Personal Artificial Intelligence (P.A.I.) developer alt Inc.(https://alt.ai/en/) (Head office: Minato-ku, Tokyo, Japan, CEO: Kazutaka Yonekura), in collaboration with Video Research, Inc., the first result of this experiment was the successful generation of digital clones (virtual survey respondents) that reproduce survey monitors (actual survey respondents).
By combining Video Research's vast survey data with Alt's digital clones, it is possible to predict the evaluation of program viewers at the planning stage and detailed targets, which has been difficult to do in the past.
Successful Generation of Digital Clones of Real Survey Monitors
By using Alt's “Clone Modeling Engine,” we have succeeded in generating digital clones based on the data obtained from survey monitors (actual survey respondents) owned by Video Research.
The digital clones generated here were installed in Alt's “Nulltitude,” and questionnaires were conducted, and it was confirmed that responses could be obtained stably.
Using the distributed computing platform “Emeth”
The generation of digital clones and their responses usually requires a very long computation time. In order to run Nulltitude with a large number of digital clones at a practical speed, we used the distributed computing platform “Emeth” that Alt has built.
This enabled Nulltitude to process a huge number of operations at high speed.
The data obtained here will be used to ensure stable operation when more digital clones are used and will also be used to make adjustments for the opening of Emeth to the public.
Video Research's “Research 4.0” Enables Prediction of Program Evaluation at the Planning Stage.
Since February 2021, Alt has been conducting joint research with Video Research to realize a next-generation marketing research method that utilizes AI technology, which Video Research has named as “Research 4.0”*3.
By realizing Research 4.0, we believe that it will be possible to create programs that are highly evaluated by viewers by reflecting simulations of viewer evaluations at the program planning stage.
In addition, by generating a large number of digital clones with various characteristics (values) that can be freely answered, it will be possible to confirm detailed target responses that have been difficult to survey in the past, and also to delve deeper into impressions of the evaluations.
Based on the success of this experiment, Alt will conduct in-depth analysis and research with Video Research on the results, and will build a highly accurate and convenient digital clone questionnaire system by implementing a variety of question formats and responses.