Enhance dataset card: Add paper, project page links, task categories, and tags

#1
by nielsr HF Staff - opened
Files changed (1) hide show
  1. README.md +15 -3
README.md CHANGED
@@ -1,9 +1,23 @@
1
  ---
2
  license: cc-by-4.0
 
 
 
 
 
 
 
 
 
 
 
3
  ---
4
 
5
  # CloudAnoBench
6
 
 
 
 
7
  CloudAnoBench is a large-scale benchmark for context-aware anomaly detection in cloud environments, jointly incorporating both metrics and logs to more faithfully reflect real-world conditions. It consists of 1,252 labeled cases spanning 28 anomalous scenarios and 16 deceptive normal scenarios (approximately 200K lines), with explicit annotations for anomaly type and scenario. By including deceptive normal cases where anomalous-looking metric patterns are explained by benign log events, the benchmark introduces higher ambiguity and difficulty compared to prior datasets, thereby providing a rigorous testbed for evaluating both anomaly detection and scenario identification.
8
 
9
  ## 📁 Dataset Structure
@@ -162,6 +176,4 @@ for i in range(1, 17): # scenarios 1-16
162
  Where:
163
  - `scenario_id`: Scenario number
164
  - `case_number`: Case number within the scenario
165
- - Extension: `.csv` for structured data, `.log` for raw logs
166
-
167
-
 
1
  ---
2
  license: cc-by-4.0
3
+ task_categories:
4
+ - TIME_SERIES_FORECASTING
5
+ - TEXT_CLASSIFICATION
6
+ language:
7
+ - en
8
+ tags:
9
+ - anomaly-detection
10
+ - cloud-environments
11
+ - metrics
12
+ - logs
13
+ - time-series
14
  ---
15
 
16
  # CloudAnoBench
17
 
18
+ Paper: [Towards Generalizable Context-aware Anomaly Detection: A Large-scale Benchmark in Cloud Environments](https://huggingface.co/papers/2508.01844)
19
+ Project Page: https://jayzou3773.github.io/cloudanobench-agent/
20
+
21
  CloudAnoBench is a large-scale benchmark for context-aware anomaly detection in cloud environments, jointly incorporating both metrics and logs to more faithfully reflect real-world conditions. It consists of 1,252 labeled cases spanning 28 anomalous scenarios and 16 deceptive normal scenarios (approximately 200K lines), with explicit annotations for anomaly type and scenario. By including deceptive normal cases where anomalous-looking metric patterns are explained by benign log events, the benchmark introduces higher ambiguity and difficulty compared to prior datasets, thereby providing a rigorous testbed for evaluating both anomaly detection and scenario identification.
22
 
23
  ## 📁 Dataset Structure
 
176
  Where:
177
  - `scenario_id`: Scenario number
178
  - `case_number`: Case number within the scenario
179
+ - Extension: `.csv` for structured data, `.log` for raw logs