Commit
·
f46de4e
1
Parent(s):
0f9a3c5
Update README.md
Browse files
README.md
CHANGED
@@ -63,21 +63,23 @@ English.
|
|
63 |
|
64 |
### Data Fields
|
65 |
|
66 |
-
candidates (
|
67 |
-
cue (string).
|
68 |
-
associations (string).
|
69 |
-
score_fool_the_ai (int64).
|
70 |
-
num_associations (int64).
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
ID (int64).
|
76 |
|
77 |
### Data Splits
|
78 |
-
|
79 |
-
|
80 |
-
--
|
|
|
|
|
|
|
81 |
|
82 |
## Dataset Creation
|
83 |
|
|
|
63 |
|
64 |
### Data Fields
|
65 |
|
66 |
+
candidates (list): ["bison", "shelter", "beard", "flea", "cattle", "shave"] - list of image candidates.
|
67 |
+
cue (string): pogonophile - the generated cue.
|
68 |
+
associations (string): ["bison", "beard", "shave"] - the images associated with the cue selected by the user.
|
69 |
+
score_fool_the_ai (int64): 80 - the spymaster score (100 - model score) for fooling the AI, with CLIP RN50 model.
|
70 |
+
num_associations (int64): 3 - The number of images selected as associative with the cue.
|
71 |
+
num_candidates (int64): 6 - the number of total candidates.
|
72 |
+
solvers_jaccard_mean (float64): 1.0 - three solvers scores average on the generated association instance.
|
73 |
+
solvers_jaccard_std (float64): 1.0 - three solvers scores standard deviation on the generated association instance
|
74 |
+
ID (int64): 367 - association ID.
|
|
|
75 |
|
76 |
### Data Splits
|
77 |
+
There is a single TEST split. In the accompanied paper and code we sample it to create different training sets, but the intended use is to use winogavil as a test set.
|
78 |
+
There are different number of candidates, which creates different difficulty levels:
|
79 |
+
-- With 5 candidates, random chance for success is 38%.
|
80 |
+
-- With 6 candidates, random chance for success is 34%.
|
81 |
+
-- With 10 candidates, random chance for success is 24%.
|
82 |
+
-- With 12 candidates, random chance for success is 19%.
|
83 |
|
84 |
## Dataset Creation
|
85 |
|