Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
csv
Languages:
Japanese
Size:
10K - 100K
ArXiv:
License:
Update README.md
Browse files
README.md
CHANGED
|
@@ -51,7 +51,7 @@ configs:
|
|
| 51 |
path: "data/a3-recognition/test.tsv"
|
| 52 |
---
|
| 53 |
|
| 54 |
-
#
|
| 55 |
|
| 56 |
[](https://arxiv.org/abs/2408.05906)
|
| 57 |
[](https://github.com/cyberagentailab/adtec)
|
|
@@ -60,13 +60,16 @@ configs:
|
|
| 60 |
[](https://github.com/CyberAgentAILab/AdTEC/blob/main/materials/NAACL2025-slides.pdf)
|
| 61 |
[](https://youtu.be/3QmKidnlkiI)
|
| 62 |
|
|
|
|
| 63 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 64 |
|
| 65 |
The AdTEC dataset is designed to evaluate the quality of ad texts from multiple aspects, considering practical advertising operations.
|
| 66 |
|
| 67 |
-
> [!NOTE]
|
| 68 |
-
> We are excited to announce that our paper has been accepted to the main track of [NAACL 2025](https://aclanthology.org/2025.naacl-long.391/)! We are making the data from this research publicly available in this repository.
|
| 69 |
-
|
| 70 |
## Experiments and Tasks Considered in the Paper
|
| 71 |
|
| 72 |
This dataset includes five tasks:
|
|
|
|
| 51 |
path: "data/a3-recognition/test.tsv"
|
| 52 |
---
|
| 53 |
|
| 54 |
+
# Dataset Card for AdTEC
|
| 55 |
|
| 56 |
[](https://arxiv.org/abs/2408.05906)
|
| 57 |
[](https://github.com/cyberagentailab/adtec)
|
|
|
|
| 60 |
[](https://github.com/CyberAgentAILab/AdTEC/blob/main/materials/NAACL2025-slides.pdf)
|
| 61 |
[](https://youtu.be/3QmKidnlkiI)
|
| 62 |
|
| 63 |
+
## Dataset Description
|
| 64 |
|
| 65 |
+
- **Homepage:** [Project Page](https://cyberagentailab.github.io/AdTEC/)
|
| 66 |
+
- **Repository:** [GitHub](https://github.com/cyberagentailab/adtec/)
|
| 67 |
+
- **Paper:** [AdTEC: A Unified Benchmark for Evaluating Text Quality in Search Engine Advertising](https://arxiv.org/abs/2408.05906)
|
| 68 |
+
- **Leaderboard:**
|
| 69 |
+
- **Point of Contact:** [Peinan Zhang](https://github.com/peinan)
|
| 70 |
|
| 71 |
The AdTEC dataset is designed to evaluate the quality of ad texts from multiple aspects, considering practical advertising operations.
|
| 72 |
|
|
|
|
|
|
|
|
|
|
| 73 |
## Experiments and Tasks Considered in the Paper
|
| 74 |
|
| 75 |
This dataset includes five tasks:
|