Datasets:

Modalities:
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
pandas
Punchwe commited on
Commit
0c44556
·
verified ·
1 Parent(s): 93e50bc

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +11 -2
README.md CHANGED
@@ -84,7 +84,16 @@ dataset_info:
84
 
85
  This is the dataset repository for SCAN MCD splits. In total, there are three splits - mcd1, mcd2, and mcd3
86
 
 
 
 
 
 
 
 
 
 
 
 
87
  You can load them by:
88
  ```datasets.load_dataset("Punchwe/SCAN_MCDSplits", name="mcd1", split="train")```
89
-
90
- [More Information needed](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
 
84
 
85
  This is the dataset repository for SCAN MCD splits. In total, there are three splits - mcd1, mcd2, and mcd3
86
 
87
+ SCAN is a set of simple language-driven navigation tasks for studying compositional learning and zero-shot generalization. The SCAN tasks were inspired by the CommAI environment, which is the origin of the acronym (Simplified versions of the CommAI Navigation tasks).
88
+
89
+ The relevant SCAN paper is:
90
+
91
+ [Generalization without systematicity: On the compositional skills of sequence-to-sequence recurrent networks](https://arxiv.org/pdf/1711.00350). ICML 2018.
92
+
93
+ The relevant MCD split paper is:
94
+
95
+ [Measuring Compositional Generalization: A Comprehensive Method on Realistic Data](https://arxiv.org/pdf/1912.09713). ICLR 2020.
96
+
97
+
98
  You can load them by:
99
  ```datasets.load_dataset("Punchwe/SCAN_MCDSplits", name="mcd1", split="train")```