Search

Maritime autonomy, specifically the use of autonomous and semi-autonomous maritime vessels, is a key enabling technology supporting a set of diverse and critical research areas, including coastal and environmental resilience, assessment of waterway health, ecosystem/asset monitoring and maritime port security. Critical to the safe, efficient and reliable operation of an autonomous maritime vessel is its ability to perceive on-the-fly the external environment through onboard sensors. In this paper, buoy detection for LiDAR images is explored by using several tools and techniques: machine learning methods, Unity Game Engine (herein referred to as Unity) simulation, and traditional image processing. The Unity Game Engine (herein referred to as Unity) simulation data was used for the training and testing of a Pointnet neural network model while the labeled real-world maritime environment point cloud data was used for the model validation. Fitting the Pointnet model on the simulation data, after some data alignment with the LiDAR images allowed for accurate classification of buoys on the real-world data with the 93% of accuracy. A traditional image processing approach using 2D occupancy maps to detect the buoys by shape was used as well and is outlined in the paper.

We develop a framework to establish benchmarks for machine learning and deep neural networks analyses of exclusive scattering cross sections (FemtoNet). Within this framework we present an extraction of Compton form factors for deeply virtual Compton scattering from an unpolarized proton target. Critical to this effort is a study of the effects of physics constraint built into machine learning (ML) algorithms. We use the Bethe-Heitler process, which is the QED radiative background to deeply virtual Compton scattering, to test our ML models and, in particular, their ability to generalize information extracted from data. We then use these techniques on the full cross section and compare the results to analytic model calculations. We propose a quantification technique, the random targets method, to begin understanding the separation of aleatoric and epistemic uncertainties as they are manifest in exclusive scattering analyses. We propose a set of both physics driven and machine learning based benchmarks providing a stepping stone towards applying explainable machine learning techniques with controllable uncertainties in a wide range of deeply virtual exclusive processes., 16 pages, 12 figures

Multi-omics clustering methods are used for the stratification of patients into sub-groups of similar molecular characteristics. In recent years, a wide range of methods has been developed for this purpose. However, due to the high diversity of cancer-related data, a single method may not perform sufficiently well in all cases. Here, we propose a comprehensive framework for multi-omics hierarchical ensemble clustering. We provide a flexible environment that allows to build hierarchical clustering ensembles suitable for the available data and research goals. Survival analyses for data from The Cancer Genome Atlas (TCGA) indicate that our proposed ensembles provide more robust, and thus more reliable results than the state-of-the-art. We have implemented our architecture within the R-package HC-fused, which is freely available on Github.

We present a physics informed deep learning technique for Deeply Virtual Compton Scattering (DVCS) cross sections from an unpolarized proton target using both an unpolarized and polarized electron beam. Training a deep learning model typically requires a large size of data that might not always be available or possible to obtain. Alternatively, a deep learning model can be trained using additional knowledge gained by enforcing some physics constraints such as angular symmetries for better accuracy and generalization. By incorporating physics knowledge to our deep learning model, our framework shows precise predictions on the DVCS cross sections and better extrapolation on unseen kinematics compared to the basic deep learning approaches.

In many scientific applications, Inverse problems are challenging. An inverse problem is the process of inferring unknown parameters from observable ones. In this poster, we present our prototype using Point Cloud-based Variational Autoencoder mapping. Data that connects parameters to detector level events is used to train the proposed model. A point cloud is used to describe a series of events that keeps the permutation invariant property and geometric correlations of the events while being flexible with the number of events in the input. The trained Point Cloud-based Variational Autoencoder functions as an effective inverse function from detector level events to parameter space and can be utilized as the final step in inferring QCFs model parameters from experimental detector level events. Using a point cloud-based variational autoencoder, our results show that Sigma1 and Sigma2 are within one standard division of the predicted Sigma1 (pc) and Sigma2 (pc). The suggested model can be extended to high-dimensional events with permutation invariant features in future work.

As an acute zoonosis, rabies has a fatality rate of nearly 100%. Since rabies surveillance data over the years show that the rabies epidemic areas in China expand from southern provinces to central and northern provinces, we selected Hubei Province, an area with a high incidence of rabies in Central China, to investigate national rabies prevention and control work by developing a visual analysis platform (VAP-HRHB: http://www.combio-lezhang.online/rabies/index.html). VAP-HRHB employs bioinformatics methods to predict the future developmental trend for rabies incidence and exposure numbers, thereby comprehensively improving the prevention and control of infectious diseases in China.

In this paper, we present a fast implementation of the Singular Value Thresholding (SVT) algorithm for matrix completion. A rank-revealing randomized singular value decomposition (R3SVD) algorithm is used to adaptively carry out partial singular value decomposition (SVD) to fast approximate the SVT operator given a desired, fixed precision. We extend the R3SVD algorithm to a recycling rank revealing randomized singular value decomposition (R4SVD) algorithm by reusing the left singular vectors obtained from the previous iteration as the approximate basis in the current iteration, where the computational cost for partial SVD at each SVT iteration is significantly reduced. A simulated annealing style cooling mechanism is employed to adaptively adjust the low-rank approximation precision threshold as SVT progresses. Our fast SVT implementation is effective in both large and small matrices, which is demonstrated in matrix completion applications including image recovery and movie recommendation system.

In this paper, we revisit the generalized block power methods for approximating the eigenvector associated with $\lambda_1 = 1$ of a Markov chain transition matrix. Our analysis of the block power method shows that when $s$ linearly independent probability vectors are used as the initial block, the convergence of the block power method to the stationary distribution depends on the magnitude of the $(s+1)$th dominant eigenvalue $\lambda_{s+1}$ of $P$ instead of that of $\lambda_2$ in the power method. Therefore, the block power method with block size $s$ is particularly effective for transition matrices where $|\lambda_{s+1}|$ is well separated from $\lambda_1 = 1$ but $|\lambda_2|$ is not. This approach is particularly useful when visiting the elements of a large transition matrix is the main computational bottleneck over matrix--vector multiplications, where the block power method can effectively reduce the total number of times to pass over the matrix. To further reduce the overall computational cost, we combine the block power method with a sliding window scheme, taking advantage of the subsequent vectors of the latest $s$ iterations to assemble the block matrix. The sliding window scheme correlates vectors in the sliding window to quickly remove the influences from the eigenvalues whose magnitudes are smaller than $|\lambda_{s}|$ to reduce the overall number of matrix--vector multiplications to reach convergence. Finally, we compare the effectiveness of these methods in a Markov chain model representing a stochastic luminal calcium release site.

In this paper, we present a Rank Revealing Randomized Singular Value Decomposition (R3SVD) algorithm to incrementally construct a low-rank approximation of a potentially large matrix while adaptively estimating the appropriate rank that can capture most of the actions of the matrix. Starting from a low-rank approximation with an initial guessed rank, R3SVD adopts an orthogonal Gaussian sampling approach to obtain the dominant subspace within the leftover space, which is used to add up to the existing low-rank approximation. Orthogonal Gaussian sampling is repeated until an appropriate low-rank approximation with satisfactory accuracy, measured by the overall energy percentage of the original matrix, is obtained. While being a fast algorithm, R3SVD is also a memory-aware algorithm where the computational process can be decomposed into a series of sampling tasks that use constant amount of memory. Numerical examples in image compression and matrix completion are used to demonstrate the effectiveness of R3SVD in low-rank approximation.

Asthma is a common chronic lung disease. National guidelines encourage a stepwise approach to pharmacotherapy, and as such, an individual's current treatment step can be considered as a severity categorization proxy. BTS/SIGN steps can be estimated from electronic prescription records, however substantial data processing is required including extracting information from free-text drug descriptions and dose instructions. Almost 4.5 million asthma controller inhalers were prescribed for people in the Asthma Learning Health System (ALHS) Scottish cohort between 2009 and 2017. Asthma treatment regimens were identified and categorized by the combination of medications prescribed in the 120 days preceding prescribing events. 26% of prescriptions had no primary controller (inhaled corticosteroid) prescriptions in the previous 120 days and were thus assigned Step 0. 16% of prescriptions were assigned to BTS/SIGN Step 1, 7% to Step 2, 21% to Step 3, and 30% to Step 4. We developed a robust methodology enabling researchers to easily replicate BTS/SIGN asthma treatment step estimates, to both describe the severity of asthma in a population and to demonstrate changes over time. This can provide valuable insights into population and patient-specific trajectories, to improve understanding and management of symptoms.

There are multiple papers published about different AI models for the COVID-19 diagnosis with promising results. Unfortunately according to the reviews many of the papers do not reach the level of sophistication needed for a clinically usable model. In this paper I go through multiple review papers, guidelines, and other relevant material in order to generate more comprehensive requirements for the future papers proposing a AI based diagnosis of the COVID-19 from chest X-ray data (CXR). Main findings are that a clinically usable AI needs to have an extremely good documentation, comprehensive statistical analysis of the possible biases and performance, and an explainability module., Accepted to AI&BDvsPandemics BIBM 2021 Workshop

Funding Information: This work is supported by the Major Science and Technology Innovation Projects of Key R&D Programs of Shandong Province in 2019 (grant No.2019JZZY010113) Publisher Copyright: © 2021 IEEE. EEG based emotion recognition technology is currently an important concept in artificial intelligence, and also holds great potential in emotional health care. Nevertheless, one major limitation of the prior approaches is they do not capture the relationships between different time-series and channels explicitly, resulting in inevitable low performance, especially in subject-independent recognition settings. In this paper, we propose a novel graph attention network based model to address this issue. Our framework includes dual-pipeline Graph Attention Network layers in parallel to learn the complex dependencies of multi-channel EEG in both temporal and spatial dimensions. The proposed method outperforms other state-of-the-art models on benchmark SEED dataset. Further analysis shows that our method also has good interpretability. As far as we know, it is the first work that introduce graph attention network into EEG based emotion detection research.

Funding Information: Acknowledgment This work is supported by National Science Foundation of China under grant No. 62006212, the fund of State Key Lab. for Novel Software Technology in Nanjing University under grant No.KFKT2021B41. Publisher Copyright: © 2021 IEEE. The COVID-19 pandemic has had a severe impact on humans' lives and and healthcare systems worldwide. How to early, fastly and accurately diagnose infected patients via multimodal learning is now a research focus. The central challenges in this task mainly lie on multi-modal data representation and multi-modal feature fusion. To solve such challenges, we propose a medical knowledge enriched multi-modal sequence to sequence learning model, termed MedSeq2Seq. The key components include two attention mechanisms, viz. intra-modal (Ia) and inter-model (Ie) attentions, and a medical knowledge augmentation mechanism. The former two mechanisms are to learn multi-modal refined representation, while the latter aims to incorporate external medical knowledge into the proposed model. The experimental results show the effectiveness of the proposed MedSeq2Seq framework over state-of-the-art baselines with a significant improvement of 1%-2%.

Funding Information: This work is supported by National Science Foundation of China under grant No. 62006212, the fund of State Key Lab. for Novel Software Technology in Nanjing University under grant No.KFKT2021B41. Publisher Copyright: © 2021 IEEE. Depression is a typical mood disease that makes people a persistent feeling of sadness and loss of interest and pleasure. Emotion thus comes into sight and is tightly entangled with depression in that one helps the understanding of the other. Depression and emotion detection has been a new research task. The central challenges in this task are multi-modal interaction and multi-task correlation. The existing approaches treat them as two separate tasks, and fail to model the relationships between them. In this paper, we propose an attentive multi-modal multitask learning framework, called AMM, to generically address such issues. The core modules are two attention mechanisms, viz. inter-modal (I {mathrm{e}}) and inter-task (I {t}) attentions. The main motivation of I {mathrm{e}} attention is to learn multi-modal fused representation. In contrast, It attention is proposed to learn the relationship between depression detection and emotion recognition. Extensive experiments are conducted on two large scale datasets, i.e., DAIC and multi-modal Getty Image depression (MGID). The results show the effectiveness of the proposed AMM framework, and also shows that AMM obtains better performance for the main task, i.e., depression detection with the help of the secondary emotion recognition task.

Sequence alignments have been studied for decades under the simplified model of a consensus sequence representing a chromosome. A natural question is if there is some more accurate notion of alignment for diploid (and in general, polyploid) organisms. We have developed such a notion in our recent work, but unfortunately the computational complexity remains open for such a diploid pair-wise alignment; only a trivial exponential algorithm is known that goes over all possible diploid alignments. In this paper, we shed some light on the complexity of diploid alignments by showing that a haplotyping version, involving three diploid inputs, is polynomial time solvable.