{"id":396,"date":"2019-02-05T13:43:08","date_gmt":"2019-02-05T13:43:08","guid":{"rendered":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/?p=396"},"modified":"2019-02-05T13:43:10","modified_gmt":"2019-02-05T13:43:10","slug":"reading-reflection-3-02-05-alon-bendelac","status":"publish","type":"post","link":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/2019\/02\/05\/reading-reflection-3-02-05-alon-bendelac\/","title":{"rendered":"Reading Reflection #3 &#8211; [02\/05] &#8211; [Alon Bendelac]"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\"><strong>Summary<\/strong>:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a><strong>Automated Hate Speech Detection<\/strong><\/a>: This study is about differentiating hate speech from other\nforms of offensive language. This study uses crowd-sourcing to create a dataset\nof tweets classified into three categories: hate speech, offensive language,\nand neither. Multiple models were tested: logistic regression, na\u00efve Bayes,\ndecision trees, random forests, and linear SVMs. The results show that 40% of\nhate speech was misclassified as offensive or neither.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a><strong>Early Public Responses<\/strong><\/a>: This paper studies the presence of conspiracy theories on YouTube,\nspecifically in Zika virus-related videos. The study looks for differences\nbetween user activity, sentiment, and content of two classifications of videos:\ninformational and conspiracy. It is found that user activity and sentiment are\nsimilar between the two classifications, but content is different.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Reflection<\/strong>:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Accuracy of CrowdFlower<\/strong>: I wonder how similar is the\nlabeling of different CrowdFlower workers. This could be tested as follows: Let\nall the workers classify the same set of 100 tweets. For each tweet, calculate\nthe variance in the classifications made by the workers. High variances mean\nthat the workers are classifying the same tweets differently.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Flagged posts<\/strong>: I think it would be interesting to\ninvestigate patterns in posts that get flagged for review or reported as inappropriate.\nDo people tend to flag posts excessively, or not enough? I think these results might\naffect how social media sites, such as Facebook and Twitter, develop policies\non hate speech.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Hate speech in retweets and replies<\/strong>: The paper didn\u2019t mention if the\ndataset they studied contained only original tweets, or also retweets and\nreplies. I think it would be interesting to study how hate speech differs\nbetween these types of tweets. Where is hate speech most prevalent on Twitter?<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">I think the conclusion\nof the \u201cAutomated Hate Speech Detection\u201d study can be improved. The\nsignificance of the findings and future work should both be a lot clearer and\nconcrete.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">In the \u201cEarly\nPublic Responses,\u201d I think the data could be presented better. Bar graphs would\nprobably be easier to understand than tables.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Small sample size<\/strong>: The sample size is very small (n=23\nfor information and n=12 for conspiracy). I think the paper should have talked\nmore about how this might affect their results.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Conspiracy theories on other types of\nsocial media platforms<\/strong>: I think the same type of study can be done on social media platforms\nother than YouTube. For example, we can study the prevalence of conspiracy\ntheories in Twitter. Number of views would be replaced by number of retweets,\nand replies and likes would stay the same.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Summary: Automated Hate Speech Detection: This study is about differentiating hate speech from other forms of offensive language. This study uses crowd-sourcing to create a dataset of tweets classified into three categories: hate speech, offensive language, and neither. Multiple models were tested: logistic regression, na\u00efve Bayes, decision trees, random forests, and linear SVMs. The results [&hellip;]<\/p>\n","protected":false},"author":247,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-396","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/wp-json\/wp\/v2\/posts\/396","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/wp-json\/wp\/v2\/users\/247"}],"replies":[{"embeddable":true,"href":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/wp-json\/wp\/v2\/comments?post=396"}],"version-history":[{"count":1,"href":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/wp-json\/wp\/v2\/posts\/396\/revisions"}],"predecessor-version":[{"id":398,"href":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/wp-json\/wp\/v2\/posts\/396\/revisions\/398"}],"wp:attachment":[{"href":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/wp-json\/wp\/v2\/media?parent=396"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/wp-json\/wp\/v2\/categories?post=396"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/wordpress.cs.vt.edu\/cs4984spring19\/wp-json\/wp\/v2\/tags?post=396"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}