Event definition missing in v1.0

[tommasoc80]

Event definition missing in v1.0

[tommasoc80]

Thanks for signalling it.

"rehab" should not be an event on its own. It should be merged with "checks into".

Will take care in the next release of the dataset.

[sjtuprog]

There are another two issues.

1. How to decide whether an extended link is falling_action or precondition?
   I found that in some case:
   Event A = Event mentions (a1,a2,a3)
   Event B = Event mentions (b1,b2,b3)
   Annotated: a1 fall_action b1.
   Annotated: a3 precondition b2.
   How to decide a2 fall_action? b3.

2. How to handle the case when causal relation and event coreference happens between same event mentions?
   Event A = Event mentions (a1,a2,a3)
   a1 falling_action a2.

Thanks!

[tommasoc80]

CASE 1:

• we are not assuming transitivity among the relations at the moment;
• extended links have been created to "gain" extra annotations between event mentions;
• actually, there should not be annotated as valid the case you are reporting: if (a1,a2,a3) are all mentions of the same events, and so it is the case for (b1,b2,b3) it should not be the case that the same event set can be both falling_action and precondition with the same target event. If this has been annotated, it needs to be checked and solved (the error could be in the coreference annotations or in the plot_link annotation)
• Can you point to a specific case in the data?

CASE 2:

• as stated above, if this is the case there is an error, either in the coreference annotation or in the plot_link annotation
• Can you point a specific case in the data?

Thanks!!

[sjtuprog]

CASE 1:
In EventStoryLine v0.9,

1_21ecbplus.xml.xml t1_checking_into_morningside_or_showing_up t1_comply
['5', '4', '6'] ['7']

Notice that 5->7 is annotated as FALLING_ACTION but 6->7 is annotated as precondition. However, 5,6 are both t1_checking_into_morningside_or_showing_up.

For more examples:

12_3ecbplus.xml.xml t12_thwarted t12_pirates_arrested
['12', '8'] ['10', '1']
13_4ecbplus.xml.xml t13_church_fire t13_foul_play
['5', '10', '16'] ['13', '19']
14_9ecbplus.xml.xml t14_arson t14_fire
['2', '5', '8'] ['11', '13', '12', '14']
19_9ecbplus.xml.xml t19_killed t19_riot
['11', '12', '6'] ['10', '14']
22_12ecbplus.xml.xml t22_charged t22_murder
['3', '9'] ['10', '5', '4']
3_6ecbplus.xml.xml t3_escape_nov_30 t3_escape_nov_30
['2', '4', '7', '8', '18'] ['2', '4', '7', '8', '18']
3_8ecbplus.xml.xml t3_escape_nov_30 t3_escape_nov_30
['5', '7', '11', '12', '16', '18'] ['5', '7', '11', '12', '16', '18']
30_1ecbplus.xml.xml t30_repair t30_repair
['14', '17', '3', '5', '6'] ['14', '17', '3', '5', '6']
32_11ecbplus.xml.xml t32_double_murder t32_double_murder
['20', '22', '1', '7'] ['20', '22', '1', '7']
32_3ecbplus.xml.xml t32_double_murder 34
['2', '13', '30'] ['34']
33_7ecbplus.xml.xml 73 10 plot link annotation error. Entity involved.
35_3ecbplus.xml.xml t35_jamal_wills_arrested t35_jamal_wills_DUI
['23', '10', '19', '18'] ['8', '15', '30']
35_5ecbplus.xml.xml t35_jamal_wills_arrested t35_jamal_wills_DUI
['1', '18'] ['2', '6', '17']
37_14ecbplus.xml.xml t37_aceh_earthquake t37_girl_killed
['21', '16', '18'] ['22', '7']
37_1ecbplus.xml.xml t37_aceh_earthquake t37_landslides
['24', '23', '32', '31'] ['28', '29']
8_3ecbplus.xml.xml t8_rioting t8_rioting
['25', '27', '23', '3', '2', '6', '14', '17', '16', '31'] ['25', '27', '23', '3', '2', '6', '14', '17', '16', '31']
8_4ecbplus.xml.xml t8_rioting t8_rioting
['21', '9', '8', '13', '15', '14', '17', '19'] ['21', '9', '8', '13', '15', '14', '17', '19']

CASE 2:
In EventStoryLine v0.9,
Those are the within event coreference causal links.

1_18ecbplus.xml.xml 45 92_93 t1_approved
1_18ecbplus.xml.xml 92_93 150 t1_approved
12_8ecbplus.xml.xml 272 280 t12_operation
16_5ecbplus.xml.xml 55 74 t15_shot
18_10ecbplus.xml.xml 101 114 t18_shooting
18_4ecbplus.xml.xml 35 38 t18_shooting
19_2ecbplus.xml.xml 47 51 t19_killed
20_10ecbplus.xml.xml 33 51 40519
3_10ecbplus.xml.xml 128 207 t3_escape_nov_30
3_4ecbplus.xml.xml 160 165 t3_escape_nov_30
3_5ecbplus.xml.xml 63 158 t3_escape_nov_30
3_6ecbplus.xml.xml 90 175_176 t3_escape_nov_30
3_6ecbplus.xml.xml 90 95_96 t3_escape_nov_30
3_7ecbplus.xml.xml 67 126 t3_escape_nov_30
3_8ecbplus.xml.xml 121 144 t3_escape_nov_30
3_8ecbplus.xml.xml 155 208 t3_escape_nov_30
30_1ecbplus.xml.xml 33 67 t30_repair
30_1ecbplus.xml.xml 33 96 t30_repair
30_1ecbplus.xml.xml 33 50 t30_repair
32_11ecbplus.xml.xml 77_78 89 t32_double_murder
32_11ecbplus.xml.xml 41 89 t32_double_murder
32_11ecbplus.xml.xml 35_36 89 t32_double_murder
32_11ecbplus.xml.xml 35_36 41 t32_double_murder
33_10ecbplus.xml.xml 85 94 t33_shielded_children
33_11ecbplus.xml.xml 35 151 t33_shooting
33_7ecbplus.xml.xml 73 10 plot link annotation error. Entity involved.
41_1ecbplus.xml.xml 34 54 t41_drop_bombs_camp
41_2ecbplus.xml.xml 16_17 37 t41_drop_bombs_camp
41_2ecbplus.xml.xml 16_17 66 t41_drop_bombs_camp
41_4ecbplus.xml.xml 22 99 t41_drop_bombs_camp
41_4ecbplus.xml.xml 50 99 t41_drop_bombs_camp
41_8ecbplus.xml.xml 35 66 t41_drop_bombs_camp
5_1ecbplus.xml.xml 59 94 t5_replacing_coaches
5_9ecbplus.xml.xml 215 232 t5_replacing_coaches
7_5ecbplus.xml.xml 52 55 t7_klitschko_ends _fight
7_5ecbplus.xml.xml 68 90 t7_klitschko_ends _fight
8_3ecbplus.xml.xml 104 106 t8_rioting
8_3ecbplus.xml.xml 621 694 t8_rioting
8_3ecbplus.xml.xml 612_613 694 t8_rioting
8_3ecbplus.xml.xml 47 72 t8_rioting
8_3ecbplus.xml.xml 47 81 t8_rioting
8_3ecbplus.xml.xml 47 75 t8_rioting
8_4ecbplus.xml.xml 125 154 t8_rioting
8_4ecbplus.xml.xml 94 106_113 t8_rioting
8_4ecbplus.xml.xml 119 125 t8_rioting
8_4ecbplus.xml.xml 77 94 t8_rioting
8_4ecbplus.xml.xml 77 79 t8_rioting
8_9ecbplus.xml.xml 57 62 t8_bank_workers_died

[tommasoc80]

can you point out to examples for version 1.0 as well? version 0.9 should not be used.

[sjtuprog]

CASE 1:
In EventStoryLine v1.0,

1_21ecbplus.xml.xml t1_checking_into_morningside_or_showing_up t1_comply
['5', '4', '6'] ['7']

Notice that 5->7 is annotated as FALLING_ACTION but 6->7 is annotated as precondition. However, 5,6 are both t1_checking_into_morningside_or_showing_up.

CASE 2:
1_18ecbplus.xml.xml 45 92_93 t1_approved

I checked the first 2 example in v0.9, they also exist in v1.0.

Thanks.

[tommasoc80]

Thanks!
I will look into that - I am working on a new release of the data.