Innovative Media Intelligence Platform to Monitor Claims, Rumours, Opinions, Arguments and Stances Online

自然语言处理的创新并没有发展到基于关键字，情感，主题和实体检测提供社交倾听品牌。传播机构发现，很难优化竞选信息，以应对迅速发展的在线时代精神。他们还因对产品或品牌的恶意诽谤，或从未发生过的虚构事件而遭受声誉损害。1998年，英国医生安德鲁·韦克菲尔德发表了一项研究，表明MMR疫苗可能引发自闭症，这导致英国2岁儿童的MMR疫苗接种率下降了20%。现有的社交倾听工具监测品牌提及，跟踪关键主题内文章的受欢迎程度以及情绪的演变。然而，他们并没有提取实际所说的内容。例如，他们无法检测到正在改变公众对一个问题的看法的热门论点。他们无法跟踪公众对产品的立场是如何变化的。他们没有为该机构提供有用的反诉/论据，以推动公共对话朝着某个方向发展。总而言之，他们并没有主动帮助品牌优化品牌资产，也没有保护他们免受傀儡的虚假信息宣传。该项目将产生一个创新的媒体分析系统，可以监控在线特定社会问题的索赔，谣言，意见，论点和立场（“CROAS”）。这将作为一套算法沿着出现在终端客户端。核心创新是建立一个注释平台，可以将CROAS呈现给疫苗接种和其他社会问题中的特定专家社区团体/领域专家。这些人拥有必要的领域知识，可以准确地对文本中的CROAS进行标记和分类，比非专家更好。使用这个系统，Factmata希望部署上级算法与新的训练/测试数据集的问题的CROAS检测。Factmata在自动事实核查、自然语言处理（“NLP”）、数据注释等领域拥有丰富的经验，并成功构建了检测仇恨言论和宣传语言的算法。15个以上的通信机构已经证明有兴趣建立一个CROAS监测系统。该团队在内容审核行业拥有丰富的销售经验，并希望为公关/通信行业打造一款新产品。该项目将带来显著的出口导向型增长，可观的投资回报率，增加就业机会和进一步的研发投资机会。

Innovation in natural language processing has not evolved much beyond providing social listening to brands based on keyword, sentiment, topic and entity detection. Communications agencies find it hard to optimise campaign messaging to account for rapidly evolving zeitgeists online. They also suffer reputational damage from toxic trolling about products or brands, or made-up events that never took place. In 1998, UK doctor Andrew Wakefield published a study suggesting the MMR vaccine could trigger autism, which led to a 20% drop in MMR vaccination rates among 2-year-olds in England.Existing social listening tools monitor brand mentions, track the popularity of articles within key topics, and the evolution of sentiment. However, they don't extract what is actually being said. For example, they cannot detect the top trending arguments that are changing public opinion on an issue. They cannot track how public stance towards a product has been changing. They do not provide helpful counter-claims/arguments for the agency to nudge public dialogue in a certain direction. In summary, they don't provide proactive help for brands to optimise brand equity, or protect them against disinformation campaigns by sock-puppets.The project will produce an innovative media analysis system that can monitor claims, rumours, opinions, arguments and stances ("CROAS") on specific societal issues online. This will be surfaced as a suite of algorithms along with an end-client. The core innovation is building an annotation platform that can surface CROAS to specific expert community groups/domain specialists within vaccination and other societal issues. These individuals have the necessary domain knowledge to accurately tag and classify CROAS in text, better than non-experts. Using this system, Factmata hopes to deploy superior algorithms with novel train/test datasets for the problem of CROAS detection. The project will help communications agencies execute campaigns which align better with public debate, as well as protect their clients against disinformation.Factmata has experience in the fields of automated fact checking, natural language processing ("NLP"), data annotation and more, and has successfully built algorithms to detect hate speech and propaganda language. It has proven interest from 15+ communications agencies to build a CROAS monitoring system. The team has experience selling into the content moderation industry, and is looking to build a new product for the PR/ communications industry.The project will deliver significant export-led growth, a substantial ROI, increased employment and further opportunity for R&D investment.