Social and behavioral determinants of health (SBDoH) have important roles in shaping peoples health. In clinical research studies, especially comparative effectiveness studies, failure to adjust for SBDoH factors will potentially cause confounding is
sues and misclassification errors in either statistical analyses and machine learning-based models. However, there are limited studies to examine SBDoH factors in clinical outcomes due to the lack of structured SBDoH information in current electronic health record (EHR) systems, while much of the SBDoH information is documented in clinical narratives. Natural language processing (NLP) is thus the key technology to extract such information from unstructured clinical text. However, there is not a mature clinical NLP system focusing on SBDoH. In this study, we examined two state-of-the-art transformer-based NLP models, including BERT and RoBERTa, to extract SBDoH concepts from clinical narratives, applied the best performing model to extract SBDoH concepts on a lung cancer screening patient cohort, and examined the difference of SBDoH information between NLP extracted results and structured EHRs (SBDoH information captured in standard vocabularies such as the International Classification of Diseases codes). The experimental results show that the BERT-based NLP model achieved the best strict/lenient F1-score of 0.8791 and 0.8999, respectively. The comparison between NLP extracted SBDoH information and structured EHRs in the lung cancer patient cohort of 864 patients with 161,933 various types of clinical notes showed that much more detailed information about smoking, education, and employment were only captured in clinical narratives and that it is necessary to use both clinical narratives and structured EHRs to construct a more complete picture of patients SBDoH factors.
The newly emerged transformer technology has a tremendous impact on NLP research. In the general English domain, transformer-based models have achieved state-of-the-art performances on various NLP benchmarks. In the clinical domain, researchers also
have investigated transformer models for clinical applications. The goal of this study is to systematically explore three widely used transformer-based models (i.e., BERT, RoBERTa, and XLNet) for clinical relation extraction and develop an open-source package with clinical pre-trained transformer-based models to facilitate information extraction in the clinical domain. We developed a series of clinical RE models based on three transformer architectures, namely BERT, RoBERTa, and XLNet. We evaluated these models using 2 publicly available datasets from 2018 MADE1.0 and 2018 n2c2 challenges. We compared two classification strategies (binary vs. multi-class classification) and investigated two approaches to generate candidate relations in different experimental settings. In this study, we compared three transformer-based (BERT, RoBERTa, and XLNet) models for relation extraction. We demonstrated that the RoBERTa-clinical RE model achieved the best performance on the 2018 MADE1.0 dataset with an F1-score of 0.8958. On the 2018 n2c2 dataset, the XLNet-clinical model achieved the best F1-score of 0.9610. Our results indicated that the binary classification strategy consistently outperformed the multi-class classification strategy for clinical relation extraction. Our methods and models are publicly available at https://github.com/uf-hobi-informatics-lab/ClinicalTransformerRelationExtraction. We believe this work will improve current practice on clinical relation extraction and other related NLP tasks in the biomedical domain.
Sequential recommender systems (SRS) have become a research hotspot due to its power in modeling user dynamic interests and sequential behavioral patterns. To maximize model expressive ability, a default choice is to apply a larger and deeper network
architecture, which, however, often brings high network latency when generating online recommendations. Naturally, we argue that compressing the heavy recommendation models into middle- or light- weight neural networks is of great importance for practical production systems. To realize such a goal, we propose AdaRec, a knowledge distillation (KD) framework which compresses knowledge of a teacher model into a student model adaptively according to its recommendation scene by using differentiable Neural Architecture Search (NAS). Specifically, we introduce a target-oriented distillation loss to guide the structure search process for finding the student network architecture, and a cost-sensitive loss as constraints for model size, which achieves a superior trade-off between recommendation effectiveness and efficiency. In addition, we leverage Earth Movers Distance (EMD) to realize many-to-many layer mapping during knowledge distillation, which enables each intermediate student layer to learn from other intermediate teacher layers adaptively. Extensive experiments on real-world recommendation datasets demonstrate that our model achieves competitive or better accuracy with notable inference speedup comparing to strong counterparts, while discovering diverse neural architectures for sequential recommender models under different recommendation scenes.
It is important to learn joint embedding for 3D shapes and text in different shape understanding tasks, such as shape-text matching, retrieval, and shape captioning. Current multi-view based methods learn a mapping from multiple rendered views to tex
t. However, these methods can not analyze 3D shapes well due to the self-occlusion and limitation of learning manifolds. To resolve this issue, we propose a method to learn joint embedding of point clouds and text by matching parts from shapes to words from sentences in a common space. Specifically, we first learn segmentation prior to segment point clouds into parts. Then, we map parts and words into an optimized space, where the parts and words can be matched with each other. In the optimized space, we represent a part by aggregating features of all points within the part, while representing each word with its context information, where we train our network to minimize the triplet ranking loss. Moreover, we also introduce cross-modal attention to capture the relationship of part-word in this matching procedure, which enhances joint embedding learning. Our experimental results outperform the state-of-the-art in multi-modal retrieval under the widely used benchmark.
Noninvasive optical imaging through dynamic scattering media has numerous important biomedical applications but still remains a challenging task. While standard methods aim to form images based upon optical absorption or fluorescent emission, it is a
lso well-established that the temporal correlation of scattered coherent light diffuses through tissue much like optical intensity. Few works to date, however, have aimed to experimentally measure and process such data to demonstrate deep-tissue imaging of decorrelation dynamics. In this work, we take advantage of a single-photon avalanche diode (SPAD) array camera, with over one thousand detectors, to simultaneously detect speckle fluctuations at the single-photon level from 12 different phantom tissue surface locations delivered via a customized fiber bundle array. We then apply a deep neural network to convert the acquired single-photon measurements into video of scattering dynamics beneath rapidly decorrelating liquid tissue phantoms. We demonstrate the ability to record video of dynamic events occurring 5-8 mm beneath a decorrelating tissue phantom with mm-scale resolution and at a 2.5-10 Hz frame rate.
Making accurate recommendations for cold-start users has been a longstanding and critical challenge for recommender systems (RS). Cross-domain recommendations (CDR) offer a solution to tackle such a cold-start problem when there is no sufficient data
for the users who have rarely used the system. An effective approach in CDR is to leverage the knowledge (e.g., user representations) learned from a related but different domain and transfer it to the target domain. Fine-tuning works as an effective transfer learning technique for this objective, which adapts the parameters of a pre-trained model from the source domain to the target domain. However, current methods are mainly based on the global fine-tuning strategy: the decision of which layers of the pre-trained model to freeze or fine-tune is taken for all users in the target domain. In this paper, we argue that users in RS are personalized and should have their own fine-tuning policies for better preference transfer learning. As such, we propose a novel User-specific Adaptive Fine-tuning method (UAF), selecting which layers of the pre-trained network to fine-tune, on a per-user basis. Specifically, we devise a policy network with three alternative strategies to automatically decide which layers to be fine-tuned and which layers to have their parameters frozen for each user. Extensive experiments show that the proposed UAF exhibits significantly better and more robust performance for user cold-start recommendation.
With the continuous increase of the spectrum and antennas, endogenous sensing is now possible in the fifth generation and future wireless communication systems. However, sensing is a highly complex task for a heterogeneous communication network with
massive connections. Seeking multi-domain cooperation is necessary. In this article, we present an integrated sensing and communication (ISAC) system that performs active, passive, and interactive sensing in different stages of communication through hardware and software. We also propose different methods about how multi-user and multi-frequency band cooperate to further enhance the ISAC systems performance. Finally, we elaborate on the advantages of multi-domain cooperation from the physical layer to the network layer for the ISAC system.
Whole-brain surface extraction is an essential topic in medical imaging systems as it provides neurosurgeons with a broader view of surgical planning and abnormality detection. To solve the problem confronted in current deep learning skull stripping
methods lacking prior shape information, we propose a new network architecture that incorporates knowledge of signed distance fields and introduce an additional Laplacian loss to ensure that the prediction results retain shape information. We validated our newly proposed method by conducting experiments on our brain magnetic resonance imaging dataset (111 patients). The evaluation results demonstrate that our approach achieves comparable dice scores and also reduces the Hausdorff distance and average symmetric surface distance, thus producing more stable and smooth brain isosurfaces.
Recently, reconfigurable intelligent surfaces (RISs) have drawn intensive attention to enhance the coverage of millimeter wave (mmWave) communication systems. However, existing works mainly consider the RIS as a whole uniform plane, which may be unre
alistic to be installed on the facade of buildings when the RIS is extreme large. To address this problem, in this paper, we propose a sparse array of sub-surface (SAoS) architecture for RIS, which contains several rectangle shaped sub-surfaces termed as RIS tiles that can be sparsely deployed. An approximated ergodic spectral efficiency of the SAoS aided system is derived and the performance impact of the SAoS design is evaluated. Based on the approximated ergodic spectral efficiency, we obtain an optimal reflection coefficient design for each RIS tile. Analytical results show that the received signal-to-noise ratios can grow quadratically and linearly to the number of RIS elements under strong and weak LoS scenarios, respectively. Furthermore, we consider the visible region (VR) phenomenon in the SAoS aided mmWave system and find that the optimal distance between RIS tiles is supposed to yield a total SAoS VR nearly covering the whole blind coverage area. The numerical results verify the tightness of the approximated ergodic spectral efficiency and demonstrate the great system performance.
We propose a continuous neural network architecture, termed Explainable Tensorized Neural Ordinary Differential Equations (ETN-ODE), for multi-step time series prediction at arbitrary time points. Unlike the existing approaches, which mainly handle u
nivariate time series for multi-step prediction or multivariate time series for single-step prediction, ETN-ODE could model multivariate time series for arbitrary-step prediction. In addition, it enjoys a tandem attention, w.r.t. temporal attention and variable attention, being able to provide explainable insights into the data. Specifically, ETN-ODE combines an explainable Tensorized Gated Recurrent Unit (Tensorized GRU or TGRU) with Ordinary Differential Equations (ODE). The derivative of the latent states is parameterized with a neural network. This continuous-time ODE network enables a multi-step prediction at arbitrary time points. We quantitatively and qualitatively demonstrate the effectiveness and the interpretability of ETN-ODE on five different multi-step prediction tasks and one arbitrary-step prediction task. Extensive experiments show that ETN-ODE can lead to accurate predictions at arbitrary time points while attaining best performance against the baseline methods in standard multi-step time series prediction.
