amritasaha1812 / CSQA_Code

59 stars 20 forks source link

How to get the knowledge graph #15

Closed SRL94 closed 2 years ago

SRL94 commented 2 years ago

Hi, How can I get the corresponding knowledge graph for this dataset?

vardaan123 commented 2 years ago

Hi, the knowledge graph can be obtained from the Wikidata dump available at https://dumps.wikimedia.org/wikidatawiki/latest/wikidatawiki-latest-pages-articles.xml.bz2 . Note, we used an older version of this dump for constructing this dataset.

On Mon, Dec 6, 2021 at 1:56 AM Sirui @.***> wrote:

Hi, How can I get the corresponding knowledge graph for this dataset?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/amritasaha1812/CSQA_Code/issues/15, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEEM5DNGILLMQ4W4BPRUECTUPRNCZANCNFSM5JN3WZ2A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Thanks and Warm Regards

Vardaan Pahuja

SRL94 commented 2 years ago

Thanks.

Step4: Download the wikidata JSONs from the link https://zenodo.org/record/4052427#.X2_hWXRKhQI and put them in a folder wikidata_dir

Is it the knowledge graph you used?

vardaan123 commented 2 years ago

Yes, the one mentioned in Step 4 is the processed version of Wikidata in json format. I thought you meant the raw version and hence the previous link.

On Mon, Dec 6, 2021 at 11:50 PM Sirui @.***> wrote:

Thanks.

Step4: Download the wikidata JSONs from the link https://zenodo.org/record/4052427#.X2_hWXRKhQI and put them in a folder wikidata_dir Is it the knowledge graph you used?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/amritasaha1812/CSQA_Code/issues/15#issuecomment-987566533, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEEM5DMRQFVRHMYDRE6STATUPWHDBANCNFSM5JN3WZ2A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Thanks and Warm Regards

Vardaan Pahuja

SRL94 commented 2 years ago

Hi, Does the KG https://zenodo.org/record/4052427#.X2_hWXRKhQI contain all entities and paths involved in the question-answering task?

Best regards Sirui

vardaan123 commented 2 years ago

Yes, it does.

On Tue, Jan 25, 2022 at 11:49 PM Sirui @.***> wrote:

Reopened #15 https://github.com/amritasaha1812/CSQA_Code/issues/15.

— Reply to this email directly, view it on GitHub https://github.com/amritasaha1812/CSQA_Code/issues/15#event-5954333803, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEEM5DNBY3OT7YX7U6OESTDUX54MLANCNFSM5JN3WZ2A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

You are receiving this because you commented.Message ID: @.***>

-- Thanks and Warm Regards

Vardaan Pahuja

SRL94 commented 2 years ago

In https://zenodo.org/record/4052427#.X2_hWXRKhQI, there is a set of json files. Which one is the KG? wikidata_short_1 and wikidata_short_2?

SRL94 commented 2 years ago

Hi, also, what is "active_set" in the CSQA_v9? Looking forward to your reply. Thank you.

vardaan123 commented 2 years ago

The KG is the union of triples contained in wikidata_short_1 and wikidata_short_2. The active set is a symbolic representation of the "entire" set of triples involved in answering a particular question. For example: active_set=(c(Q8386)|c(Q181394), P129, Q1949517) means head entity corresponds to type Q8386 or Q181394, property P129 and tail entity Q1949517.

SRL94 commented 2 years ago

Thanks for the reply. I noticed some items are with empty active sets. Does it mean the answer is untrusted?

vardaan123 commented 2 years ago

No. The active set field is provided for experimental purpose in this dataset but it doesn't cover all question types. It is mainly of interest for logical reasoning, comparative and quantitative reasoning questions. This field can be used for some case studies on these specific question types, and not meant for evaluation in general.