Log anomaly detection is a critical aspect of Artificial Intelligence for IT Operations (AIOps), as it enables the timely identification of system failures, thereby facilitating program understanding throughout the entire software maintenance and engineering life cycles. While existing methods only leverage raw log information for anomaly detection, they struggle to address challenges such as log differences due to log evolution, noise from log parsing, and stylistic differences between logs and natural language. To overcome these limitations, we propose GELog, an innovative log anomaly detection method. Specifically, GELog initially employs GPT to semantically enhance log templates. Subsequently, it extracts semantic vectors using pretrained sentence-bert and introduces an attention-based semantic fusion module that integrates the semantic representations of both the original and enhanced logs. Finally, GELog utilizes a Transformer-based model for anomaly detection. We evaluated the performance of GELog on four publicly available datasets, and the experimental results demonstrate that GELog significantly enhances the semantic representation of logs, achieving superior anomaly detection performance.

Android application markets face severe threats of malware attacks. Existing learning-based malware detection approaches rely on easily obfuscated features or unscalable sophisticated graph analysis techniques. In this paper, we propose DroidRadar, an accurate Android malware detection system based on lightweight graph embedding. The key insight of our method is constructing an entire Android application collection as a global graph schema and using sensitive APIs as bridge nodes to propagate inter-application information. We conduct statistical correlation analysis from different perspectives to model the application's usage pattern of sensitive APIs, then apply graph convolution network (GCN) to perform node embedding and malware detection. We evaluate DroidRadar on large scale datasets spanning nine years. Results show that DroidRadar has an average detection accuracy of 98.57% and a false-positive rate of 1.4 % on different time periods, which outperforms the state-of-the-art approaches, and it has strong robustness when detecting obfuscated malware variants.

The inter-symbol interference (ISI) caused by multipath propagation in wireless channel is one of the main reasons for high bit error rate (BER) in wireless communication system. Chaos was proved to be an ideal communication baseband signal because of its special properties, including the corresponding simple matched filter and multipath resistance ability. Although the ISI caused by the multipath is promised to be completely eliminated by using the optimal symbol decoding threshold, the future symbols are needed to be known for calculating the optimal threshold, which is hard to be practically implemented. To deal with such problem, an echo state network (ESN), because of its short-term memory ability, is proposed to predict the future chaotic baseband signal based on the signal received after the matched filter. Thus, the first future symbol is estimated, which is used to improve the accuracy of the decoding threshold in the chaotic baseband wireless communication system (CBWCS) in order to further relieve the effect of ISI. Compared to the sub-optimal decoding threshold considering only past symbols, the improved threshold proposed here considers not only the ISI from past symbols, but also the ISI from one future symbol. The simulation in both the static and time-varying wireless channels are performed, the results show that the BER performance of CBWCS is improved more significantly in multipath channels as compared to single path channel. The experiments based on wireless open-access research platform verify the ISI resistance performance under the practical scenarios. Both simulation and experimental results show the effectiveness and the superiority of the proposed method.

This paper deals with the blind source separation (BSS) problem in the overdetermined case, i.e. the number of receive antennas is larger than that of sources. First, a method, which converts the over-determined BSS into complete BSS, is presented. Then a maximum ratio combining (MRC) based BSS algorithm (MRC-FastICA) is proposed. The proposed algorithm includes source separation, estimated signals classification and MRC of multiple estimated signals corresponding to a source signal. Theoretical analysis and simulation results show that the proposed algorithm can obtain diversity gain, and has better performance than traditional methods.

Cooperative communications achieve tremendous improvements in system performance. Meanwhile, it tends to change the conventional access and transmission schemes in wireless networks. Thus, CoopMAC is proposed and analyzed for fully connected WLANs. It improves the performance of the network dramatically by enabling additional collaboration from other nodes. However, how does it work in multi-hop Ad Hoc networks is not well-understood yet. In this paper, we first propose an analytical model to evaluate the performance of CoopMAC for multi-hop Ad Hoc networks with consideration of hidden node problem. Our analysis results show that the cooperative transmissions are frequently interrupted by hidden nodes' interference. To mitigate the effect of hidden node, an enhanced CoopMAC (ECoopMAC) is further proposed without introducing any overhead and complexity, which jointly optimizes the order of handshake and the helper selection metric. By increasing the probability of successful cooperative transmission and decreasing the interference to other flows, the saturated throughput and access delay of multi-hop Ad Hoc networks are improved by ECoopMAC remarkably. Extensive simulations evaluate the performance of our analytical model and protocol, the results of analysis and simulation match well. Compared with CoopMAC, the saturated throughput is improved up to 12% on average.

Network-assisted diversity multiple access (NDMA) was recently developed to resolve collision and achieve high throughput. The blind NDMA method using independent component analysis (ICA), named ICA-NDMA, has been shown to overcome the difficulties of synchronization and orthogonal identification codes required by the training-based NDMA protocols. In this paper, we propose a novel blind collision resolution strategy by employing cooperative transmission in ICA-NDMA. When K-node collision happened, a cooperative transmission mechanism is adopted to collect different superposition of the K nodes' packets in the following slots. Moreover, a cooperative blind detection algorithm is intro- duced to determine the number of collision nodes. Finally, with ICA, the destination can retrieve the original packets through processing the collided packet and the signals forwarded by relays. Simulation results show that, the proposed method can work effectively under both fast-varying and slow-varying channels, the performances of average packet delay and maximum stable throughput are better than ICA-NDMA.

In this paper, we propose the optimal resource allocation strategies for energy constrained coded cooperative networks. By combining power control and multi-relays selection with LOC adjustment, our schemes aim at providing higher QoS and extending network lifetime. We consider the TDMA based scenario where one node acts as the source and the other nodes can be selected as relay nodes in a time slot and each node is limited by separate power constraint. Firstly, we build an optimization model for minimizing and balancing the energy consumption under QoS constrains. After that, when the LOC is constant, a closed form optimal solution is got by KKT conditions, which results in a dynamic resource allocation strategy. Moreover, based on solution for the fixed LOC, a suboptimal scheme is further proposed for a variable LOC via one dimension searching. The simulation results show that our schemes can improve network QoS and prolong network lifetime dramatically comparing with other existing cooperative resource allocation schemes.

Topology control via per-node transmission power adjustment has been shown effective in extending network lifetime. However, most of existing algorithms construct static topologies which fail to consider the residual energy of network nodes, thus in fact they can not balance energy consumption efficiently. To address this problem, a lightweight distributed topology control algorithm EDTC (Energy-aware Dynamic Topology Control) is proposed in this paper. Based on the link metric reflecting both the energy consumption rates and residual energy levels at the two end nodes, EDTC generates a dynamic network topology that changes with the variation of node energy. In addition, without the aid of location information, each node determines its transmission power according to local network information, which reduces the complexity and overhead of EDTC greatly. Theoretic analysis and experiment results show that EDTC preserves network connectivity and minimum-cost property and compared with other algorithms, it can extend network lifetime more remarkably.

In cognitive radio system, secondary users should have capability of detecting the signal from primary user to evade interfering harmfully with any primary receiver. The duration used for spectrum sensing influences the performance of secondary user system. In this paper, we show disadvantages of one threshold judgment in cognitive radio spectrum sensing. For improving this defect, we propose a double thresholds scheme and prove this scheme indeed could gain more throughputs under certain condition. Additionally, two instantiations are taken for general illustrating that our scheme performs better than traditional judgment at the conditions shown in these two examples. It is proved this condition is satisfied universally in cognitive radio system. Results indicate the maximal throughput increases while adopting our judgment scheme when parameters meet the condition.

Biconnectivity is the baseline graph theoretic metric of fault tolerance to node failures which can keep network connectivity while allowing unexpected node failures. In this paper we analyze this property for one-dimensional wireless Ad Hoc networks with finite nodes, which finds many applications in the real world, such as bus networks built along freeways and sensor networks deployed along frontiers. With common adopted assumption of uniform node distribution for static networks, the formula for the critical transmitting range that realizes network biconnectivity with certain probability is derived. Simulation results validate the accuracy of our conclusion and confirm its efficiency in practical network design.