Magnetic Resonance Imaging in Diagnosis of Complications of Renal and Ureteral Injuries in Different Periods of Traumatic Disease

The aim of this study was to evaluate the effectiveness of MRI in diagnosing combined renal and ureteral injuries at different periods of traumatic disease (TD). Methods and Results: We analyzed the results of diagnostics and treatment of 139 patients (80 women and 59 men) with renal and ureteral injuries aged between 18 and 72 years. There were 67(48.2%) patients in the period of acute reaction to trauma, 40(28.8%) patients with early manifestations, and 32(23%) patients in the period of late manifestations. In 127(91.4%) patients, an urgent plain abdominal X-ray was performed without any preliminary preparation. USI of the abdominal and retroperitoneal space was performed in 108(77.7%) patients in the stage of the primary assessment of renal injury as it was a rapid non-invasive investigation. A whole-body MSCT was performed in 131(94.2%) patients, using the nonionic contrast agents Ultravist (350mg I/ml) and Omnipaque (350mg I/ml). MRI was performed in 125(89.9%) patients, including cases of pregnancy and a medical history of allergies. Contrast-enhanced MSCT had a high diagnostic efficiency in assessing complications in kidney and ureteral injuries at different periods of TD (accuracy of 89.2% for acute reaction, 88.8% for early manifestations, and 89.5% for late manifestations). MRI of the kidneys and ureters was indicated in periods of early and late manifestations of TB to detect renal complications in cases with a discrepancy between clinical manifestations and the results obtained by ultrasound and MSCT (accuracy of 87.5% for early manifestations and 89.9% for late manifestations).

Frequency Comb Optoacoustic Tomography

Optoacoustics (OA) is overwhelmingly implemented in the Time Domain (TD) to achieve a high Signal-to-Noise-Ratio (SNR). Implementations in the Frequency Domain (FD) have been proposed, but have not offered competitive advantages over TD methods to reach high dissemination. It is therefore commonly believed that the TD represents the optimal way of performing optoacoustics. Here, we introduce a novel optoacoustic concept based on frequency comb and theoretically demonstrate its superiority to the TD. Then, using recent advances in laser diode illumination, we launch Frequency Comb Optoacoustic Tomography (FCOT), at multiple wavelengths, and experimentally demonstrate its advantages over TD methods in phantoms and in-vivo. We demonstrate that FCOT optimizes the SNR of spectral measurements over TD methods by benefiting from signal acquisition in the TD and processing in the FD, and that it reaches the fastest multi-spectral operation ever demonstrated in optoacoustics while reducing performance compromises present in TD systems.

Beyond Unfolding: Exact Recovery of Latent Convex Tensor Decomposition Under Reshuffling

Exact recovery of tensor decomposition (TD) methods is a desirable property in both unsupervised learning and scientific data analysis. The numerical defects of TD methods, however, limit their practical applications on real-world data. As an alternative, convex tensor decomposition (CTD) was proposed to alleviate these problems, but its exact-recovery property is not properly addressed so far. To this end, we focus on latent convex tensor decomposition (LCTD), a practically widely-used CTD model, and rigorously prove a sufficient condition for its exact-recovery property. Furthermore, we show that such property can be also achieved by a more general model than LCTD. In the new model, we generalize the classic tensor (un-)folding into reshuffling operation, a more flexible mapping to relocate the entries of the matrix into a tensor. Armed with the reshuffling operations and exact-recovery property, we explore a totally novel application for (generalized) LCTD, i.e., image steganography. Experimental results on synthetic data validate our theory, and results on image steganography show that our method outperforms the state-of-the-art methods.

Fixed-Horizon Temporal Difference Methods for Stable Reinforcement Learning

We explore fixed-horizon temporal difference (TD) methods, reinforcement learning algorithms for a new kind of value function that predicts the sum of rewards over a fixed number of future time steps. To learn the value function for horizon h, these algorithms bootstrap from the value function for horizon h−1, or some shorter horizon. Because no value function bootstraps from itself, fixed-horizon methods are immune to the stability problems that plague other off-policy TD methods using function approximation (also known as “the deadly triad”). Although fixed-horizon methods require the storage of additional value functions, this gives the agent additional predictive power, while the added complexity can be substantially reduced via parallel updates, shared weights, and n-step bootstrapping. We show how to use fixed-horizon value functions to solve reinforcement learning problems competitively with methods such as Q-learning that learn conventional value functions. We also prove convergence of fixed-horizon temporal difference methods with linear and general function approximation. Taken together, our results establish fixed-horizon TD methods as a viable new way of avoiding the stability problems of the deadly triad.

Assumed Density Filtering Q-learning

While off-policy temporal difference (TD) methods have widely been used in reinforcement learning due to their efficiency and simple implementation, their Bayesian counterparts have not been utilized as frequently. One reason is that the non-linear max operation in the Bellman optimality equation makes it difficult to define conjugate distributions over the value functions. In this paper, we introduce a novel Bayesian approach to off-policy TD methods, called as ADFQ, which updates beliefs on state-action values, Q, through an online Bayesian inference method known as Assumed Density Filtering. We formulate an efficient closed-form solution for the value update by approximately estimating analytic parameters of the posterior of the Q-beliefs. Uncertainty measures in the beliefs not only are used in exploration but also provide a natural regularization for the value update considering all next available actions. ADFQ converges to Q-learning as the uncertainty measures of the Q-beliefs decrease and improves common drawbacks of other Bayesian RL algorithms such as computational complexity. We extend ADFQ with a neural network. Our empirical results demonstrate that ADFQ outperforms comparable algorithms on various Atari 2600 games, with drastic improvements in highly stochastic domains or domains with a large action space.

Proximal Gradient Temporal Difference Learning: Stable Reinforcement Learning with Polynomial Sample Complexity

In this paper, we introduce proximal gradient temporal difference learning, which provides a principled way of designing and analyzing true stochastic gradient temporal difference learning algorithms. We show how gradient TD (GTD) reinforcement learning methods can be formally derived, not by starting from their original objective functions, as previously attempted, but rather from a primal-dual saddle-point objective function. We also conduct a saddle-point error analysis to obtain finite-sample bounds on their performance. Previous analyses of this class of algorithms use stochastic approximation techniques to prove asymptotic convergence, and do not provide any finite-sample analysis. We also propose an accelerated algorithm, called GTD2-MP, that uses proximal "mirror maps" to yield an improved convergence rate. The results of our theoretical analysis imply that the GTD family of algorithms are comparable and may indeed be preferred over existing least squares TD methods for off-policy learning, due to their linear complexity. We provide experimental results showing the improved performance of our accelerated gradient TD methods.

Outcomes after Early Neurological Deterioration and Transitory Deterioration in Acute Ischemic Stroke Patients

Background and Purpose: Early neurological deterioration (END) occurs in 10-40% of acute ischemic stroke (AIS) patients and has been associated with worse outcome. Recent improvements in treatment may have reduced the prevalence of END. A single early control or repeated observations have been applied to detect END close to occurrence, in order to improve the poor outcome associated with END, as clinical interventions may still be effective. Deterioration detected through repeated observations may be transitory or lead to END. Our aim was to study outcome after END and transitory deterioration (TD). Methods: In acute ischemic stroke patients, key Scandinavian Stroke Scale (SSS) items were scored 12 times from admission to 72 h. END was defined as ≥2 point decrease in any key SSS item from admission to 72 h. Early deterioration episode was defined as similar worsening between two consecutive assessments within 72 h, and TD as early deterioration episode in patients without END. Main outcome measures were odds ratios (OR) for worse functional outcome (including death) measured by modified Rankin scale at 90 days for END and TD compared with stable patients. Results: 368 patients were included. 13.9% had END and 28.3% had TD. Both deterioration groups were associated with worse outcome at 12 weeks compared with stable patients, with ORs of 35.1 (95% CI 8.8-140) for death/dependency and 5.8 (95% CI 1.8-19.4) for death in END patients and ORs of 2.3 (95% CI 1.1-4.8) for death/dependency and 1.9 (95% CI 0.5-6.3) for death in patients with TD. LOS increased by 6.4 days for END (p < 0.001) and 1.1 days for TD (p = 0.014) compared with stable patients. Conclusion: We found a strong association between END and worse outcome, and even TD doubled the OR for death/dependency compared to stable patients. Early deterioration episodes identified through frequent observations are therefore clinically significant and such frequent observations may detect worsening sufficiently close to occurrence for potentially effective treatment to be applied.

Repeatability and comparison of clinical tests in children with spastic diplegia and with typical development,

Objective The aims of this study were to evaluate the reliability of three range of motion tests (Straight Leg Raise, Modified Thomas, and Duncan-Ely) using photographic measurements in children with spastic diplegic cerebral palsy (SD) and with typical development (TD). Methods A cross-sectional test-retest design was applied to compare muscle-tendon unit shortening tests between groups. Results The tests showed reliability that ranged from good to excellent (ICC > 0.8), except for the Thomas Test for the bi-articular hip flexor muscle-tendon unit of the TD group (ICC = 0.7). Differences between groups were found in all tests (p < 0.05), except when the range of motion of the bi-articular hip flexor muscles was assessed using the Thomas test (p > 0.05). Conclusion Children with SD presents a smaller range of motion than the TD group. However, the Thomas Test for the bi-articular hip flexor muscles was unable to determine differences between children with spastic diplegic cerebral palsy from that with typical development.

Random-TD Function Approximator

In this paper, adaptive controller architecture based on a combination of temporal-difference (TD) learning and an on-line variant of Random Forest (RF) classifier is proposed. We call this implementation Random-TD. The approach iteratively improves its control strategies by exploiting only relevant parts of action and is able to learn completely in on-line mode. Such capability of on-line adaptation would take us closer to the goal of more robust and adaptable control. To illustrate this and to demonstrate the applicability of the approach, it has been applied to a non-linear, non-stationary control task, Cart-Pole balancing and on high-dimensional control problems –Ailerons, Elevator, Kinematics, and Friedman–. The results demonstrate that our hybrid approach is adaptable and can significantly improves the performance of TD methods while speeding up the learning process.