Data augmentation is an intuitive solution to increase the diversity of training instances in the machine learning community. Mixup is acknowledged as an effective and efficient mix-based data augmentation method, following a linear alignment assumption that the linear interpolations of features align the corresponding linear interpolations of labels. Unfortunately, this assumption can be violated in many complex scenarios, resulting in augmented instances with noisy labels, especially for regression problems. To solve this problem, we propose an easy-to-implement mixup method, namely DEnosing MIXUP (DE-mixup), which iteratively corrects the noisy response targets by leveraging an auxiliary noise estimation task with mixup deep features. Additionally, we suggest an efficient optimization method with alternating direction method of multipliers. We compare DE-mixup with the existing mixup variants and other prevalent data augmentation methods across benchmark regression datasets. Empirical results indicate the effectiveness of DE-mixup under the in-distribution and out-of-distribution cases.

Positive-Unlabeled (PU) learning refers to a specific weakly-supervised learning paradigm that induces a binary classifier with a few positive labeled instances and massive unlabeled instances. To handle this task, the community has proposed dozens of PU learning methods with various techniques, demonstrating strong potential. In this paper, we conduct a comprehensive study to investigate the basic characteristics of current PU learning methods. We organize them into two fundamental families of PU learning, including *disambiguation-free empirical risks*, which approximate the expected risk of supervised learning, and *pseudo-labeling methods*, which estimate pseudo-labels for unlabeled instances. First, we make an empirical analysis on disambiguation-free empirical risks such as uPU, nnPU, and DistPU, and suggest a novel risk-consistent set-aware empirical risk from the perspective of aggregate supervision. Second, we make an empirical analysis of pseudo-labeling methods to evaluate the potential of pseudo-label estimation techniques and widely applied generic tricks in PU learning. Finally, based on those empirical findings, we propose a general framework of PU learning by integrating the set-aware empirical risk with pseudo-labeling. Compared with existing PU learning methods, the proposed framework can be a practical benchmark in PU learning.

Software security has drawn extensive attention as software projects have grown increasingly large and complex. Since the traditional manual or equipment vulnerability detection technology cannot meet today's software development needs, there is a recognized need to create more effective techniques to address security issues. Although various vulnerability detection systems have been proposed, most are based only on serialization or graph representation, to inadequate effect. We propose a system, MSGVUL, that provides superior vulnerability detection using a new multi-semantic approach. MSGVUL uses versatile and efficient code slicing employing a search algorithm based on sensitive data and functions and innovatively constructs an SSVEC model to fully integrate the semantic and structural information into the code. We also developed a novel BAG model, made up of BAP and PAG frameworks, that enables the hierarchical extraction of code vulnerability representations from the graph and sequence levels. The MSGVUL model is evaluated on slice-level and function-level vulnerability datasets, and the results demonstrate that the MSGVUL method outperforms other state-of-the-art methods.

Blind face restoration (BFR) aims to restore high-quality face images from inputs with complex degradation, which is key to extensive applications. Existing methods usually learn a black-box mapping to achieve the goal, which however often produce over-smoothed results. This work proposes a novel feature-guided framework via leveraging prior from a pre-trained Generative Adversarial Network (GAN) model to recover reasonable textures. Furthermore, we design a feature fusion module to guide the generator with low-level spatial content information in the degraded input, for the sake of holding the consistency on both facial structure and background details. The proposed method can be seamlessly integrated with a learned GAN through a simple yet effective principle to recover realistic results under complex degradation circumstances. Extensive comparisons demonstrate the superiority of our strategy over other state-of-the-art methods in terms of restoration quality and training cost.