{"id":990,"date":"2021-03-23T01:51:08","date_gmt":"2021-03-23T01:51:08","guid":{"rendered":"https:\/\/sail.usc.edu:\/~ccmi\/?p=990"},"modified":"2022-02-07T01:37:23","modified_gmt":"2022-02-07T01:37:23","slug":"actions-agents-and-patients","status":"publish","type":"post","link":"https:\/\/sail.usc.edu:\/ccmi\/actions-agents-and-patients\/","title":{"rendered":"Actions, Agents and Patients"},"content":{"rendered":"\n<section class = \"section section--small section--white\">\n  <div class=\"container\">\n    <div class=\"grid grid--1-to-3\">\n      <h2>Project Description<\/h2>\n      <div> <p> We develop models, data and statistical analysis necessary to uncover the pervasiveness of stereotypical portrayals in characters&#8217; actions at a scale.<\/p> <\/div>\n    <\/div>\n  <\/div>\n<\/section>\n\n\n\n<section class=\"section section--highlight\">\n  <div class=\"container text-center\">\n    <h4 class=\"callout caps\"><\/h4>\n    <h2> WHY IS IT IMPORTANT?<\/h2>\n    <p> To understand the influential relation between exposure to media content and accepted social norms and expectations through large-scale quantitative measurements of characters\u2019 attributes (e.g., age, gender, or race)<\/p>\n  <\/div>\n<\/section>\n\n\n\n<figure class=\"wp-block-video aligncenter\"><video controls src=\"https:\/\/sail.usc.edu\/~mica\/ccmi-demos\/2021_10_16_14_32_50.mp4\"><\/video><\/figure>\n\n\n\n<section class = \"section section--small section--white\">\n  <div class=\"container\">\n    <header class=\"grid\">\n      <h2>  Dataset stats <\/h2>\n    <\/header>\n    <div class=\"grid grid--3\">\n      <div>\n        <h2 class=\"stats-number\"> 1.2M+<\/h2>\n        <h3 class=\"caps\"> Sentences<\/h3>\n        <p>  Action descriptions<\/p>\n      <\/div>\n      <div>\n        <h2 class=\"stats-number\"> 50k+<\/h2>\n        <h3 class=\"caps\"> Actions<\/h3>\n        <p> Distinct actions<\/p>\n      <\/div>\n      <div>\n        <h2 class=\"stats-number\"> 25K+<\/h2>\n        <h3 class=\"caps\"> Characters<\/h3>\n        <p> Identified characters.<\/p>\n      <\/div>\n    <\/div>\n  <\/div>\n<\/section>\n\n\n\n<section class = \"section section--small section--white\">\n  <div class=\"container\">\n    <div class=\"grid\">\n      <h2>Automatic Identification of Characters and Actions<\/h2>\n      <p class=\"text-center small\"><img decoding=\"async\" class=\"full-width\" src=\"https:\/\/sail.usc.edu\/ccmi\/wp-content\/uploads\/2021\/03\/movieSRL-BERT.svg\" alt=\"\">\n      <br> Our proposed SRL system. Starting at the bottom, the input to the system is an action description in natural language. The output, shown at the top of the figure, is a sequence of labels (one per word). Labels indicate whether this word is playing the role of action, agent, patient or none. From its inputs, our model obtains a highly-contextualized representation for each word using the BERT transformer. Each representation corresponds to a high dimensional dense vector that encodes the semantics of that word and the context it plays within the sentence. The sequence of vector representations is then fed into a recurrent neural network and a softmax layer for sequence labeling. As a post-processing step, a set of heuristics aggregate multi-word expressions to handle the case of groups of agents or patients.<\/p>\n    <\/div>\n  <\/div>\n<\/section>\n\n\n\n<section class = \"section section--small section--white\">\n  <div class=\"container\">\n    <header class=\"grid\">\n      <h2>Statistical Analysis<\/h2>\n    <\/header>\n    <div class=\"grid grid--2\">\n      <div>\n        <img decoding=\"async\" src=\"https:\/\/sail.usc.edu\/ccmi\/wp-content\/uploads\/2021\/03\/CodeCogsEqn-1.svg\" class=\"full-width\" alt=\"\">\n      <\/div>\n      <div class=\"our-team\">\n        <p>We propose a statistical model to identify significant differences in the frequency of the action portrayals due to the role and gender of its participants. We use a Poisson-regression generalized linear mixed model (GLMM). GLMMs are an extension of generalized linear models (e.g., logistic regression) to include both fixed and random effects. <\/p>\n      <\/div>\n    <\/div>\n  <\/div>\n<\/section>\n\n\n<section class=\"section section--small section--white\">\n  <div class=\"container container--small\">\n    <header class=\"grid\">\n      <h2>Results<\/h2>\n    <\/header>\n    <div class=\"tabs\">\n      <div class=\"tabs__labels\">\n        <label class=\"tabs__label active\" for=\"hd1\">\n          Agency        <\/label>\n        <label class=\"tabs__label\" for=\"hd2\">\n          Male Gaze        <\/label>\n        <label class=\"tabs__label\" for=\"hd3\">\n          Disability        <\/label>\n        <label class=\"tabs__label\" for=\"hd4\">\n          Emotion        <\/label>\n        <label class=\"tabs__label\" for=\"hd5\">\n                  <\/label>\n      <\/div>\n      <div class=\"tabs__content\">\n        <input type=\"radio\" checked=\"\" class=\"tab__button\" name=\"current\" id=\"hd1\">\n        <div class=\"tab\">\n\t\t\tThe proportion of male agents is significantly higher than that of female agents. Additionally, the proportion of female patients is larger than that of male patients.        <\/div>\n        <input type=\"radio\" class=\"tab__button\" name=\"current\" id=\"hd2\">\n        <div class=\"tab\">\n\t\t\tOur results highlight differences in the emphasis placed on the female appearance and sexual objectification of women actors.          \n        <\/div>\n        <input type=\"radio\" class=\"tab__button\" name=\"current\" id=\"hd3\">\n        <div class=\"tab\">\n\t\t\tWe found that whenever the character requires someone to push their wheelchair, male agents are unlikely to \u2018wheel\u2019 neither female patients nor male patients        <\/div>\n        <input type=\"radio\" class=\"tab__button\" name=\"current\" id=\"hd4\">\n        <div class=\"tab\">\n\t\t\tMale characters are less likely to display affection, particularly when the agent and patients happen to be male.        <\/div>\n        <input type=\"radio\" class=\"tab__button\" name=\"current\" id=\"hd5\">\n        <div class=\"tab\">\n\t\t\t        <\/div>\n      <\/div>\n    <\/div>\n  <\/div>\n<\/section>\n\n\n\n<section class = \"section section--small section--white\">\n  <div class=\"container\">\n    <div class=\"grid grid--1-to-3\">\n      <h2>Summary<\/h2>\n      <div> <p> This work presents a novel large-scale analysis on the actions taken by the characters, and how these actions are related to gender-biases in media.<\/p> <\/div>\n    <\/div>\n  <\/div>\n<\/section>\n\n\n\n\n<section class=\"section section--small section--white\">\n    <div class=\"container container--small\">\n        <header class=\"grid\" >\n            <h2> Related Links <\/h2>\n        <\/header>\n        <a href=\"https:\/\/github.com\/usc-sail\/mica-actions-agents-and-patients\" target=\"_blank\" class=\"link-block\">\n            <div class=\"link-block__thumbnail link-block__icon\" \n                    style=\"background-image:url(https:\/\/sail.usc.edu\/ccmi\/wp-content\/uploads\/2020\/06\/GitHub-Mark-64px.png)\">\n            <\/div>\n            <div class=\"container-fluid\">\n            <div class=\"row\">\n                <div class=\"col\">\n                    <h3>GitHub<\/h3>\n                 <\/div>\n            <\/div>\n            \n            <div class=\"row\">\n                <div class=\"col\">\n                <p>Code Repo for large-scale analysis of characters&#8217; actions<\/p>\n                 <\/div>\n            <\/div>\n            <\/div>\n        <\/a>\n    <\/div>\n<\/section>\n","protected":false},"excerpt":{"rendered":"<p>A computational linguistic lens into gendered actions in film<\/p>\n","protected":false},"author":2,"featured_media":1041,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"inline_featured_image":false,"footnotes":""},"categories":[32,31,4],"tags":[],"acf":[],"_links":{"self":[{"href":"https:\/\/sail.usc.edu:\/ccmi\/wp-json\/wp\/v2\/posts\/990"}],"collection":[{"href":"https:\/\/sail.usc.edu:\/ccmi\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/sail.usc.edu:\/ccmi\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/sail.usc.edu:\/ccmi\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/sail.usc.edu:\/ccmi\/wp-json\/wp\/v2\/comments?post=990"}],"version-history":[{"count":31,"href":"https:\/\/sail.usc.edu:\/ccmi\/wp-json\/wp\/v2\/posts\/990\/revisions"}],"predecessor-version":[{"id":1599,"href":"https:\/\/sail.usc.edu:\/ccmi\/wp-json\/wp\/v2\/posts\/990\/revisions\/1599"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/sail.usc.edu:\/ccmi\/wp-json\/wp\/v2\/media\/1041"}],"wp:attachment":[{"href":"https:\/\/sail.usc.edu:\/ccmi\/wp-json\/wp\/v2\/media?parent=990"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/sail.usc.edu:\/ccmi\/wp-json\/wp\/v2\/categories?post=990"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/sail.usc.edu:\/ccmi\/wp-json\/wp\/v2\/tags?post=990"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}