In recent years, network traffic analysis solutions that are driven by artificial intelligence models have achieved impressive performance. The “magic spells” of these solutions come from the knowledge that they learn from large amounts of network traffic data. However, these solutions neglect the impact of the real-world network's complexity on data quality, which makes the knowledge they learn from regular network traffic data difficult to be effective on low-quality data. Considering the packet loss in real-world network environments, this paper presents TFMana to calibrate the inaccurate packet length features extracted from incomplete network traffic data. TFMana utilizes an encoder-based masked language model to predict features of lost packets, incorporating network traffic feature embeddings to enhance prediction accuracy. This approach enables the calibrated features to approximate those extracted from loss-free network traffic asymptotically. Comprehensive experiments are conducted to verify the effectiveness of the proposed method. The evaluation demonstrates that TFMana's calibration achieves recovery accuracy between 83.94 % and 85.66 %, with minimal sensitivity to packet loss rates. Integrated with four benchmark application identification models, TFMana significantly improves classification accuracy under packet loss conditions. Notably, the analysis models maintains reliable performance even at high packet loss rates of 30 %.

Network traffic classification is critical for ensuring network quality, security, and stability. However, the increasing complexity of network environments and the growth of encrypted traffic bring significant challenges. Traditional rule-based, machine learning-based, and deep learning-based approaches are limited by the scarcity of plaintext, reliance on handcrafted features, and the need for large labeled datasets. Pre-training methods have alleviated these issues, but existing models mainly focus on payload semantics and lack dedicated learning of traffic behavior patterns essential for encrypted traffic characterization. Motivated by this, we propose TrafficBM, a dual-modality pre-training framework that jointly models semantic features and traffic behavior patterns. Our approach extracts dualmodality features from network traffic and applies modalityspecific data augmentation to mitigate data imbalance and scarcity. During pre-training, BERT leverages masked bigram modeling (MBM) to capture semantic information, while Mamba uses a masked autoencoder (MAE) architecture to learn traffic behavior patterns. An adaptive gating network, together with a parameter-preserving warm-up strategy, fuses features from both pre-trained models during fine-tuning to improve downstream classification performance. TrafficBM achieves state-of-the-art results on six tasks across eight datasets, including over 0.99 accuracy on five datasets and a 10 % improvement over the best baseline on Datacon2021 Part 2, demonstrating strong generalization and robustness in network traffic classification.

Replay of interactive stories is an essential characteristic of the medium, enabling players to both see whether their choices impact the narrative, and work towards an understanding of the story and underlying system. Despite the centrality of replay to the form, little work has been done to investigate what motivates people to replay interactive stories, or when and why they stop replaying. In this paper, we report on the results of an exploratory qualitative observational study of 12 participants (aged 22 to 28) replaying the visual novel The Shadows that Run Alongside Our Car, a short work that presents a post-apocalyptic survival story from the perspective of two different playable characters and involves four dialogue choices leading to one of three possible endings. Unlike previous models of rereading, our observations suggest that players' reasons for deciding whether or not to replay change fluidly as their sense of closure changes, and players are deciding based not just on whether they have reached closure, but also on their estimate of the likelihood of seeing something new or moving closer to closure versus the effort required to replay. This suggests a need to rethink earlier, more simplistic models of rereading in interactive stories.

Self-Organizing Networks (SON) aim at automation of network parameters to make the planning, configuration, management, optimization and healing of mobile radio access networks simpler and faster, and thus also more cost efficient. SON use case in this article is Mobility Robustness Optimization (MRO), which targets automatic detection and correction of handover (HO) parameterization. To this end we present novel simulation framework called Self-Organizing Simulator 3 (SonS3) that makes use of real network data to enable an accurate performance analysis of mobility related SON use cases. In addition, we present an MRO algorithm which adapts dynamically the Cell Individual Offset (CIO), and a performance analysis within a realistic urban scenario. The proposed MRO algorithm leads to an overall reduction of more than 30% in mobility problems without causing significant increase in number of HOs.

The advancement in vehicles has enabled crowdsensing in vehicular fog computing (VFC), where vehicles are recruited to be assigned different subtasks and participate sensing activities that may disclose their sensitive information. To stimulate more participants, VFC systems should be able to provide reliable and privacy-preserving data transmission and processing mechanisms for the sensing report. To ensure the report process, we present a reliable and privacy- preserving task recomposition (REPTAR) for multiple subtasks sensing in VFC. Modified homomorphic Paillier encryption and superincreasing sequence are employed for aggregating hybrid subtasks into one ciphertext. Reliability is verified by means and variances of each aggregated subtasks from different vehicular fog nodes. Detailed security analysis and performance evaluation are provided to demonstrate the security, privacy-enhancement, efficiency and low complexity of the proposed REPTAR.