Search

Deep neural networks (DNNs) have greatly benefited direction of arrival (DoA) estimation methods for speech source localization in noisy environments. However, their localization accuracy is still far from satisfactory due to the vulnerability to nonspeech interference. To improve the robustness against interference, we propose a DNN based normalized time-frequency (T-F) weighted criterion which minimizes the distance between the candidate steering vectors and the filtered snapshots in the T-F domain. Our method requires no eigendecomposition and uses a simple normalization to prevent the optimization objective from being misled by noisy filtered snapshots. We also study different designs of T-F weights guided by a DNN. We find that duplicating the Hadamard product of speech ratio masks is highly effective and better than other techniques such as direct masking and taking the mean in the proposed approach. However, the best-performing design of T-F weights is criterion-dependent in general. Experiments show that the proposed method outperforms popular DNN based DoA estimation methods including widely used subspace methods in noisy and reverberant environments., 5 pages. Accepted at ICASSP 2023

A deep neural network using rectified linear units represents a continuous piecewise linear (CPWL) function and vice versa. Recent results in the literature estimated that the number of neurons needed to exactly represent any CPWL function grows exponentially with the number of pieces or exponentially in terms of the factorial of the number of distinct linear components. Moreover, such growth is amplified linearly with the input dimension. These existing results seem to indicate that the cost of representing a CPWL function is expensive. In this paper, we propose much tighter bounds and establish a polynomial time algorithm to find a network satisfying these bounds for any given CPWL function. We prove that the number of hidden neurons required to exactly represent any CPWL function is at most a quadratic function of the number of pieces. In contrast to all previous results, this upper bound is invariant to the input dimension. Besides the number of pieces, we also study the number of distinct linear components in CPWL functions. When such a number is also given, we prove that the quadratic complexity turns into bilinear, which implies a lower neural complexity because the number of distinct linear components is always not greater than the minimum number of pieces in a CPWL function. When the number of pieces is unknown, we prove that, in terms of the number of distinct linear components, the neural complexities of any CPWL function are at most polynomial growth for low-dimensional inputs and factorial growth for the worst-case scenario, which are significantly better than existing results in the literature., 31 pages. Accepted at NeurIPS 2022

Additional file 1: Table A1. Definitions of the different types of medication errors in the reporting system at the NTUH. Table A2. Outcome variable definitions. Table A3. Baseline characteristics of the patients between phases 1 and 2. Table A4. Baseline characteristics of the patients between the intervention and control wards. Table A5. Sensitivity analysis: effects of the NTUH-IMM model on outcome variables when shortening the observational interval to half-month. Table A6. Subgroup analysis: Effects of the NTUH-IMM model on the outcome variables in patients without prolonged lengths of stay. Table A7. Subgroup analysis: Effects of the NTUH-IMM model on the outcome variables for patients who did not expire or were transferred to the intensive care unit during hospitalization

Designing complex, dynamic yet multi-functional materials and devices is challenging because the design spaces for these materials have numerous interdependent and often conflicting constraints. Taking inspiration from advances in artificial intelligence and their applications in material discovery, we propose a computational method for designing metamorphic DNA-co-polymerized hydrogel structures. The method consists of a coarse-grained simulation and a deep learning-guided optimization system for exploring the immense design space of these structures. Here, we develop a simple numeric simulation of DNA-co-polymerized hydrogel shape change and seek to find designs for structured hydrogels that can fold into the shapes of different Arabic numerals in different actuation states. We train a convolutional neural network to classify and score the geometric outputs of the coarse-grained simulation to provide autonomous feedback for design optimization. We then construct a genetic algorithm that generates and selects large batches of material designs that compete with one another to evolve and converge on optimal objective-matching designs. We show that we are able to explore the large design space and learn important parameters and traits. We identify vital relationships between the material scale size and the range of shape change that can be achieved by individual domains and we elucidate trade-offs between different design parameters. Finally, we discover material designs capable of transforming into multiple different digits in different actuation states., LIPIcs, Vol. 238, 28th International Conference on DNA Computing and Molecular Programming (DNA 28), pages 4:1-4:14

Kuan-Lin Chen,1–3 Cheng-Te Chen,4 Chien-Ho Lin,5 Chien-Yu Huang,6 Ya-Chen Lee7 1Department of Occupational Therapy, College of Medicine, National Cheng Kung University, Tainan City 701, Taiwan (R.O.C); 2Department of Physical Medicine and Rehabilitation, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan City 701, Taiwan (R.O.C); 3Institute of Allied Health Sciences, College of Medicine, National Cheng Kung University, Tainan City 701, Taiwan (R.O.C); 4Department of Educational Psychology and Counseling, National Tsing Hua University, Hsinchu, Taiwan (R.O.C); 5Department of Psychiatry, Chi Mei Medical Center, Youngkang Dist., Tainan City 710, Taiwan (R.O.C); 6Department of Occupational Therapy, I-Shou University, Yanchao District, Kaohsiung City 824, Taiwan (R.O.C); 7Department of Occupational Therapy, College of Medical and Health Science, Asia University, Wufeng, Taichung, Taiwan (R.O.C)Correspondence: Ya-Chen LeeDepartment of Occupational Therapy, College of Medical and Health Science, Asia University, No. 500, Lioufeng Road, Wufeng, Taichung, Taiwan (R.O.C)Tel +886-4-23323456 Ext. 20007Email jennyleenet@hotmail.comBackground: Children with autism spectrum disorder (ASD) often exhibit deficits in pretend play and have less playfulness. The purpose of this study was to evaluate the relationship between pretend play and playfulness in children with autism spectrum disorder, while controlling for severity of autism behaviors, verbal comprehension, and age.Methods: A sample of 72 children with ASD aged between 3 and 12 years were assessed with the Child-Initiated Pretend Play Assessment, Test of Playfulness, and Childhood Autism Rating Scale, respectively, for their pretend play, playfulness, and severity of autism behaviors. Correlation and multiple regression analyses were conducted.Results: The results of Pearson correlation coefficients revealed that the pretend play variables had mild to moderate associations with the playfulness variables (r = −0.25 to 0.68). The multiple regression analyses showed that, overall, the internal locus of control was the significant predictor of the pretend play variables (accounting for 5–47% of the variance, p < 0.001). The six pretend play variables were all important predictors of all playfulness variables (explaining 41–76% of the variance, p < 0.001–0.047). Particularly, the elaborate pretend play action was a significant predictor of all four playfulness variables. Our findings indicated that the more children with ASD engaged in pretend play, the more they experienced playfulness.Conclusion: Clinicians could help children with ASD improve their feeling of being in charge of their play in order to develop better performance in pretend play. Assisting children with ASD to engage in pretend play is important to promote their internal experience of playfulness.Keywords: autism spectrum disorder, pretend play, playfulness, severity of autism behaviors

Models recently used in the literature proving residual networks (ResNets) are better than linear predictors are actually different from standard ResNets that have been widely used in computer vision. In addition to the assumptions such as scalar-valued output or single residual block, these models have no nonlinearities at the final residual representation that feeds into the final affine layer. To codify such a difference in nonlinearities and reveal a linear estimation property, we define ResNEsts, i.e., Residual Nonlinear Estimators, by simply dropping nonlinearities at the last residual representation from standard ResNets. We show that wide ResNEsts with bottleneck blocks can always guarantee a very desirable training property that standard ResNets aim to achieve, i.e., adding more blocks does not decrease performance given the same set of basis elements. To prove that, we first recognize ResNEsts are basis function models that are limited by a coupling problem in basis learning and linear prediction. Then, to decouple prediction weights from basis learning, we construct a special architecture termed augmented ResNEst (A-ResNEst) that always guarantees no worse performance with the addition of a block. As a result, such an A-ResNEst establishes empirical risk lower bounds for a ResNEst using corresponding bases. Our results demonstrate ResNEsts indeed have a problem of diminishing feature reuse; however, it can be avoided by sufficiently expanding or widening the input space, leading to the above-mentioned desirable property. Inspired by the DenseNets that have been shown to outperform ResNets, we also propose a corresponding new model called Densely connected Nonlinear Estimator (DenseNEst). We show that any DenseNEst can be represented as a wide ResNEst with bottleneck blocks. Unlike ResNEsts, DenseNEsts exhibit the desirable property without any special architectural re-design., Comment: 24 pages. Accepted by NeurIPS 2021. Remark 1 clarified and typos corrected

Rationale: Shoulder pain is a common complaint among patients, and rotator cuff disorders are the most common diagnoses. Humeral head osteonecrosis is easily masked by other more common diagnoses and concomitant conditions. Patient concerns: This challenging diagnostic report consists of 2 cases. Case 1 was that of a 59-year-old man who presented with right shoulder pain that had lasted for >1 year. Case 2 was that of a 52-year-old man who complained of right shoulder pain lasting for 6 months. They both presented with chronic right shoulder pain without relevant trauma history, and the physical examination showed a tenderness point over the right greater tuberosity. Diagnosis: These 2 patients were diagnosed with osteonecrosis involving the right greater tuberosity region via magnetic resonance imaging. Interventions: In case 1, the patient underwent cord decompression and artificial bone grafting with C-arm guidance. In case 2, the patient refused surgical intervention and decided to continue receiving physical therapy for symptom control. Outcomes: In case 1, the patient responded well to cord decompression and artificial bone grafting. After the surgery, the active range of motion was restored and the pain in the right shoulder diminished further. In case 2, conservative treatment helped alleviate the patient's shoulder pain but did not entirely eliminate it. Lessons: Physicians should always have a high index of suspicion for osteonecrosis, especially when treating chronic shoulder pain, regardless of whether there are typical symptoms/known risk factors or not.

There are several challenges for search and rescue robots: mobility, perception, autonomy, and communication. Inspired by the DARPA Subterranean (SubT) Challenge, we propose an autonomous blimp robot, which has the advantages of low power consumption and collision-tolerance compared to other aerial vehicles like drones. This is important for search and rescue tasks that usually last for one or more hours. However, the underground constrained passages limit the size of blimp envelope and its payload, making the proposed system resource-constrained. Therefore, a careful design consideration is needed to build a blimp system with on-board artifact search and SLAM. In order to reach long-term operation, a failure-aware algorithm with minimal communication to human supervisor to have situational awareness and send control signals to the blimp when needed.

Digital servitization is gradually having the operational impacts in conventional B2B manufacturing firms, because the firms’ operational focus needs to move towards product-service excellence. In this particular scenario, practitioners experience the causality dilemma and are often trapped in so-called servitization paradox. This experimental study intends to demonstrate how to overcome such dilemma. A data-driven approach is used to provide solid evidences about business potentials to be considered when taking on further operational actions. Design science research method is used to develop a product-service platform. Taking with the service-dominant (S-D) logic thinking, we consolidate the heterogeneous data, from the internal ERP system and the external Internet of Things/Internet of Services platforms. The intention is to promote product-service thinking cross the organization through this platform and recognize value-adding tasks to pursue product-service business goals. Our approach was tested with industrial data from a Danish manufacturing firm.

Robotic software and hardware systems of autonomous surface vehicles have been developed in transportation, military, and ocean researches for decades. Previous efforts in RobotX Challenges 2014 and 2016 facilitates the developments for important tasks such as obstacle avoidance and docking. Team NCTU is motivated by the AI Driving Olympics (AI-DO) developed by the Duckietown community, and adopts the principles to RobotX challenge. With the containerization (Docker) and uniformed AI agent (with observations and actions), we could better 1) integrate solutions developed in different middlewares (ROS and MOOS), 2) develop essential functionalities of from simulation (Gazebo) to real robots (either miniaturized or full-sized WAM-V), and 3) compare different approaches either from classic model-based or learning-based. Finally, we setup an outdoor on-surface platform with localization services for evaluation. Some of the preliminary results will be presented for the Team NCTU participations of the RobotX competition in Hawaii in 2018.

Rationale: Bilateral anterior shoulder dislocation is clinically rare and has been reported to be related to high-speed sports. Patient concerns: A 76-year-old woman presented with bilateral shoulders pain after traditional Chinese manipulation. Diagnoses: She was diagnosed with bilateral anterior dislocations, and a closed reduction was immediately performed. Intervention: The patient was referred for rehabilitation 3 days later, and bilateral rotator cuff injuries were identified from musculoskeletal ultrasound. After 4 weeks of physical therapy, the patient's shoulder pain had reduced and the passive ROM was nearly full. Outcomes: At 1-year follow-up, only mild intermittent shoulder pain was noted, and there was no limitation of shoulder ROM. Lessons: This case illustrates that patients with acute shoulder injuries who receive proper diagnosis and treatment can achieve good outcomes. Therefore, patients with musculoskeletal disorders should seek qualified specialists for accurate diagnosis and appropriate management.

Background Cardiovascular disease (CVD)accounts for 31% of all global deaths. Some CVD mortalities can be attributed to environmental factors such as particulate matters(PMs). Coal fired power plant is one of the major contributors of PM. However, the short-term effect of coal fired power plant on cardiovascular disease is not well studied. In this study, we investigate an association between coal capacity and CVD mortality from a global perspective. Method Age and Sex-adjusted CVD mortalities of 111 countries were followed from 1998 to 2012.Coal capacity was defined as total capacity of coal fired power plants in a given country in a given year, from Utility Data Institute World Electric Power Plants(UDI WEPP)Database while CVD mortality were obtained from WHO mortality data. We applied mix model and adjusted other risk factors for analysis. Results The average coal capacity around the world is increasing globally, but coal percentage used has been fairly constant(8555.18 MW(15.99%) in 1998–2002,12071.11 MW(16.71%)in 2003–2007 and 16394.05 MW(16.58%) in 2008–2012. One log coal capacity (unit: log MW) was associated with an increase in CVD mortality by 22.98(p=0.076) to 55.74(p Conclusion The result of the current study indicated that after adjusting for commonly known risk factors of cardiovascular disease, coal fired power plants emission is correlated with country specific short term cardiovascular mortality.