-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathcorrelations.py
211 lines (139 loc) · 7.61 KB
/
correlations.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
import numpy as np
import pandas as pd
import scipy.stats as stats
from scipy.stats import pearsonr, spearmanr, kendalltau
import json
from tabulate import tabulate
def sample_level_correlation_summeval(human_metric):
print(f'Human metric: {human_metric}')
assert human_metric in ['coherence', 'relevance', 'consistency', 'fluency']
with open('data/summeval.json', 'r', encoding='utf-8') as f:
data = json.load(f)
auto_metrics=['rouge1_f', 'rouge2_f', 'rougel_f', 'bert_score_f', 'mover_score', 'prism_src_hypo', 'bart_score_src_hypo', 'bart_score_cnn_src_hypo', 'bart_score_para_src_hypo', 'chatgpt_%s'%human_metric]
headers = ['metric', 'spearman', 'pearsonr', 'kendalltau']
metric_with_corr = []
for metric in auto_metrics:
correlations = []
for doc_id in data:
target_scores = []
prediction_scores = []
sys_summs = data[doc_id]['sys_summs']
for sys_name in sys_summs:
prediction_scores.append(sys_summs[sys_name]['scores'][metric])
target_scores.append(sys_summs[sys_name]['scores'][human_metric])
if len(set(prediction_scores)) == 1 or len(set(target_scores)) == 1:
continue
correlations.append([
spearmanr(target_scores, prediction_scores)[0],
pearsonr(target_scores, prediction_scores)[0],
kendalltau(target_scores, prediction_scores)[0],
])
corr_mat = np.array(correlations)
spearman, pearman, ktau = np.mean(corr_mat[:, 0]), np.mean(corr_mat[:, 1]), np.mean(corr_mat[:, 2])
metric_with_corr.append([metric, spearman, pearman, ktau])
print(tabulate(metric_with_corr, headers=headers, tablefmt='simple'))
def dataset_level_correlation_summeval(human_metric):
print(f'Human metric: {human_metric}')
assert human_metric in ['coherence', 'relevance', 'consistency', 'fluency']
with open('data/summeval.json', 'r', encoding='utf-8') as f:
data = json.load(f)
auto_metrics=['rouge1_f', 'rouge2_f', 'rougel_f', 'bert_score_f', 'mover_score', 'prism_src_hypo', 'bart_score_src_hypo', 'bart_score_cnn_src_hypo', 'bart_score_para_src_hypo', 'chatgpt_%s'%human_metric]
headers = ['metric', 'spearman', 'pearsonr', 'kendalltau']
metric_with_corr = []
for metric in auto_metrics:
correlations = []
target_scores = []
prediction_scores = []
for doc_id in data:
sys_summs = data[doc_id]['sys_summs']
for sys_name in sys_summs:
prediction_scores.append(sys_summs[sys_name]['scores'][metric])
target_scores.append(sys_summs[sys_name]['scores'][human_metric])
correlations.append([
spearmanr(target_scores, prediction_scores)[0],
pearsonr(target_scores, prediction_scores)[0],
kendalltau(target_scores, prediction_scores)[0],
])
corr_mat = np.array(correlations)
spearman, pearman, ktau = np.mean(corr_mat[:, 0]), np.mean(corr_mat[:, 1]), np.mean(corr_mat[:, 2])
metric_with_corr.append([metric, spearman, pearman, ktau])
print(tabulate(metric_with_corr, headers=headers, tablefmt='simple'))
def sample_level_correlation_openmeva():
""" Evaluate summaries. Conduct summary-level correlations w.r.t each document """
auto_metrics = ['rouge1-r', 'rouge2-r', 'rougel-r', 'bertscore-r', 'bartscore', 'bartscore_cnn', 'bartscore_cnn_pa', 'fwppl', 'chatgpt']
with open('data/openmeva.json', 'r', encoding='utf-8') as f:
data = json.load(f)
headers = ['metric', 'spearman', 'pearsonr', 'kendalltau']
metric_with_corr = []
for metric in auto_metrics:
correlations = []
for item in data:
target_scores = []
prediction_scores = []
sys_summs = item['gen']
for sys_name in sys_summs:
prediction_scores.append(sys_summs[sys_name]['score'][metric])
target_scores.append(sys_summs[sys_name]['score']['human'])
if len(set(prediction_scores)) == 1 or len(set(target_scores)) == 1:
continue
correlations.append([
spearmanr(target_scores, prediction_scores)[0],
pearsonr(target_scores, prediction_scores)[0],
kendalltau(target_scores, prediction_scores)[0]
])
corr_mat = np.array(correlations)
spearman, pearman, ktau = np.mean(corr_mat[:, 0]), np.mean(corr_mat[:, 1]), np.mean(corr_mat[:, 2])
metric_with_corr.append([metric, spearman, pearman, ktau])
print(tabulate(metric_with_corr, headers=headers, tablefmt='simple'))
def dataset_level_correlation_openmeva():
""" Evaluate summaries. Conduct summary-level correlations w.r.t each document """
auto_metrics = ['rouge1-r', 'rouge2-r', 'rougel-r', 'bertscore-r', 'bartscore', 'bartscore_cnn', 'bartscore_cnn_pa', 'fwppl', 'chatgpt']
with open('data/openmeva.json', 'r', encoding='utf-8') as f:
data = json.load(f)
headers = ['metric', 'spearman', 'pearsonr', 'kendalltau']
metric_with_corr = []
for metric in auto_metrics:
correlations = []
target_scores = []
prediction_scores = []
for item in data:
sys_summs = item['gen']
for sys_name in sys_summs:
prediction_scores.append(sys_summs[sys_name]['score'][metric])
target_scores.append(sys_summs[sys_name]['score']['human'])
if len(set(prediction_scores)) == 1 or len(set(target_scores)) == 1:
continue
correlations.append([
spearmanr(target_scores, prediction_scores)[0],
pearsonr(target_scores, prediction_scores)[0],
kendalltau(target_scores, prediction_scores)[0]
])
corr_mat = np.array(correlations)
spearman, pearman, ktau = np.mean(corr_mat[:, 0]), np.mean(corr_mat[:, 1]), np.mean(corr_mat[:, 2])
metric_with_corr.append([metric, spearman, pearman, ktau])
print(tabulate(metric_with_corr, headers=headers, tablefmt='simple'))
def dataset_level_correlation_bagel(human_metric): # ['rouge1_r', 'rouge2_r', 'rougel_r', 'bert_score_r', 'mover_score']
""" Evaluate summaries. Conduct summary-level correlations w.r.t each document """
print(f'Human metric: {human_metric}')
assert human_metric in ['informativeness', 'naturalness', 'quality']
with open('data/bagel.json', 'r', encoding='utf-8') as f:
data = json.load(f)
auto_metrics=['rouge1_p', 'rouge2_p', 'rougel_p', 'bert_score_p', 'mover_score', 'prism_ref_hypo', 'bart_score_ref_hypo', 'bart_score_cnn_ref_hypo', 'bart_score_para_ref_hypo', 'chatgpt_%s'%human_metric]
headers = ['metric', 'spearman', 'pearsonr', 'kendalltau']
metric_with_corr = []
for metric in auto_metrics:
correlations = []
target_scores = []
prediction_scores = []
for doc_id in data:
prediction_scores.append(data[doc_id]['scores'][metric])
target_scores.append(data[doc_id]['scores'][human_metric])
correlations.append([
spearmanr(target_scores, prediction_scores)[0],
pearsonr(target_scores, prediction_scores)[0],
kendalltau(target_scores, prediction_scores)[0],
])
corr_mat = np.array(correlations)
spearman, pearman, ktau = np.mean(corr_mat[:, 0]), np.mean(corr_mat[:, 1]), np.mean(corr_mat[:, 2])
metric_with_corr.append([metric, spearman, pearman, ktau])
print(tabulate(metric_with_corr, headers=headers, tablefmt='simple'))