search

neukg / GRTE

52 stars 15 forks source link

How to deal with subjects or objects that appear multiple times in the text？ #2

Closed liuyijiang1994 closed 2 years ago

liuyijiang1994 commented 2 years ago

Thanks for your wonderful work! I have a question. Since there is only text of subject and object in the dataset and there is no location information, is the first found location used as the subject or object when reading data? We look forward to your reply!

liuyijiang1994 commented 2 years ago

In addition, we note that in some datasets (e.g. NYT24 train.json), there are some following instances:

{
        "text": "She has played Edith Piaf in a one-woman show seen in the United States and France , and possesses some of Piaf 's spirit as well as the throbbing mixture of steel and cotton in the voice .",
        "triple_list": [
            [
                "\u00c9dith Piaf",
                "/people/person/nationality",
                "France"
            ]
        ]
    }

These triples don't seem to be matched by the search function in main.py, how is this handled? Thanks!

neukg commented 2 years ago

Thanks for your wonderful work! I have a question. Since there is only text of subject and object in the dataset and there is no location information, is the first found location used as the subject or object when reading data? We look forward to your reply!

Yes, we only consider the location where the entity first appears as the standard location. There are certainly better ways of doing this, but that is beyond the scope of this article.

neukg commented 2 years ago

In addition, we note that in some datasets (e.g. NYT24 train.json), there are some following instances:

{
        "text": "She has played Edith Piaf in a one-woman show seen in the United States and France , and possesses some of Piaf 's spirit as well as the throbbing mixture of steel and cotton in the voice .",
        "triple_list": [
            [
                "\u00c9dith Piaf",
                "/people/person/nationality",
                "France"
            ]
        ]
    }

These triples don't seem to be matched by the search function in main.py, how is this handled? Thanks!

We simply ignore the wrong triplet.

sssssajfsd commented 2 years ago

In addition, we note that in some datasets (e.g. NYT24 train.json), there are some following instances:

{
        "text": "She has played Edith Piaf in a one-woman show seen in the United States and France , and possesses some of Piaf 's spirit as well as the throbbing mixture of steel and cotton in the voice .",
        "triple_list": [
            [
                "\u00c9dith Piaf",
                "/people/person/nationality",
                "France"
            ]
        ]
    }

These triples don't seem to be matched by the search function in main.py, how is this handled? Thanks!

Thanks for your wonderful work! I have a question. Since there is only text of subject and object in the dataset and there is no location information, is the first found location used as the subject or object when reading data? We look forward to your reply!

Hi, may I ask what the version of bert4keras is

liuyijiang1994 commented 2 years ago

In addition, we note that in some datasets (e.g. NYT24 train.json), there are some following instances:

{
        "text": "She has played Edith Piaf in a one-woman show seen in the United States and France , and possesses some of Piaf 's spirit as well as the throbbing mixture of steel and cotton in the voice .",
        "triple_list": [
            [
                "\u00c9dith Piaf",
                "/people/person/nationality",
                "France"
            ]
        ]
    }

These triples don't seem to be matched by the search function in main.py, how is this handled? Thanks!

Thanks for your wonderful work! I have a question. Since there is only text of subject and object in the dataset and there is no location information, is the first found location used as the subject or object when reading data? We look forward to your reply!

Hi, may I ask what the version of bert4keras is

I didn't run the code, I just looked :(

sssssajfsd commented 2 years ago

In addition, we note that in some datasets (e.g. NYT24 train.json), there are some following instances:

{
        "text": "She has played Edith Piaf in a one-woman show seen in the United States and France , and possesses some of Piaf 's spirit as well as the throbbing mixture of steel and cotton in the voice .",
        "triple_list": [
            [
                "\u00c9dith Piaf",
                "/people/person/nationality",
                "France"
            ]
        ]
    }

These triples don't seem to be matched by the search function in main.py, how is this handled? Thanks!

Thanks for your wonderful work! I have a question. Since there is only text of subject and object in the dataset and there is no location information, is the first found location used as the subject or object when reading data? We look forward to your reply!

Hi, may I ask what the version of bert4keras is

I didn't run the code, I just looked :(

OK,thank you!