Discover, Connect & Collaborate at TECHINNOVATION 2021
Explainable Artificial Intelligence Systems for Energy Consumption and Occupant Thermal Comfort Evaluation
Do you trust AI (artificial intelligence)? Will you deploy AI if you do not trust it? AI has been a popular word in those years. Companies and institutes cannot afford to lag behind in the wave of AI. Research effort for AI has been significant; however, the real-world adoption of AI is not that widespread. Why? Because people do not trust AI. Without sufficient understanding and trust, AI benefits will not be approachable with the fear of any hazard and damaging consequences. Our team is aware of the challenge, and we have designed, developed, and tested the explainable AI systems for smart buildings and smart grid for energy consumption and occupant thermal comfort evaluation. We have the demonstrated the effectiveness of our system and our research achievements have already been published and recognized in top-tier publications. We are expanding the scope of our explainable AI to the other smart building and smart grid applications to enable great diversity and large impact. We look forward to hearing from different sectors and companies that are concerned about the black-box nature of AI. We can collaborate and work together to tackle the challenge and demystify AI to realize AI’s full potential in smart buildings and smart grid. Partnerships can be flexible, and we welcome the initiatives about both academic research and industry development.
Technology Features, Specifications and Advantages
Explainable AI is one of the latest topics in the AI domain. Discussion and tutorials have become active in recent years’ AI-related conferences and workshops. AI research in its beginning stage is mostly accuracy driven; however, people start realizing that accuracy is not all about AI and more comprehensive considerations about robustness, explainability, and so on have been raised recently. Our team is in the frontier of explainable AI research in applied domains and aims to bring the latest explainable AI technologies to smart buildings and smart grid. Our team are competitive in those domains with our multidisciplinary expertise in AI and domain knowledge. Our top-tier publications in those domains can be strong justifications. Last but not least, explainable AI is necessary to clear the government regulations and meet related policies, as black-box AI systems are prohibited to be deployed in many sectors. The benefits, or advantages, of explainable AI can be diverse. For building and grid system operators, explainable AI helps them understand and validate the system dynamics and explains the AI decision-making process. For developers, explanations and interpretations enable them to perform system inspection and optimization. For end-users like occupants and energy users, explainable AI can also promote acceptance and instill user behavior changes. Those advantages are not easily available in the existing accuracy-driven AI in the building and grid sectors.
Our progress on smart buildings and smart grid covers several aspects, e.g., energy consumption and occupant thermal comfort evaluation. We believe our solution is ready to take off in those applications. Explainable AI also serves different purposes, e.g., validation and optimization for system developers, monitoring and inspection for system operators, and recommendation and behavior change for consumers. We believe we can make our system more comprehensive to cover various aspects and meet the requirements of different users.
Here, we would like to highlight that explainable AI can be applied to various areas as long as existing AI solutions are not explainable and interpretable. We can build up trust in AI systems and introduce transparency and explainability to different smart buildings and smart grid applications. Our ongoing effort covers other domains like smart transportation and smart manufacturing. We foresee the research outcome will be promising with a huge impact in those areas.
We look forward to discussing with the industry to broaden the scope of our explainable AI system for smart buildings and grid. We believe the market of explainable AI is significant given the fact that AI will be one of the dominant technologies and can reshape industry and society in the coming decades.
Building and grid software developers: explainable AI-enabled system validation and optimization.
Building and grid managers and operators: explainable AI-enabled system monitoring and inspection.
End-users like occupants: explainable AI-enabled recommendation engine to instill behavior change.
AI providers for building and grid sector: meet regulatory requirements with explainable, transparent, and trustful AI systems.