首页|Patent Issued for Detecting affective characteristics of text with gated convolutional encoder-decoder framework (USPTO 11886480)
Patent Issued for Detecting affective characteristics of text with gated convolutional encoder-decoder framework (USPTO 11886480)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
According to news reporting originating from Alexandria, Virginia, by NewsRx journalists, a patent by the inventors Chawla, Kushal (Kadubeesanahalli, IN), Chhaya, Niyati Himanshu (Hyderabad, IN), Khosla, Sopan (Kadubeesanahalli, IN), filed on August 29, 2022, was published online on January 30, 2024. The assignee for this patent, patent number 11886480, is Adobe Inc. (San Jose, California, United States). Reporters obtained the following quote from the background information supplied by the inventors: “Human expressions, such as written or verbal communication, typically include both a factual component and a non-factual component. The human expressions are sometimes analyzed to detect the non-factual component as a method to determine an effectiveness of delivering the factual component of the expression. Personalization of websites, targeted communications, and targeted marketing materials all rely on an accurate characterization of the non-factual component of the human expression. Analysis of the nonfactual component using machine-learning techniques is useful to filter expressions that are provided to a target. In one example, the analysis provided by a machine-learning technique provides an indication that the non-factual component of an analyzed human expression is not appropriate for a specific target group because the non-factual component lacks an affective characteristic that is particularly relevant to the target group. In such an example, the affective characteristic identified by the machine-learning technique indicates that the human expression is too informal for a communication targeted to a business acquaintance or too formal for a communication targeted to a teenager.
NETL
NSTL
