Cross-platform binary code similarity detection aims at detecting whether two or more pieces of binary code are similar or not. Existing approaches that combine control flow graphs (CFGs)-based function representation and graph convolutional network (GCN)-based similarity analysis are the best-performing ones. Due to a large amount of convolutional computation and the loss of structural information, the use of convolution networks will inevitably bring problems such as high overhead and sometimes inaccuracy. To address these issues, we propose a fast cross-platform binary code similarity detection framework that takes advantage of natural language processing (NLP) and inductive graph neural network (GNN) for basic blocks embedding and function representation respectively by simulating extracting structural features and temporal features. GNN’s node-centric and small batch is a suitable training way for large CFGs, it can greatly reduce computational overhead. Various NLP basic block embedding models and GNNs are evaluated. Experimental results show that the scheme with long short term memory (LSTM) for basic blocks embedding and inductive learning-based GraphSAGE(GAE) for function representation outperforms the state-of-the-art works. In our framework, we can take only 45% overhead. Improve efficiency significantly with a small performance trade-off.

In practical abnormal traffic detection scenarios, traffic often appears as drift, imbalanced and rare labeled streams, and how to effectively identify malicious traffic in such complex situations has become a challenge for malicious traffic detection. Researchers have extensive studies on malicious traffic detection with single challenge, but the detection of complex traffic has not been widely noticed. Queried adaptive random forests (QARF) is proposed to detect traffic streams with concept drift, imbalance and lack of labeled instances. QARF is an online active learning based approach which combines adaptive random forests method and adaptive margin sampling strategy. QARF achieves querying a small number of instances from unlabeled traffic streams to obtain effective training. We conduct experiments using the NSL-KDD dataset to evaluate the performance of QARF. QARF is compared with other state-of-the-art methods. The experimental results show that QARF obtains 98.20% accuracy on the NSL-KDD dataset. QARF performs better than other state-of-the-art methods in comparisons.