We propose a new approach, called self-motivated pyramid curriculum domain adaptation (PyCDA), to facilitate the adaptation of semantic segmentation neural networks from synthetic source domains to real target domains. Our approach draws on an insight connecting two existing works: curriculum domain adaptation and self-training. Inspired by the former, PyCDA constructs a pyramid curriculum which contains various properties about the target domain. Those properties are mainly about the desired label distributions over the target domain images, image regions, and pixels. By enforcing the segmentation neural network to observe those properties, we can improve the network's generalization capability to the target domain. Motivated by the self-training, we infer this pyramid of properties by resorting to the semantic segmentation network itself. Unlike prior work, we do not need to maintain any additional models (e.g., logistic regression or discriminator networks) or to solve min max problems which are often difficult to optimize. We report state-of-the-art results for the adaptation from both GTAV and SYNTHIA to Cityscapes, two popular settings in unsupervised domain adaptation for semantic segmentation.

Trained with the standard cross entropy loss, deep neural networks can achieve great performance on correctly labeled data. However, if the training data is corrupted with label noise, deep models tend to overfit the noisy labels, thereby achieving poor generation performance. To remedy this issue, several loss functions have been proposed and demonstrated to be robust to label noise. Although most of the robust loss functions stem from Categorical Cross Entropy (CCE) loss, they fail to embody the intrinsic relationships between CCE and other loss functions. In this paper, we propose a general framework dubbed Taylor cross entropy loss to train deep models in the presence of label noise. Specifically, our framework enables to weight the extent of fitting the training labels by controlling the order of Taylor Series for CCE, hence it can be robust to label noise. In addition, our framework clearly reveals the intrinsic relationships between CCE and other loss functions, such as Mean Absolute Error (MAE) and Mean Squared Error (MSE). Moreover, we present a detailed theoretical analysis to certify the robustness of this framework. Extensive experimental results on benchmark datasets demonstrate that our proposed approach significantly outperforms the state-of-the-art counterparts.

No abstract

A botnet is one of the most grievous threats to network security since it can evolve into many attacks, such as Denial-of-Service (DoS), spam, and phishing. However, current detection methods are inefficient to identify unknown botnet. The high-speed network environment makes botnet detection more difficult. To solve these problems, we improve the progress of packet processing technologies such as New Application Programming Interface (NAPI) and zero copy and propose an efficient quasi-real-time intrusion detection system. Our work detects botnet using supervised machine learning approach under the high-speed network environment. Our contributions are summarized as follows: (1) Build a detection framework using PF_RING for sniffing and processing network traces to extract flow features dynamically. (2) Use random forest model to extract promising conversation features. (3) Analyze the performance of different classification algorithms. The proposed method is demonstrated by well-known CTU13 dataset and nonmalicious applications. The experimental results show our conversation-based detection approach can identify botnet with higher accuracy and lower false positive rate than flow-based approach.

No abstract