, we propose a novel and general cross-domain learning framework that can exploit the intra-affinity of classes to perform intra-class knowledge transfer. The proposed framework, referred to as Stratified Transfer Learning (STL), can dramatically improve the classification accuracy for cross-domain activity recognition. Specifically, STL first obtains pseudo labels for the target domain via majority voting technique. Then, it performs intra-class knowledge transfer iteratively to transform both domains into the same subspaces. Finally, the labels of target domain are obtained via the second annotation. To evaluate the performance of STL, we conduct comprehensive experiments on three large public activity recognition datasets (i.e. OPPORTU-NITY, PAMAP2, and UCI DSADS), which demonstrates that STL significantly outperforms other state-of-the-art methods w.r.t. classification accuracy (improvement of 7.68%). Furthermore, we extensively investigate the performance of STL across different degrees of similarities and activity levels between domains. And we also discuss the potential of STL in other pervasive computing applications to provide empirical experience for future research.

No abstract

Northeast trending Yong'an Basin, southeast South China Craton, preserves a Permian-Jurassic, marine to continental, siliciclastic-dominated, retroarc foreland basin succession. Modal and detrital zircon data, along with published paleocurrent data, sedimentary facies, and euhedral to subhedral detrital zircon shapes, indicate derivation from multicomponent, nearby sources with input from both the interior of the craton to the northwest and from an inferred arc accretionary complex to the southeast. The detrital zircon U-Pb age spectra range in age from Archean to early Mesozoic, with major age groups at 2000-1700 Ma, 1200-900 Ma, 400-340 Ma, and 300-240 Ma. In addition, Early Jurassic strata include zircon detritus with ages of 200-170 Ma. Regional geological relations suggest that Precambrian and Early Paleozoic detritus was derived from the inland Wuyi Mountain region and Yunkai Massif of the South China Craton. Sources for Middle Paleozoic to early Mesozoic detrital zircons include input from beyond the currently exposed China mainland. Paleogeographic reconstruction in East Asia suggests derivation from an active convergent plate margin along the southeastern rim of the craton that incorporated part of Southwest Japan and is related to the subduction of the Paleo-Pacific Ocean. Integration of the geologic and provenance records of the Yong'an Basin with the time equivalent Yongjiang and Shiwandashan basins that lie to the southwest and south, respectively, provides an integrated record of the subduction of the Paleo-Pacific Ocean along the southeast margin of the South China Craton and termination of subduction of the Paleo-Tethys beneath its southwest margin in Permo-Triassic.