TrafficInstruct dataset and benchmark for instruction tuning on network traffic released
Original title: TrafficInstruct: A Multi-Task Network Traffic Dataset and Benchmark for Instruction Tuning
55,000 structured records support 10 detection tasks with task-specific LoRA adapters improving performance.
The authors present TrafficInstruct, a dataset of raw packet captures and structured records for 10 network traffic analysis tasks. An automated workflow captures traffic from controlled executions and attack experiments. Task-specific adaptation with LoRA adapters improves classification performance.
- Dataset supports malware traffic detection and botnet detection
- Partitioned by collection sessions for evaluation
- Task-specific LoRA adapters improve performance
- Benchmark defines fixed partitions and scoring rules
From the original article
1 min read at arxiv.orgInstruction tuning for multi-task network traffic analysis requires data with clearly defined sources, supervision targets, and evaluation conditions.
By Zhengdao Li · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.