recent network intrusion datasets

 Here are some recent network intrusion datasets suitable for machine learning training, along with their sources and URLs:

1. **UNSW-NB15 Dataset**: This dataset contains nine different types of attacks, including DoS, worms, backdoors, and fuzzers, along with raw network packets. The training set includes 175,341 records and the testing set 82,332 records from different types, both attack and normal【41†source】. 

   - URL: [UNSW-NB15 Dataset](https://paperswithcode.com/dataset/unsw-nb15)

2. **CICIDS2017 Dataset**: The CICIDS2017 dataset consists of labeled network flows, including full packet payloads in pcap format, along with labeled flows and CSV files for machine and deep learning purposes【48†source】.

   - URL: [CICIDS2017 Dataset](https://paperswithcode.com/dataset/cicids2017)

3. **UQ NIDS Datasets (FlowMeter Format)**: Introduced by Sarhan et al., this dataset is specifically formatted for use with FlowMeter, a tool for extracting flow-based features from network traffic【55†source】.

   - URL: [UQ NIDS Datasets (FlowMeter Format)](https://paperswithcode.com/dataset/uq-nids-datasets-flowmeter-format)

4. **CIC IoT Dataset 2022**: Aimed at profiling, behavioral analysis, and vulnerability testing of different IoT devices, this dataset encompasses various experiments capturing network traffic of IoT devices under different conditions, including power, idle, interactions, scenarios, active use, and attacks【62†source】【63†source】.

   - URL: [CIC IoT Dataset 2022](https://paperswithcode.com/dataset/cic-iot-dataset-2022)

5. **IoT Benign and Attack Traces Dataset**: This dataset includes data collected for research on detecting volumetric attacks on IoT devices. It contains flow data and annotations for both benign and attack scenarios【73†source】【74†source】.

   - URL: [IoT Benign and Attack Traces Dataset](https://paperswithcode.com/dataset/iot-benign-and-attack-traces)

These datasets are valuable resources for training and evaluating machine learning models for network intrusion detection, especially considering the diverse nature of network attacks and behaviors they encompass.