SingularityNET is backed by the SingularityNET Foundation, which believes that the benefits of AI should not accrue only to a small set of powerful institutions, but rather should be shared by all. A key goal of SingularityNET is ensuring that the technology is benevolent by human standards and that the network is designed to incentivize and reward beneficial players. A Decentralized, Open Market and Network for AIs, Whitepaper 2.0: February 2019
Some of the specific AI services under development following: in the social robotics track are the 17 18 46 Action Dialogue GHOST dialogue engine based on OpenCog AI Expression Facial expression generation Perception Visual Faces Face recognition/tracking Face identification Gaze tracking Facial expression and emotion recognition Visual speaking and non-speaking detector Bodies Pose tracker Robust person detection Gesture recognition Gait characterization Auditory Voice Speech recognition Voice identification Voice activity detection Laughter detection Multiple-speaker speech separation 47 Detection of what language is being spoken
id: 1e3895c9e064bf97fc4464cbf077156a - page: 47
4.2.2.5 Mind-Modeling and Loving AI Development We are currently working on extending the Loving AI pilot program to add new functions to the OpenCog AI system related to unconditional love. These functions could be manifested via any reasonably flexible robot or avatar; however, for the immediate future we will continue using Sophia for experimentation and testing. Sophia has particularly strong emotional expression capability and a global media presence, and she is a system our team is familiar with.
id: 8248d04c85f05ab4090ec7848fc1aece - page: 49
The work in 2019 will focus on giving OpenCog/Sophia a genuine model of mind for the first time. The goal will be for the AI/robot to build a working internal model of thoughts, feelings, motivations, intentions, etc. of the person with whom it is interacting. This would give a significant boost to the robots understanding of people with whom she interacts, which we believe is required for her to genuinely express unconditional love. Of course this initial model of mind will not be the same as a typical humans model of other humans minds, but it will be a start in this important direction. One of the motivations of this work is to lay the groundwork to extend the Loving AI protocol so that the AI can learn about the person as they are interacting and to some extent bring that learning to bear in its statements and questions to the person. This will also allow for repeated sessions with the same person with continuity, as the robot builds and improves its model of the person.
id: e2d9b7c50c7bcfae31fc979f9567a30f - page: 49
This mind-modeling work integrates emotion-modeling as a result of leveraging and improving the AI frameworks emotion regulation, enabling the AIs emotions to better reflect and respond to the humans emotions that it models. This will give Sophia (and any other robots or avatars controlled by the software) richer emotional expression, better emotional connection with others, and the beginnings of an understanding of human emotion in general. All of these are steps toward skillful means in expressing and eventually feeling unconditional love. 4.2.2.6
id: 15300f43bbc84efbc3dc926a28730b01 - page: 49