Do Transformers Actually Help Intrusion Detection? A Temporal Sequence Evaluation on CIC-IDS2017

Recent deep learning approaches for network intrusion detection increasingly incorporate temporal architectures such as recurrent networks and Transformers, often reporting near-perfect performance on CIC-IDS2017. However, many existing studies neither supply their temporal modules with genuine sequence inputs nor evaluate under realistic, leakage-free conditions, making it unclear whether reported gains arise from true sequence-modeling capability. In this work, we reformulate CIC-IDS2017 as a temporal intrusion-detection task by constructing ordered flow sequences from network conversations and benchmarking nine classical and deep learning architectures under a random split, two leakage-free splits, and a padding-scheme ablation. The central finding is that padding convention, not architecture, determines the Transformer's performance: on genuinely sequential (non-padded) windows the Transformer achieves the highest macro-F1 of any model in the experiment (0.89); under zero-pad+mask evaluation it drops markedly (-0.24 macro-F1), while LSTM, GRU, and 1D-CNN remain stable. Under leakage-free group evaluation the Random Forest is the most robust model (+0.009), while the Transformer's false-alarm rate grows from 0.04% to 2.7%, a 67-fold increase invisible under conventional protocols. These findings demonstrate that evaluation methodology -- specifically padding convention and split protocol -- has a larger effect on reported performance than architectural choice, and that widely used random splits with repeat-last padding can overestimate model robustness by up to 0.24 macro-F1. We advocate leakage-free splits, explicit padding disclosure, and sequence-aware benchmarking as standard practice in future IDS research. Code and implementation details are available at https://github.com/zachmocz/temporal-ids-bench.

Automotive Speed Estimation: Sensor Types and Error Characteristics from OBD-II to ADAS

Modern on-road navigation systems heavily depend on integrating speed measurements with inertial navigation systems (INS) and global navigation satellite systems (GNSS). Telemetry-based applications typically source speed data from the On-Board Diagnostic II (OBD-II) system. However, the method of deriving speed, as well as the types of sensors used to measure wheel speed, differs across vehicles. These differences result in varying error characteristics that must be accounted for in navigation and autonomy applications. This paper addresses this gap by examining the diverse speed-sensing technologies employed in standard automotive systems and alternative techniques used in advanced systems designed for higher levels of autonomy, such as Advanced Driver Assistance Systems (ADAS), Autonomous Driving (AD), or surveying applications. We propose a method to identify the type of speed sensor in a vehicle and present strategies for accurately modeling its error characteristics. To validate our approach, we collected and analyzed data from three long real road trajectories conducted in urban environments in Toronto and Kingston, Ontario, Canada. The results underscore the critical role of integrating multiple sensor modalities to achieve more accurate speed estimation, thus improving automotive navigation state estimation, particularly in GNSS-denied environments.