Conventional methods for computing maximum-likelihood estimators (MLE) often converge slowly in practical situations, leading to a search for simplifying methods that rely on additional assumptions for their validity. In this work, we provide a fast and reliable algorithm for maximum likelihood reconstruction that avoids this slow convergence. Our method utilizes the state-of-the-art convex optimization scheme-an accelerated projected-gradient method-that allows one to accommodate the quantum nature of the problem in a different way than in the standard methods. We demonstrate the power of our approach by comparing its performance with other algorithms for n-qubit state tomography. In particular, an 8-qubit situation that purportedly took weeks of computation time in 2005 can now be completed in under a minute for a single set of data, with far higher accuracy than previously possible. This refutes the common claim that MLE reconstruction is slow, and reduces the need for alternative methods that often come with difficult-to-verify assumptions. In fact, recent methods assuming Gaussian statistics or relying on compressed sensing ideas are demonstrably inapplicable for the situation under consideration here. Our algorithm can be applied to general optimization problems over the quantum state space; the philosophy of projected gradients can further be utilized for optimization contexts with general constraints. PACS numbers: 03.65.Wj, 03.67.-a, 02.60.PnIntroduction.-Efficient and reliable characterization of properties of a quantum system, e.g., its state or the process it is undergoing, is needed for any quantum information processing task. Such are the goals of quantum tomography [1], broadly classified into state tomography and process tomography. Process tomography can be recast as state tomography via the Choi-Jamiolkowski isomorphism [2, 3]; we hence restrict our attention to state tomography. Tomography is a two-step process: the first is data gathering from measurements of the quantum system; the second is the estimation of the state from the gathered data. This second step is the focus of this article.A popular estimation strategy is that of the maximumlikelihood estimator (MLE) [4] from standard statistics. Computing the MLE for quantum tomography is, however, not straightforward due to the constraints imposed by quantum mechanics. While general-purpose convex optimization toolboxes (e.g., CVX [5,6]) are available for small-sized problems, specially adapted MLE algorithms are needed for tackling useful system sizes. Past MLE algorithms [7,8] incorporate the quantum constraints by going to the factored space (see definition later) where the constraints are satisfied by an appropriate parameterization. Gradient methods are then straightforwardly employed in the now-unconstrained factored space. These algorithms can be slow in practice, with an extreme

Our results suggest that overexpressed miR-191 is associated with ICC progression through the miR-191/TET1/p53 pathway. (Hepatology 2017;66:136-151).

BackgroundStudies investigating the association between hepatitis C virus (HCV) infections and the occurrence of cholangiocarcinoma (CCA), especially intrahepatic cholangiocarcinoma (ICC), have shown inconsistent findings. Although previous meta-analyses referred to HCV and CCA, they mainly focused on ICC rather than CCA or extrahepatic cholangiocarcinoma (ECC). Since then, relevant new studies have been published on the association between HCV and ICC. Since the different anatomic locations of CCA have distinct epidemiologic features and different risk factors, it is necessary to evaluate the relationship between HCV infection and ICC, ECC, and CCA.MethodsRelevant studies were identified by searching PUBMED, EMBASE, and MEDLINE databases prior to 1 August 2013. Pooled risk estimates were calculated with random-effects models using STATA 11.0.ResultsA total of 16 case-control studies were included in the final analysis. Pooled risk estimates showed a statistically significant increasing risk of CCA (odds ratio (OR) = 5.44, 95% CI, 2.72 to 10.89). The pooled risk estimate of ICC (OR = 3.38, 95% CI, 2.72 to 4.21) was higher than that of ECC (OR = 1.75, 95% CI, 1.00 to 3.05). In a subgroup analysis, the pooled risk estimate of ICC in studies from North America was obviously higher than in Asia (6.48 versus 2.01). The Begg funnel plot and Egger test showed no evidence of publication bias.ConclusionsHCV infection is associated with the increasing risk of CCA, especially ICC.

No abstract

Infective factors cause the perpetuation of inflammation as a result of the permanent exposure of the immune system to exogenous or endogenous products of virus or bacteria. Mesenchymal stem cells (MSCs) can be exposed to this infective environment, which may change the characteristics and therapeutic potency of these MSCs. MSCs have the ability to repair damaged and inflamed tissues and regulate immune responses. In this study, we demonstrated that MSCs express functional Toll-like receptors (TLR) 3 and 4, the Toll-like receptor families that recognize the signals of viral and bacterial mimics, respectively. The specific stimulations did not affect the self-renewal and apoptosis capabilities of MSCs but instead promoted their differentiation into the adipocytes and osteoblasts with the TLR3 ligand. The reverse of these results were obtained with the TLR4 ligand. The migration of the MSCs to stimulate either of the two specific ligands was inhibited at different times, whereas the immunogenicity and immunosuppressive properties of the MSCs were not weakened unlike in the MSCs group. These results suggest that TLR3 and TLR4 stimulation affect the characterization of MSCs.

Deep learning (DL) has been applied extensively in many computational imaging problems, often leading to superior performance over traditional iterative approaches. However, two important questions remain largely unanswered: first, how well can the trained neural network generalize to objects very different from the ones in training? This is particularly important in practice, since large-scale annotated examples similar to those of interest are often not available during training. Second, has the trained neural network learnt the underlying (inverse) physics model, or has it merely done something trivial, such as memorizing the examples or point-wise pattern matching? This pertains to the interpretability of machine-learning based algorithms. In this work, we use the Phase Extraction Neural Network (PhENN) [Optica 4, 1117-1125 (2017)], a deep neural network (DNN) for quantitative phase retrieval in a lensless phase imaging system as the standard platform and show that the two questions are related and share a common crux: the choice of the training examples. Moreover, we connect the strength of the regularization effect imposed by a training set to the training process with the Shannon entropy of images in the dataset. That is, the higher the entropy of the training images, the weaker the regularization effect can be imposed. We also discover that weaker regularization effect leads to better learning of the underlying propagation model, i.e. the weak object transfer function, applicable for weakly scattering objects under the weak object approximation. Finally, simulation and experimental results show that better cross-domain generalization performance can be achieved if DNN is trained on a higher-entropy database, e.g. the ImageNet, than if the same DNN is trained on a lower-entropy database, e.g. MNIST, as the former allows the underlying physics model be learned better than the latter.

No abstract

Antiviral therapy improved the prognosis of hepatitis B virus-related HCC up to 3 cm. Antiviral therapy should be considered a standard postoperative adjuvant therapy of hepatitis B virus-related HCC.