When India fell into the Covid-19 crisis towards the end of April, #indianeedSoxen and #PakistanStandswithia began to become trends in neighboring countries.
It is a strong support show from the other side of seven decades of hostility equations.
But at that time, #eendiasaySortokashmir began to become a trend together.
In the end, the split was reduced, indicating the first data set that publicly analyzed the geopolitical relationship between the two countries in the middle of a raging pandemic; 85% of Tweet from Pakistan about the Covid-19 crisis in our country using the Hashtags that support India.
Researchers, from Carnegie Mellon University, Maulana Abul Kalam Azad University, and Blockchain and AI Company Onai, passed more than 3 lakh tweets by 1.5 lakh users between April 21 and May 4 this year.
“Most supports tweets, with supportive tweets that receive almost 1.5 times more like and retweet than those who do not support.
(From a data set), 85% use a supportive tag,” Lead Author Dr.
Ashiqur Khudabukhsh from Carnegie Mellon University’s The Language Technology Institute to Ti.
In two weeks of Twitter activities in India and Pakistan they studied, they found around 21,600 tweets with #PakistanStandswithIndia hashtag and around 3,800 with #PakistanStandwithIndia hashtag.
#IndianeedSoxen has almost 20,000 tweets from Pakistan and #indianeedoxen, 2,400.
In comparison, trend tags that are presented #eendiasaySortokashmir only have 8,000 tweets from Pakistan and #Indiasay sorry to Tokashmir, only about 170.
“We further see that the tweet containing supportive tags originating from Pakistan received more than they are from India,” The paper is received by the association for computing linguistic workshops on natural language processing for positive impacts, said.
With this, Khudabukhsh and his team had made a conversation from one centered on hatred speech for one about the speech of hope, a project he began in 2019, right after the crisis of the Pulwama.
“Along with the deceased Prof.
Jaime Carloonell, Shriphani (Palakodety, Co-Author) and I have been widely working on the web manifestation of political tensions between India and Pakistan in the context of the 2019 Indian-Pakistani conflict,” Khudabukhsh said.
At that time, they had analyzed comments on YouTube videos to train Hope Speech Classifier – Machine Learning Techniques to identify “administrative-spread, peace content”.
For this study, they used the Hope Speech Classifier they had made in 2019 along with Empathy-Distress Classifier trained in response to news footage to identify a tweet pattern.
Because they must be careful – tags can be hijacked.
“Many of the (tweet) will not contain text or only images or links, or can post unrelated content (eg product promotion) or can express content that does not support,” Khudabukhsh said.
It runs two directions.
For example, someone tweeted, “India deserves this.
You face what you do on Kashmir and fool Pakistan supports India about this, you are only a slave for England,” with Tagar #indianeedSoxee.
Likewise, other tweets said, “Political differences have their place but US prayers Pakistan without brothers and Indian sisters.
May God provide health to everyone,” with Hastag #eendiasaysorto to Kashmir who is negative.
But their classification detects positive tweets 83% of the time, much higher than the existing method.
“While we are working on this project, we have friends and family back in India who suffer from a crisis.
We are proud to work on this humanitarian project that focuses on national well-being under the health crisis,” Khudabukhsh said.
“In this case and our previous study, we found that both countries (India and Pakistan) have web users who behave with each other.
In a world full of divisive content, I think focus on positive content to combat hatred that can become an effective strategy.
“