Density ratio estimation (DRE) is a paramount task in machine learning, for its broad applications across multiple domains, such as covariate shift adaptation, causal inference, independence tests and beyond. Parametric methods for estimating the density ratio possibly lead to biased results if models are misspecified, while conventional non-parametric methods suffer from the curse of dimensionality when the dimension of data is large. To address these challenges, in this paper, we propose a novel approach for DRE based on the projection pursuit (PP) approximation. The proposed method leverages PP to mitigate the impact of high dimensionality while retaining the model flexibility needed for the accuracy of DRE. We establish the consistency and the convergence rate for the proposed estimator. Experimental results demonstrate that our proposed method outperforms existing alternatives in various applications.

The capability of video super-resolution (VSR) to synthesize high-resolution (HR) video from ideal datasets has been demonstrated in many works. However, applying the VSR model to real-world video with unknown and complex degradation remains a challenging task. First, existing degradation metrics in most VSR methods are not able to effectively simulate real-world noise and blur. On the contrary, simple combinations of classical degradation are used for real-world noise modeling, which led to the VSR model often being violated by out-of-distribution noise. Second, many SR models focus on noise simulation and transfer. Nevertheless, the sampled noise is monotonous and limited. To address the aforementioned problems, we propose a Negatives augmentation strategy for generalized noise modeling in Video Super-Resolution (NegVSR) task. Specifically, we first propose sequential noise generation toward real-world data to extract practical noise sequences. Then, the degeneration domain is widely expanded by negative augmentation to build up various yet challenging real-world noise sets. We further propose the augmented negative guidance loss to learn robust features among augmented negatives effectively. Extensive experiments on real-world datasets (e.g., VideoLQ and FLIR) show that our method outperforms state-of-the-art methods with clear margins, especially in visual quality. Project page is available at: https://negvsr.github.io/.

Siyuan Chen, Meilin Wang, Minghao Lv, Zhiling Zhang, Qianqian Ju, Dejiyangla, Yujia Peng, Kenny Q. Zhu, Mengyue Wu. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.

Deep learning technologies have demonstrated their effectiveness in removing cloud cover from optical remote-sensing images. Convolutional Neural Networks (CNNs) exert dominance in the cloud removal tasks. However, constrained by the inherent limitations of convolutional operations, CNNs can address only a modest fraction of cloud occlusion. In recent years, diffusion models have achieved state-of-the-art (SOTA) proficiency in image generation and reconstruction due to their formidable generative capabilities. Inspired by the rapid development of diffusion models, we first present an iterative diffusion process for cloud removal (IDF-CR), which exhibits a strong generative capabilities to achieve component divide-and-conquer cloud removal. IDF-CR consists of a pixel space cloud removal module (Pixel-CR) and a latent space iterative noise diffusion network (IND). Specifically, IDF-CR is divided into two-stage models that address pixel space and latent space. The two-stage model facilitates a strategic transition from preliminary cloud reduction to meticulous detail refinement. In the pixel space stage, Pixel-CR initiates the processing of cloudy images, yielding a suboptimal cloud removal prior to providing the diffusion model with prior cloud removal knowledge. In the latent space stage, the diffusion model transforms low-quality cloud removal into high-quality clean output. We refine the Stable Diffusion by implementing ControlNet. In addition, an unsupervised iterative noise refinement (INR) module is introduced for diffusion model to optimize the distribution of the predicted noise, thereby enhancing advanced detail recovery. Our model performs best with other SOTA methods, including image reconstruction and optical remote-sensing cloud removal on the optical remote-sensing datasets.

SUMMARY: Colorectal cancer is a heterogeneous disease with diverse prognoses between left-sided and right-sided patients; therefore, it is necessary to precisely evaluate the survival probability of side-specific colorectal cancer patients. Here, we collected multi-omics data from The Cancer Genome Atlas program, including gene expression, DNA methylation and microRNA expression. Specificity measure and robust likelihood-based survival analysis were used to identify 6 left-sided and 28 right-sided prognostic biomarkers. Compared to the performance of clinical prognostic models, the addition of these biomarkers could significantly improve the discriminatory ability and calibration in predicting side-specific 5-year survival for colorectal cancer. Additional dataset derived from Gene Expression Omnibus was used to validate the prognostic value of side-specific genes. Finally, we constructed colorectal cancer side-specific molecular database (CoSMeD), a user-friendly interface for estimating side-specific colorectal cancer 5-year survival probability, which can lay the basis for personalized management of left-sided and right-sided colorectal cancer patients. AVAILABILITY AND IMPLEMENTATION: CoSMeD is freely available at https://mulongdu.shinyapps.io/cosmed. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Dimensionality reduction is of high importance in hyperspectral data processing, which can effectively reduce the data redundancy and computation time for improved classification accuracy. Band selection and feature extraction methods are two widely used dimensionality reduction techniques. By integrating the advantages of the band selection and feature extraction, the authors propose a new method for reducing the dimension of hyperspectral image data. First, a new and fast band selection algorithm is proposed for hyperspectral images based on an improved determinantal point process (DPP). To reduce the amount of calculation, the dual‐DPP is used for fast sampling representative pixels, followed by k‐nearest neighbour‐based local processing to explore more spatial information. These representative pixel points are used to construct multiple adjacency matrices to describe the correlation between bands based on mutual information. To further improve the classification accuracy, two‐dimensional singular spectrum analysis is used for feature extraction from the selected bands. Experiments show that the proposed method can select a low‐redundancy and representative band subset, where both data dimension and computation time can be reduced. Furthermore, it also shows that the proposed dimensionality reduction algorithm outperforms a number of state‐of‐the‐art methods in terms of classification accuracy.

Genome-wide association studies (GWASs) generally focus on a single marker, which limits the elucidation of the genetic architecture of complex traits. Herein, we present a new computational framework, termed probabilistic natural mapping (PALM), for performing gene-level association tests. PALM robustly reveals the inherent genomic structures of genes and generates feature representations that can be seamlessly incorporated into conventional statistic tests. Our approach substantially improves the effectiveness of uncovering associations derived from a subgroup of variants with weak effects, which represents a known challenge associated with existing methods. We applied PALM in a gastric cancer GWAS and identified two additional gastric cancer-associated susceptibility genes, NOC3L and RUNDC2A. The robust susceptibility discoveries of PALM are widely supported by existing studies from other biological perspectives. PALM will be useful for further GWAS analytical strategies that use gene-level analyses.

To improve the accuracy of analysis results is one of the hard challenges for static analysis. Especially, static analyzers generally analyze all paths of a program, including infeasible paths, which undoubtedly decreases the analysis accuracy. To mitigate the issue, we design and implement a static analyzer, called ABAZER-SE, which is based on the meta-compilation and the GCC abstract syntax tree. ABAZER-SE combines symbolic execution and static analysis techniques to detect bugs in the source code. In addition, it allows users to write a custom checker for a specific bug.

In this paper, an optimal design of a multibit interpolative sigma delta modulator (SDM) based on the absolute stability criterion is proposed. There are two types of midtread quantizers, namely, the uniform midtread quantizer and the nonuniform midtread quantizer. It is shown in this paper that the uniform midtread quantizer is the optimal one between these two types of the midtread quantizers in the sense of minimizing the maximum output input ratio of the midtread quantizer. A loop filter of the multibit interpolative SDM is designed based on the minimization of the energy of the noise transfer function in signal band subject to the strictly stable or the marginally stable condition of the loop filter, the absolute stability criterion, and the specifications on both the noise transfer function in the signal band and the magnitude response of the loop filter outside the signal band. Computer numerical simulation results show that our proposed multibit interpolative SDM achieves a broader stability margin and a higher signal-to-noise ratio compared to the state-of-art designs.

Many diseases can be detected via performing the electrocardiogram diagnosis. Although the electrocardiograms can be acquired in the public clinics, hospitals or even at homes by using the existing devices, these existing devices are not tiny enough to carry on the bodies. Hence, it is difficult to acquire the electrocardiograms all the time. In order to acquire the electrocardiograms all the time, this paper proposes a wearable electrocardiogram acquisition system. In particular, the R points of the electrocardiograms are detected by computing the regularities of the signals. If the regularity of a signal at a point v is about -1.5 and it is the minimum around its neighbor, then v is assumed to be the R point. Similarly, the Q points and the R points can be detected accordingly. Here, it is not required to compute the modulus maxima. Hence, the required computational power is very low. The simulation results show that the proposed algorithm outperforms the existing methods in terms of the required computational complexity.

This paper proposes a novel method of killing birds and applying vaccines for suppressing the spread of avian influenza via a nonlinear switching control approach. The switching strategy is based on the population of the susceptible birds and the population of the susceptible humans. There are four switching cases. For the first three switching cases, the elimination control force and the quarantine control force are equal to either zero or one. For the last switching case, they are equal to one minus a scalar divided by the population of the susceptible birds and one minus another scalar divided by the population of the susceptible humans, respectively. The system state vectors of the avian influenza model are guaranteed to reach the desirable equilibrium point. Also, the positivity requirements on the system states as well as the constraints on both the lower bounds and the upper bounds of both the elimination control force and the quarantine control force are guaranteed to be satisfied. Computer numerical simulation results show that the proposed control strategy is very effective and efficient.