Search

We investigate the role of projection heads, also known as projectors, within the encoder-projector framework (e.g., SimCLR) used in contrastive learning. We aim to demystify the observed phenomenon where representations learned before projectors outperform those learned after -- measured using the downstream linear classification accuracy, even when the projectors themselves are linear. In this paper, we make two significant contributions towards this aim. Firstly, through empirical and theoretical analysis, we identify two crucial effects -- expansion and shrinkage -- induced by the contrastive loss on the projectors. In essence, contrastive loss either expands or shrinks the signal direction in the representations learned by an encoder, depending on factors such as the augmentation strength, the temperature used in contrastive loss, etc. Secondly, drawing inspiration from the expansion and shrinkage phenomenon, we propose a family of linear transformations to accurately model the projector's behavior. This enables us to precisely characterize the downstream linear classification accuracy in the high-dimensional asymptotic limit. Our findings reveal that linear projectors operating in the shrinkage (or expansion) regime hinder (or improve) the downstream classification accuracy. This provides the first theoretical explanation as to why (linear) projectors impact the downstream performance of learned representations. Our theoretical findings are further corroborated by extensive experiments on both synthetic data and real image data.

Matrix completion aims to estimate missing entries in a data matrix, using the assumption of a low-complexity structure (e.g., low rank) so that imputation is possible. While many effective estimation algorithms exist in the literature, uncertainty quantification for this problem has proved to be challenging, and existing methods are extremely sensitive to model misspecification. In this work, we propose a distribution-free method for predictive inference in the matrix completion problem. Our method adapts the framework of conformal prediction, which provides confidence intervals with guaranteed distribution-free validity in the setting of regression, to the problem of matrix completion. Our resulting method, conformalized matrix completion (cmc), offers provable predictive coverage regardless of the accuracy of the low-rank model. Empirical results on simulated and real data demonstrate that cmc is robust to model misspecification while matching the performance of existing model-based methods when the model is correct.

Text image machine translation (TIMT) aims to translate texts embedded in images from one source language to another target language. Existing methods, both two-stage cascade and one-stage end-to-end architectures, suffer from different issues. The cascade models can benefit from the large-scale optical character recognition (OCR) and MT datasets but the two-stage architecture is redundant. The end-to-end models are efficient but suffer from training data deficiency. To this end, in our paper, we propose an end-to-end TIMT model fully making use of the knowledge from existing OCR and MT datasets to pursue both an effective and efficient framework. More specifically, we build a novel modal adapter effectively bridging the OCR encoder and MT decoder. End-to-end TIMT loss and cross-modal contrastive loss are utilized jointly to align the feature distribution of the OCR and MT tasks. Extensive experiments show that the proposed method outperforms the existing two-stage cascade models and one-stage end-to-end models with a lighter and faster architecture. Furthermore, the ablation studies verify the generalization of our method, where the proposed modal adapter is effective to bridge various OCR and MT models., Accepted at The 17th International Conference on Document Analysis and Recognition (ICDAR 2023)

Emerging spatial technologies, including spatial transcriptomics and spatial epigenomics, are becoming powerful tools for profiling of cellular states in the tissue context1-5. However, current methods capture only one layer of omics information at a time, precluding the possibility of examining the mechanistic relationship across the central dogma of molecular biology. Here, we present two technologies for spatially resolved, genome-wide, joint profiling of the epigenome and transcriptome by cosequencing chromatin accessibility and gene expression, or histone modifications (H3K27me3, H3K27ac or H3K4me3) and gene expression on the same tissue section at near-single-cell resolution. These were applied to embryonic and juvenile mouse brain, as well as adult human brain, to map how epigenetic mechanisms control transcriptional phenotype and cell dynamics in tissue. Although highly concordant tissue features were identified by either spatial epigenome or spatial transcriptome we also observed distinct patterns, suggesting their differential roles in defining cell states. Linking epigenome to transcriptome pixel by pixel allows the uncovering of new insights in spatial epigenetic priming, differentiation and gene regulation within the tissue architecture. These technologies are of great interest in life science and biomedical research.

Stacking fault energy (SFE) is an essential parameter for characterizing mechanical properties. However, in high entropy alloys (HEAs), the local chemical environment varies significantly across different stacking fault planes, resulting in a substantial fluctuation of SFE values rather than a unique value, which prohibits the prediction of the local SFE. Herein, we proposed an effective descriptor based on the local concentration ratio near stacking fault to quantitatively predict the local SFE of refractory HEAs. We find that the role of a given element in determining SFE strongly depends on its valence-electron number relative to other components and the contribution of its s- and d-electrons to its cohesive properties, which can be understood in the framework of the tight-binding model. Notably, the descriptor not only unifies the local nature of SFE from simple alloys to HEAs but also helps to quickly design HEAs as the involved parameters are easily accessible., 20 pages,10 figures

IntroductionChina is one of the largest consumers of agricultural antibiotics in the world. While the Chinese government has been tightening its regulations to control antimicrobial resistance (AMR) from animal sources in recent years, the extent of antimicrobial oversight and the practices of antibiotic use in animal agriculture in China has not yet been explored. This study describes the practices of antimicrobial management in eastern China and current scenarios of antibiotic use in commercial farms and smallholder backyard farming.Methods33 semi-structured interviews were conducted with government agriculture officials, veterinary drug sellers, farmers and smallholders in two contrasting areas of rural Zhejiang and Jiangsu provinces, China. Interview transcripts were analyzed in NVivo12 using a thematic approach.ResultsFindings revealed that although the governance of antibiotic use has made progress, especially in controlling irrational antibiotic use in commercial farms, smallholders are under-regulated due to a lack of resources and assumptions about their marginal role as food safety governance targets. We also found that smallholders resort to human antibiotics for the treatment of backyard animals because of economic constraints and lack of access to professional veterinary services.DiscussionMore attention needs to be devoted to the local structural needs of farmers to reduce antibiotic misuse. Considering the extensive links of AMR exposure under the One Health framework, efforts to integrate smallholders in antibiotic governance are required to address the AMR burden systematically in China.

Additional file 1: Fig. S1. Mutations T323A, M324A and R325A impaired the MUN-catalyzed transition from Munc18-1/Syx1 to the SNARE complex. Representative native gel (upper) shown was from one of three replicates. Quantification of the integrated densities of Munc18-1/Syx1 bands is shown below the native gel. Data are presented as mean values ± SD, n = 3. Fig. S2. Peak elution volume analysis of Munc18-1 WT and mutants in the presence and absence of Syx1. All analyses were performed by size-exclusion chromatography on a Superdex 200 increase 10/300 GL column. Fig. S3. Expression of Munc18-1 TMR in cultured cortical neurons. (A) Quantification of Munc18-1 expression by western blot in neurons infected with control lentivirus (Control) or lentiviruses expressing Munc18-1 shRNA alone (None) or together with wild-type Munc18-1 (WT) or Munc18-1 TMR mutants. Representative results displayed are from one of three independent replicates. (B) Quantitative analysis of the integrated densities of Munc18-1/β-actin bands. The mean values ± SD are shown, n = 3. Fig. S4. The Munc18-1 TMR mutant does not alter synapse formation or the localization of Munc18-1 targeting to synapses in cultured cortical neurons. (A) Cultured cortical neurons infected with control lentivirus or with lentivirus expressing the Munc18-1 shRNA only (None), or together with wild-type Munc18-1 (WT) or the TMR mutant. Neurons were fixed and labeled by double immunofluorescence using Synapsin-1 (to mark synapses) and Munc18-1 antibodies along with fluorescently labeled secondary antibodies (see Materials and Methods). (A) Summary graphs of Synapsin-1-specific fluorescence intensities for all conditions as described above. (B) Summary graphs of ΔF (Fsynapse− Fnon-synapse) of Munc18-1-specific fluorescence intensity for all conditions as described above. Summary graphs of synapse number (C) and size (D), which were quantified from Synapsin-1-specific puncta. Data are presented as mean values ± SEM. Neurons analyzed are from three independent cultures. Numbers of cells/independent cultures analyzed are listed in the bars. Statistical assessments were performed by Student’s t-test comparing each condition to the indicated control experiment (****, P< 0.0001). Fig. S5. Interaction between Munc18-1/Syx1 and Syb2 in the presence of the MUN domain analyzed by western blotting. Immunoblotting data are shown with Munc18-1 monoclonal antibody and His-Tag monoclonal antibody to exhibit the binding of Munc18-1 and Syx1, respectively. Bands of Munc18-1 degradation are indicated with asterisks. Degradation is common in Munc18-1 and has no effect on Syx1 binding. Representative results displayed are from one of three independent replicates.

This paper is concerned with the problem of policy evaluation with linear function approximation in discounted infinite horizon Markov decision processes. We investigate the sample complexities required to guarantee a predefined estimation error of the best linear coefficients for two widely-used policy evaluation algorithms: the temporal difference (TD) learning algorithm and the two-timescale linear TD with gradient correction (TDC) algorithm. In both the on-policy setting, where observations are generated from the target policy, and the off-policy setting, where samples are drawn from a behavior policy potentially different from the target policy, we establish the first sample complexity bound with high-probability convergence guarantee that attains the optimal dependence on the tolerance level. We also exhihit an explicit dependence on problem-related quantities, and show in the on-policy setting that our upper bound matches the minimax lower bound on crucial problem parameters, including the choice of the feature maps and the problem dimension., Comment: The first two authors contributed equally

Text image machine translation (TIMT) has been widely used in various real-world applications, which translates source language texts in images into another target language sentence. Existing methods on TIMT are mainly divided into two categories: the recognition-then-translation pipeline model and the end-to-end model. However, how to transfer knowledge from the pipeline model into the end-to-end model remains an unsolved problem. In this paper, we propose a novel Multi-Teacher Knowledge Distillation (MTKD) method to effectively distillate knowledge into the end-to-end TIMT model from the pipeline model. Specifically, three teachers are utilized to improve the performance of the end-to-end TIMT model. The image encoder in the end-to-end TIMT model is optimized with the knowledge distillation guidance from the recognition teacher encoder, while the sequential encoder and decoder are improved by transferring knowledge from the translation sequential and decoder teacher models. Furthermore, both token and sentence-level knowledge distillations are incorporated to better boost the translation performance. Extensive experimental results show that our proposed MTKD effectively improves the text image translation performance and outperforms existing end-to-end and pipeline models with fewer parameters and less decoding time, illustrating that MTKD can take advantage of both pipeline and end-to-end models., Comment: Accepted at The 17th International Conference on Document Analysis and Recognition (ICDAR 2023)

Additional file 1: Supplementary Table 1. Basic data of moderate-severe OHSS group and matched Control

We propose $\textsf{ScaledGD($\lambda$)}$, a preconditioned gradient descent method to tackle the low-rank matrix sensing problem when the true rank is unknown, and when the matrix is possibly ill-conditioned. Using overparametrized factor representations, $\textsf{ScaledGD($\lambda$)}$ starts from a small random initialization, and proceeds by gradient descent with a specific form of damped preconditioning to combat bad curvatures induced by overparameterization and ill-conditioning. At the expense of light computational overhead incurred by preconditioners, $\textsf{ScaledGD($\lambda$)}$ is remarkably robust to ill-conditioning compared to vanilla gradient descent ($\textsf{GD}$) even with overprameterization. Specifically, we show that, under the Gaussian design, $\textsf{ScaledGD($\lambda$)}$ converges to the true low-rank matrix at a constant linear rate after a small number of iterations that scales only logarithmically with respect to the condition number and the problem dimension. This significantly improves over the convergence rate of vanilla $\textsf{GD}$ which suffers from a polynomial dependency on the condition number. Our work provides evidence on the power of preconditioning in accelerating the convergence without hurting generalization in overparameterized learning.

Additional file 2: Supplementary Table 2. Ovulation induction and transplant information of moderate-severe OHSS and control.

Additional file 4. Original images of confocal imaging for Fig. 3A and B.

Additional file 2. Uncropped images of gels and Western blots shown in this paper.

This paper proposes a novel metaheuristic optimizer, named Colony Search Optimization Algorithm (CSOA). The algorithm mimics the social behavior of early humans. Early humans expanded their settlements in search of more livable places to live. In CSOA, the worst solution is used to escape from local optima. And the number of these redundant solutions’ updates is reduced to improve the performance of the algorithm. CSOA is tested with 26 mathematical optimization problems and 4 classical engineering optimization problems. The optimization results are compared with those of various optimization algorithms. The experimental results show that the CSOA is able to provide very competitive results on most of the tested problems. Then, a new effective method is provided for solving optimization problems.

Cesarean scar pregnancy (CSP) is a rare form of ectopic pregnancy, which is a long-term complication of cesarean section. Prompt and accurate diagnosis of CSP is important to decrease maternal mobility and mortality. However, it is difficult to make an early detection for CSP complicated with morbidly adherent placenta. Contrast-enhanced ultrasound with the advantage in blood flow imaging is low-cost, time-saving, safe and more accessible in clinical practice. Here, we report a case with early detection of CSP with placenta increta by contrast-enhanced ultrasound and its successful uterine-sparing surgical management.

We prove that optimistic-follow-the-regularized-leader (OFTRL), together with smooth value updates, finds an $O(T^{-1})$-approximate Nash equilibrium in $T$ iterations for two-player zero-sum Markov games with full information. This improves the $\tilde{O}(T^{-5/6})$ convergence rate recently shown in the paper Zhang et al (2022). The refined analysis hinges on two essential ingredients. First, the sum of the regrets of the two players, though not necessarily non-negative as in normal-form games, is approximately non-negative in Markov games. This property allows us to bound the second-order path lengths of the learning dynamics. Second, we prove a tighter algebraic inequality regarding the weights deployed by OFTRL that shaves an extra $\log T$ factor. This crucial improvement enables the inductive analysis that leads to the final $O(T^{-1})$ rate., Accepted to ICLR 2023

End-to-end text image translation (TIT), which aims at translating the source language embedded in images to the target language, has attracted intensive attention in recent research. However, data sparsity limits the performance of end-to-end text image translation. Multi-task learning is a non-trivial way to alleviate this problem via exploring knowledge from complementary related tasks. In this paper, we propose a novel text translation enhanced text image translation, which trains the end-to-end model with text translation as an auxiliary task. By sharing model parameters and multi-task training, our model is able to take full advantage of easily-available large-scale text parallel corpus. Extensive experimental results show our proposed method outperforms existing end-to-end methods, and the joint multi-task learning with both text translation and recognition tasks achieves better results, proving translation and recognition auxiliary tasks are complementary., Accepted at the 26TH International Conference on Pattern Recognition (ICPR 2022)

Jun Fang Chen,1 Xue Mei Zhang,1,2 Kiran Sanjel,1,2 Juan Zhang,2 Cong Ma1 1The Dermatology and Venereology Department, The Affiliated Hospital of Inner Mongolia Minzu University, Tongliao, People’s Republic of China; 2Clinical Medical School, Inner Mongolia Minzu University, Tongliao, People’s Republic of ChinaCorrespondence: Xue Mei Zhang, The Dermatology and Venereology Department, The Affiliated Hospital of Inner Mongolia Minzu University, Tongliao, Inner Mongolia, 028000, People’s Republic of China, Tel +8618504757225, Email tongliaozhangxuemei@126.comPurpose: The purpose of this study was to explore the expression of TNF-α and NF-κB/p65 in Lichen planus skin lesions and their correlation with the pathogenesis of Lichen planus.Patients and Methods: The case group consisted of 30 individuals diagnosed with LP based on clinical and histopathologic examination. The control group consisted of 10 individuals from an Orthopedic Department with normal skin. TNF-α and NF-κB/p65 expression in skin tissue samples was detected by immunohistochemistry.Results: The immunohistochemical results showed that TNF-α and NF-κB/p65 expression levels were significantly higher in LP skin lesions than normal skin tissues (P ≤ 0.05). Positive TNF-α staining mainly occurred in the cytoplasm of keratinocytes of the stratum granulosum, stratum spinosum, and stratum basale in the epidermis and lymphocytes in the superficial dermis. Positive NF-κB/p65 staining mainly occurred in the nucleus and cytoplasm of keratinocytes of the stratum spinosum and stratum basale in the epidermis and lymphocytes in the superficial dermis.Conclusion: TNF-α and NF-κB/p65 are overexpressed in cutaneous LP. The two are positively correlated in LP, suggesting that they both play essential roles in the pathogenesis of LP.Keywords: lichen planus, TNF-α, NF-κB/p65, staining, immunohistochemistry

We present a family $\{\hat{\pi}\}_{p\ge 1}$ of pessimistic learning rules for offline learning of linear contextual bandits, relying on confidence sets with respect to different $\ell_p$ norms, where $\hat{\pi}_2$ corresponds to Bellman-consistent pessimism (BCP), while $\hat{\pi}_\infty$ is a novel generalization of lower confidence bound (LCB) to the linear setting. We show that the novel $\hat{\pi}_\infty$ learning rule is, in a sense, adaptively optimal, as it achieves the minimax performance (up to log factors) against all $\ell_q$-constrained problems, and as such it strictly dominates all other predictors in the family, including $\hat{\pi}_2$., Comment: Accepted to NeurIPS 2022

We study the covariate shift problem in the context of nonparametric regression over a reproducing kernel Hilbert space (RKHS). We focus on two natural families of covariate shift problems defined using the likelihood ratios between the source and target distributions. When the likelihood ratios are uniformly bounded, we prove that the kernel ridge regression (KRR) estimator with a carefully chosen regularization parameter is minimax rate-optimal (up to a log factor) for a large family of RKHSs with regular kernel eigenvalues. Interestingly, KRR does not require full knowledge of likelihood ratios apart from an upper bound on them. In striking contrast to the standard statistical setting without covariate shift, we also demonstrate that a naive estimator, which minimizes the empirical risk over the function class, is strictly sub-optimal under covariate shift as compared to KRR. We then address the larger class of covariate shift problems where the likelihood ratio is possibly unbounded yet has a finite second moment. Here, we propose a reweighted KRR estimator that weights samples based on a careful truncation of the likelihood ratios. Again, we are able to show that this estimator is minimax rate-optimal, up to logarithmic factors., to appear in the Annals of Statistics

1H nucleus is the most commonly used nucleus for magnetic resonance imaging (MRI) due to its high natural abundance in organisms and high MR sensitivity. However, other nuclei are distinctive in physiological processes, such as 31P that is important in the energy metabolisms. The development of high-field MRI allows for MRI research based on 31P nucleus, while high-quality 1H/31P dual-tuned radio frequency (RF) coil is essential for such applications. This article summarized the researches and applications of various 1H/31P dual-tuned RF coils, and presented in vivo 1H MRI and 31P MR spectra of mouse brain at 9.4 T. Potential applications of 1H/31P dual-tuned RF coil for high field were discussed.

In high-reliability applications, DC/DC converters should realize high-efficiency and fast-transient in harsh environments. Thus, the circuit parameter design and the reliability design are both critical. The secondary-side control circuit has a faster transient and a simpler gate-drive than the primary-side control circuit, which has been used for over twenty years. However, the detailed design procedure and analysis of the high-reliability secondary-side control circuit have not been presented in the literature. This article introduces a secondary-side control circuit design and its small-signal modeling procedure with a peak-current-mode-control (PCMC) forward converter. The reliability design and evaluation for thick-film converter are demonstrated. A 100-watt prototype is manufactured, and its steady-state waveforms and transient waveforms are tested and then compared with other high-reliability products. The experimental results show that the peak efficiency reaches 90.6%; the load transient response is 330mV/ $560~\mu \text{s}$ ; and the mean-time-between-failure (MTBF) value is 481 kilo-hours, proving the fast-transient and high-reliability features of the secondary-side control circuit.

Fully Field Programmable Gate Array (FPGA)based digitizer for high-resolution time and energy measurement is an attractive low cost solution for the readout electronics in positron emission computed tomography (PET)detector. In recent years, the FPGA based time-digital converter (FPGA-TDC) has been widely used for time measurement in the commercial PET scanners. Yet, for the energy measurement, few studies have been reported on a fully FPGA based, large dynamic range and high resolution alternative to the commercial analog-digital converter (ADC). Our previous research presents a 25 Ms/s FPGA-TDC based free-running ADC (FPGA-ADC), and successfully employed it in the readout electronics for PET detector. In this work-in-progress study, by means of the time-interleaved strategy, a 50 Ms/s FPGA-ADC is presented. With only two off-chip resistors, both the A/D conversion and energy measurement are achieved on a Xilinx Kintex-7 FPGA. Therefore, this method has great advantages inimproving system integration. Initial performance tests are also presented, and we hope it can give us a possibility to develop a new FPGA-only front-end digitizer for PET in future., 4 pages, 5 figures, 1 table. arXiv admin note: substantial text overlap with arXiv:2108.08002

An increasing number of data science and machine learning problems rely on computation with tensors, which better capture the multi-way relationships and interactions of data than matrices. When tapping into this critical advantage, a key challenge is to develop computationally efficient and provably correct algorithms for extracting useful information from tensor data that are simultaneously robust to corruptions and ill-conditioning. This paper tackles tensor robust principal component analysis (RPCA), which aims to recover a low-rank tensor from its observations contaminated by sparse corruptions, under the Tucker decomposition. To minimize the computation and memory footprints, we propose to directly recover the low-dimensional tensor factors -- starting from a tailored spectral initialization -- via scaled gradient descent (ScaledGD), coupled with an iteration-varying thresholding operation to adaptively remove the impact of corruptions. Theoretically, we establish that the proposed algorithm converges linearly to the true low-rank tensor at a constant rate that is independent with its condition number, as long as the level of corruptions is not too large. Empirically, we demonstrate that the proposed algorithm achieves better and more scalable performance than state-of-the-art matrix and tensor RPCA algorithms through synthetic experiments and real-world applications.

I introduce the audience to the intrigues of floating-point arithmetic and provide my own commentary. The lecture begins with an excursion into music theory and its history, by explaining the impressive feat of calculating the square root of 2 to the 25th decimal place for the first time on the abacus, motivated by its use in early modern music. No familiarity with music theory is required; we use Python snippets and modern math notation for everything. We show the computational cost necessary for implement it (on the abacus or computer). Then we visit some modern examples in computing that are directly relevant to the audience's experiences in manipulating geometric entities as floating-point numerical data. Again, using Python for numerical experiments, we demonstrate some precision-enhacing techniques. In the end, I offer some personal takes on the meaning of it all, and treat the audience with some memorable quotes ranging from William Kahan to Dostoevsky. The Appendix contains a summary of "Platonic" vs. "existentialist" styles of error assessment.

I introduce the audience to the intrigues of floating-point arithmetic and provide my own commentary. The lecture begins with an excursion into music theory and its history, by explaining the impressive feat of calculating the square root of 2 to the 25th decimal place for the first time on the abacus, motivated by its use in early modern music. No familiarity with music theory is required; we use Python snippets and modern math notation for everything. We show the computational cost necessary for implement it (on the abacus or computer). Then we visit some modern examples in computing that are directly relevant to the audience's experiences in manipulating geometric entities as floating-point numerical data. Again, using Python for numerical experiments, we demonstrate some precision-enhacing techniques. In the end, I offer some personal takes on the meaning of it all, and treat the audience with some memorable quotes ranging from William Kahan to Dostoevsky. The Appendix contains a summary of "Platonic" vs. "existentialist" styles of error assessment.

This paper is concerned with noisy matrix completion--the problem of recovering a low-rank matrix from partial and noisy entries. Under uniform sampling and incoherence assumptions, we prove that a tuning-free square-root matrix completion estimator (square-root MC) achieves optimal statistical performance for solving the noisy matrix completion problem. Similar to the square-root Lasso estimator in high-dimensional linear regression, square-root MC does not rely on the knowledge of the size of the noise. While solving square-root MC is a convex program, our statistical analysis of square-root MC hinges on its intimate connections to a nonconvex rank-constrained estimator., Comment: Accepted to IEEE Transactions on Information Theory

Reinforcement learning (RL) provides a theoretical framework for continuously improving an agent's behavior via trial and error. However, efficiently learning policies from scratch can be very difficult, particularly for tasks with exploration challenges. In such settings, it might be desirable to initialize RL with an existing policy, offline data, or demonstrations. However, naively performing such initialization in RL often works poorly, especially for value-based methods. In this paper, we present a meta algorithm that can use offline data, demonstrations, or a pre-existing policy to initialize an RL policy, and is compatible with any RL approach. In particular, we propose Jump-Start Reinforcement Learning (JSRL), an algorithm that employs two policies to solve tasks: a guide-policy, and an exploration-policy. By using the guide-policy to form a curriculum of starting states for the exploration-policy, we are able to efficiently improve performance on a set of simulated robotic tasks. We show via experiments that JSRL is able to significantly outperform existing imitation and reinforcement learning algorithms, particularly in the small-data regime. In addition, we provide an upper bound on the sample complexity of JSRL and show that with the help of a guide-policy, one can improve the sample complexity for non-optimism exploration methods from exponential in horizon to polynomial., Comment: 20 pages, 10 figures

We study covariate shift in the context of nonparametric regression. We introduce a new measure of distribution mismatch between the source and target distributions that is based on the integrated ratio of probabilities of balls at a given radius. We use the scaling of this measure with respect to the radius to characterize the minimax rate of estimation over a family of H\"older continuous functions under covariate shift. In comparison to the recently proposed notion of transfer exponent, this measure leads to a sharper rate of convergence and is more fine-grained. We accompany our theory with concrete instances of covariate shift that illustrate this sharp difference., Comment: 22 pages, 2 figures, 1 table

Background Osteoporotic vertebral fractures are often clinically silent and unrecognized. The present study aimed to determine whether routine chest radiographs could be a potential screening tool for identifying missed vertebral fractures in men aged over 50 years or postmenopausal women, especially those with type 2 diabetes mellitus (T2DM). In this study, we aimed to determine the prevalence of undetected vertebral fractures in elderly Chinese patients with and without T2DM. Methods Clinical data and chest radiographs of 567 individuals with T2DM (T2DM group) and 583 without diabetes (nondiabetic group) at a tertiary hospital in central south China were extracted from the records. Vertebral fractures were specifically looked for on chest radiographs and classified using the Genant semi-quantitative scale. Prevalence was compared between the two groups. Results Mean age and sex composition were comparable between the two groups. Mean weight and body mass index were significantly lower in the T2DM group. In both groups, fractures mostly involved the T11–12 and L1 vertebrae. Moderate/severe fractures were identified in 33.3% individuals in the T2DM group (31.4% men and 36.0% women) versus 23.2% individuals (20.9% men and 25.5% women) in the nondiabetic group. Conclusions Routine chest radiographs could be a useful screening tool for identifying asymptomatic vertebral fractures. Trial registration The study was designed as an observational retrospective study, therefore a trial registration was not necessary.

DC Transformer (DCX) is widely used because of its high efficiency and tight conversion ratio. Although the traditional cascade converter can regulated output voltage, it causes a huge loss (over 10% efficiency). The proposed topology outputs regulated voltage with high efficiency and wide input voltage. It consists of a push-pull DCX converter in parallel with a voltage compensation circuit. The compensation circuit only processes part of input power with low stress and low loss; thus, the total efficiency is improved. Two compensation circuits are compared and discussed, then the small-signal model and the control-to-output transfer function are derived. A design example is presented, and the simulation and calculation results verify the theory. The results show that the secondary-side compensation circuit has high efficiency (92.5%), wide input voltage range (±43%), and the efficiency penalty caused by voltage compensation is less than 2.5%.

The aim of this study is to design and evaluate a simple free running Analog-Digital Converter (ADC) based on the Field Programmable Gate Array (FPGA) device to accomplish the energy and position readout of the silicon photomultiplier (SiPM) array for application as PET scanners. This simple FPGA-ADC based on a carry chain Time-Digital Converter (TDC) implemented on a Kintex-7 FPGA consists of only one off-chip resistor so it has greater advantages in improving system integration and reducing cost than commercial chips. In this paper, a FPGA-ADC based front-end electronics prototype is presented, and both the design principle and implementation considerations are discussed. Experiments were performed using an 8 x 8 (crystal size: 4 x 4 x 15 mm3 ) and a 12 x 12 (crystal size: 2.65 x2.65 x 15 mm3 ) segmented LYSO crystals coupled with an 8 x 8 SiPM (Jseries, from ON Semiconductor) array which is under 22Na point source excitation. Initial test results indicate that the energy resolution of the two detectors after correction is around 13.2% and 13.5 % at 511 keV, and the profiles of the flood histograms show a clear visualization of the discrete scintillator element. All measurements were carried out at room temperature (~25 degree), without additional cooling, 9 pages,14 figures

Seasonal algal bloom, a water supply issue worldwide, can be efficiently solved by membrane technology. However, membranes typically suffer from serious fouling, which hinders the wide application of this technology. In this study, the feasibility of adding Ca2+ to control membrane fouling in coagulation–membrane treatment of algal-rich water was investigated. According to the results obtained, the normalized membrane flux decreased by a lower extent upon increasing the concentration of Ca2+ from 0 to 10 mmol/L. Simultaneously, the floc particle size increased significantly with the concentration of Ca2+, which leads to a lower hydraulic resistance. The coagulation performance is also enhanced with the concentration of Ca2+, inducing a slight osmotic pressure-induced resistance. The formation of Ca2+ coagulation flocs resulted in a looser, thin, and permeable cake layer on the membrane surface. This cake layer rejected organic pollutants and could be easily removed by physical and chemical cleaning treatments, as revealed by scanning electron microscopy images. The hydraulic irreversible membrane resistance was significantly reduced upon addition of Ca2+. All these findings suggest that the addition of Ca2+ may provide a simple-operation, cost-effective, and environmentally friendly technology for controlling membrane fouling during coagulation–membrane process for algal-rich water treatment.

To investigate the expression and sources of high mobility group box 1 (HMGB1) protein in the maternal–fetal interface of patients with unexplained recurrent spontaneous abortion (URSA), and further to verify the role of HMGB1 in the etiology of URSA. 55 women at early pregnancy with URSA and 55 women undergoing selective termination of normal early pregnancy as control were included. The abortion tissues including villi and decidua were collected. The expression of HMGB1, CD45, CK7, and vimentin in abortion tissues was detected, and the localization and sources of HMGB1 were analyzed. Infiltrating immune cells and expression of HMGB1 were significantly increased in villi and decidua in URSA group compared with those in the control group. In the URSA group, HMGB1 was colocalized with the CD45-labeled immune cells, and it was more obvious in decidua than in villi; in addition, HMGB1 was colocalized with the vimentin-labeled decidual stromal cells, but not with the CK7- labeled villous epithelial cells. High expression of HMGB1 in the maternal–fetal interface in URSA patients was actively secreted by the infiltrating immune cells, and decidual stromal cells may passively release HMGB1 during necrosis.

TANet is one of state-of-the-art 3D object detection method on KITTI and JRDB benchmark, the network contains a Triple Attention module and Coarse-to-Fine Regression module to improve the robustness and accuracy of 3D Detection. However, since the original input data (point clouds) contains a lot of noise during collecting the data, which will further affect the training of the model. For example, the object is far from the robot, the sensor is difficult to obtain enough pointcloud. If the objects only contains few point clouds, and the samples are fed into model with the normal samples together during training, the detector will be difficult to distinguish the individual with few pointcloud belong to object or background. In this paper, we propose TANet++ to improve the performance on 3D Detection, which adopt a novel training strategy on training the TANet. In order to reduce the negative impact by the weak samples, the training strategy previously filtered the training data, and then the TANet++ is trained by the rest of data. The experimental results shows that AP score of TANet++ is 8.98 higher than TANet on JRDB benchmark., Comment: 3pages, 2 figures

Additional file 1. More details on FlipFlop algorithm. This additional file provides more explanation on the algorithm of FlipFlop and why it cannot handle variable-length sequencing fragments. It also includes the figure showing the edge abundance estimation of LPAR1 gene in the other BD and control samples.

Bin Liang, 1,* Hua Xiang, 2,* Cong Ma, 3 Bin Xiong, 1 Yilong Ma, 4 Chang Zhao, 4 Yuanhui Yao, 2 Zishu Zhang, 3 Changyong Chen, 5 Haiping Li, 5 Qingyun Long, 6 Jun Zhou, 6 Chao Luo, 7 Huaiming Qiu, 7 Hongyao Hu, 8 Hui Zhao, 8 Guofeng Zhou, 1 Chuansheng Zheng 1 1Department of Radiology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, People’s Republic of China; 2Department of Interventional Radiology, Hunan Provincial People’s Hospital, Changsha, People’s Republic of China; 3Department of Radiology, The Second Xiangya Hospital of Central South University, Changsha, People’s Republic of China; 4Department of Interventional Radiology, Cancer Hospital Affiliated to Guangxi Medical University, Nanning, People’s Republic of China; 5Department of Radiology, Xiangya Hospital General South University, Changsha, People’s Republic of China; 6Department of Interventional Radiology, Zhongnan Hospital of Wuhan University, Wuhan, People’s Republic of China; 7Department of Radiology, Wuhan General Hospital of Guangzhou Military Region, Wuhan, People’s Republic of China; 8Department of Interventional Radiology, Renmin Hospital of Wuhan University, Hubei General Hospital, Wuhan, People’s Republic of China*These authors contributed equally to this workCorrespondence: Chuansheng ZhengDepartment of Radiology, Hubei Key Laboratory of Molecular Imaging, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1277 Jiefang Road, Wuhan 430022, People’s Republic of ChinaEmail hqzcsxh@sina.com Guofeng ZhouDepartment of Radiology, Hubei Key Laboratory of Molecular Imaging, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430022, People’s Republic of ChinaEmail zhouguofeng69@126.comPurpose: This study aimed to compare the efficacy and safety between transarterial chemoembolization (TACE) with CalliSpheres® microspheres (CSM-TACE) and conventional TACE (cTACE) in patients with hepatocellular carcinoma (HCC).Patients and Methods: Three hundred and thirty-five HCC patients receiving CSM-TACE or cTACE were consecutively enrolled in this multi-center, retrospective cohort study, and then divided into CSM-TACE group and cTACE group accordingly. Complete response (CR), objective response (ORR) and disease control response (DCR) was assessed according to mRECIST criteria at 1 month (M1), 3 months(M3)and 6 months(M6) after treatment. Progression-free survival (PFS) and overall survival (OS) were assessed. Liver function indexes and adverse events (AEs) were also evaluated.Results: CR at M3 (P=0.020) and ORR at M1 (P< 0.001), M3 (P< 0.001) and M6 (P=0.017) after treatment were significantly higher in the CSM-TACE compared with cTACE group. DCRs, PFS (25.3 months vs 24.2 months, P=0.503) and OS (27.8 months vs 25.3 months, P=0.203) were similar between the two groups. CSM-TACE was independently correlated with higher ORR at M1 (P=0.002) and longer OS (P=0.023). Abnormal alkaline phosphatase (ALP) (P=0.049) was independently associated with lower ORR at M3, and history of alcohol intake (P=0.019) and largest nodule size ≥ 7 cm (P=0.015) independently correlated with lower ORR at M6 (P=0.015). Largest nodule size ≥ 7 cm (P=0.029) and abnormal albumin (ALB) (P=0.046) were independently associated with shorter PFS. Child-Pugh stage B/C (P=0.023), abnormal ALB (P=0.001), ALP (P=0.008) and alpha-fetoprotein (AFP) (P=0.005) were independently associated with shorter OS. Most liver function indexes and AEs were similar between the two groups (P> 0.05), except that ALP (P=0.005), total bilirubin (P=0.031), pain during procedure (P=0.034) and occurrence of fever post(treatment (P=0.017) were significantly elevated in the CSM-TACE compared with cTACE group.Conclusion: CSM-TACE presents with a better treatment response and similar survival profile compared with cTACE in HCC patients.Keywords: chemoembolization, CalliSpheres microspheres, hepatocellular carcinoma, treatment response, survival, predictive factor

A green bond is a type of fixed-income security that raises money to invest in predetermined climate and environmental projects, in contrast to conventional debt instruments, where the use of proceeds is not specified in the terms. The difference in yield between a green bond and an otherwise identical non-green bond of the same issuer and with the same terms is called the greenium. In this paper, we investigate this yield differential between green and conventional bonds. We estimate the greenium on the basis of the bond's asset swap spread (ASW) to investigate whether, consistent with a non-pecuniary motive for holding green assets, green labels are associated with a negative or positive yield gap with respect to ordinary bonds. We calculate and compare several descriptive statistics of green bonds and conventional bonds. Then, several statistical tests are implemented to analyze potential statistical differences between their return distributions. In our analysis, synthetic non-green bonds are constructed via interpolation of the ASW curve of non-green bonds. There are several findings: (1) From a statistical point of view, no difference between the overall distribution, the mean or median of ASW changes is detected on individual bond pairs. However, our estimation of an overall greenium exhibits a level uctuating near zero over time with an overall average around -7 bps. (2) In addition, we see indications that the volatility of some green-bonds is lower than their non-green counterparts. (3) We see a lagging effect between the greenium and stress in financial markets. This could indicate that sustainable investments like green bonds are potentially more immune to systemic crises.

Time-digital Converter (TDC) aims to measure the arrival time of the leading edge of the pulse signal. Our recent work presented a high resolution multi-phase TDC based on the Kintex-7 Field Programmable Gate Array (FPGA) device. A simple I/O tile based circular input buffer is employed to oscillate the input signal periodically, and then a multi-phase TDC based on ISERDES core with a 625 ps bin size is used to accomplish the multiple measurements for getting higher resolution performance. In this paper, the design concept, architecture, as well as kernel implementation considerations are all discussed. To evaluate the TDC's performance, we built a verification system based on Kintex-7 FPGA. Initial test results indicate that the TDC's effective bin size is successfully reduced from 625 ps to 78.125 ps, and the measured dual-channel time resolution is better than 35 ps RMS., Comment: 10 pages,9 figures,1 table,accepted by Journal of Instrumentation,2020

Additional file 2. Supplementary document on distance considerations. An extended discussion on the false positives and false negatives that may be due to lack of consideration about various distances in the genome.

We aimed to compare the treatment response, survivals and safety of drug-eluting bead (DEB) transarterial chemoembolization (TACE) with CalliSpheres(®) microspheres (CSM) and conventional TACE (cTACE) as first-line treatment in Chinese HCC patients. 192 HCC patients from multiple centers received DEB-TACE with CSM or cTACE treatment as first-line treatment were included and assigned to DEB-TACE group (N=94) or cTACE group (N=98) accordingly. Treatment response was assessed at 1 month (M1), M3 and M6 after treatment. Progression-free survival (PFS) and overall survival (OS) was evaluated. Liver function indexes and adverse events were recorded. Complete response (CR) and objective response rate (ORR) were higher, while disease control rate (DCR) rate was similar in DEB-TACE group compared with cTACE group, and further multivariate logistic regression analysis validated that DEB-TACE vs cTACE independently predicted higher ORR. For survivals, no difference in PFS or OS was observed between DEB-TACE and cTACE groups, and multivariate Cox’s proportional hazards regression revealed that DEB-TACE vs cTACE was not correlated with PFS or OS either. Additionally, no difference in liver function indexes at M1 or changes of liver function indexes from M0 to M1 between DEB-TACE and cTACE groups after treatment was observed, whereas DEB-TACE resulted in higher incidence of pain and fever during treatment or hospitalization. DEB-TACE with CSM discloses better treatment response, similar survival profiles and equal liver function injury but increased incidence of short-term adverse events than cTACE as the first-line therapy in treating HCC patients.

Current expression quantification methods suffer from a fundamental but under-characterized type of error: the most likely estimates for transcript abundances are not unique. Current quantification methods rely on probabilistic models, and the scenario where it admits multiple optimal solutions is called non-identifiability. This means multiple estimates of transcript abundances generate the observed RNA-seq reads equally likely, and the underlying true expression cannot be determined. The non-identifiability problem is further exacerbated when incompleteness of reference transcriptome and existence of unannotated transcripts are taken into consideration. State-of-the-art quantification methods usually output a single inferred set of abundances, and the accuracy of the single expression set is unknown compared to other equally optimal solutions. Obtaining the set of equally optimal solutions is necessary for evaluating and extending downstream analyses to take non-identifiability into account. We propose methods to compute the range of equally optimal estimates for the expression of each transcript, accounting for non-identifiability of the quantification model using several novel graph theoretical approaches. It works under two scenarios, one assuming the reference transcriptome is complete, another assuming incomplete reference and allowing for expression of unannotated transcripts. Our methods calculate a “confidence” range for each transcript, representing its possible abundance across equally optimal estimates. This range can be used for evaluating the reliability of detected differentially expressed (DE) transcripts, as a large overlap of confidence range between DE conditions indicates the prediction may be unreliable due to uncertainty. We observe that 5 out of 257 DE predictions are unreliable on an MCF10 cell line and 19 out of 3152 are unreliable on a CD8 T cell dataset. The source code can be found at https://github.com/Kingsford-Group/subgraphquant .

With the scale of grid-tied renewable energy generation systems increasing significantly, the grid stability and safety have to be ensured by new strict grid code. The low voltage ride through (LVRT) regulation which requires the renewable energy generation systems should connect and provide reactive power to support during grid fault and the islanding detection which ensures grid safety are two focuses in the grid code of many countries. This paper proposes a power perturbation method, which consists of dual second order generalized integrator (DSOGI) PLL, power perturbation module and islanding detection module. This method can meet the requirement of LVRT and islanding detection simultaneously by adjusting proportion of active power current and reactive power current generated by grid-tied renewable energy generation systems. Finally, validity and accuracy of the proposed method was verified by simulation results.