Projects Projects

PROJECTS

HOME
Finished Projects
群體人工智慧-群體人工智慧之用戶特徵學習、情緒運算及行為塑型研究
01
JAN
2018
31
JAN
2021
Bo-Hao Su Chin-Po Chen Jeng-Lin Li Huang-Cheng Chou Hao-Chun Yang
The use of AI-technologies have transformed business fundamentally across the globe. While AI-applications have been dominated by foreign industrial super-powers using centralized brains, i.e., big companies (Google, Facebook, and Baidu) with powerful data centers, controlling all the computational power and data informatics, Taiwan's opportunity lies in the development and prevalence of smart edge devices. The popularization of these devices presents an unique opportunity for large-scaled and precise learning of collective human intelligence. Crowd AI will not only lessen the burden of centralized data centers, but, more importantly, will also allow learning to be more user-personalized and privacy-sensitive at a massive scale. However, advanced technical developments are required to achieve robustness given these computationally-limited yet wide-spread devices enabling next generation AI-technologies.

As part of the integrated proposal of Crowd AI, in this proposal, our aim is to focus on developing deep-learning based AI-technologies embedded with human factors (e.g., emotion and social factors) using large-scale heterogeneous type of cyber-physical data collected from real-world crowd user devices. A tangible outcome of our research proposal will be to integrate the developed human-centric/emotion AI-enabled technologies to achieve:

1) Precision and contextually-meaningful user stratification through data-driven representation learning
2) Advanced user behavior analytics with emotion-AI (recognition and stimulation) technologies
3) Long-term behavior shaping strategy within real-world contextualized application domain

By bridging the technical and data resources gaps, we can imagine the possibility of a variety of next-generation human computing interface systems being utilized across small to mid-scale service companies. The integration of human factors-embedded AI-modules into the scheme of Crowd AI can help achieve massive wide-spread effect and bring competitiveness of Taiwan business globally by focusing deeply on the user-ends. By learning from the crowd, operating for the crowd, and finally delivering intelligence to the crowd, these human-centric computing technologies play an evitable role realizing the next frontier of Crowd AI.
PARTNER
科技部
AINTU