Page 729 - AI for Good Innovate for Impact
P. 729

AI for Good Innovate for Impact



               4�9�        Category 9: Accessibility





                   Use Case 1: Bridging the Digital Gap: AI Sign Language                                           4.9: Accessibility

               Interpretation Solution









               Organization: Baidu

               Country: China

               Contact Person(s):

                    Primary: Tianmin Long, intlcomm@ baidu .com
                    Secondary: Yang Liu, liuyang145@ baidu .com
                    Additional Contact: Cathy Fu, fuduanling@ baidu .com


               1      Use Case Summary Table


                Item               Details
                Category           Accessibility

                Problem Addressed Public accessibility issue for the deaf and hard-of-hearing community
                Key Aspects of Solu- Artificial Intelligence(AI) Sign Language Platform, AI Sign Language All-in-
                tion               One Machine, and digital avatar
                Technology         AI, sign language translation and interpretation, digital avatars
                Keywords

                Data Availability  Private
                Metadata (Type of  Text, audio, visual
                Data)              (Note: Data were collected through a dedicated collection app and are
                                   strictly limited to model training purposes only.)
                Model Training and  The platform has collected nearly 10 million natural sign language samples
                Fine-Tuning        to build a large-scale sign language corpus to generate natural sign
                                   language that aligns with the deaf and hard-of-hearing(DHH) communi-
                                   ty's communication habits.
                Testbeds or Pilot  University and Research Lab Pilot, Rehabilitation Center Integration,
                Deployments        Government Service Counter Trials [1]














                                                                                                    693
   724   725   726   727   728   729   730   731   732   733   734