R&D 100 Awards: Battling COVID-19 with Misinformation join-fighting mechanism enabled by AI forensics technology

R&D 100 Awards: Battling COVID-19 with Misinformation join-fighting mechanism enabled by AI forensics technology

By Institute for Information Industry (III)

Another good news comes from the pride of Taiwan in the technology field! This is the 59th year of the U.S. R&D 100 Awards, a hall of dreams for countless technology developers from all over the world. Supported by the special project of the Department of Industrial Technology (DoIT), the Ministry of Economic Affairs (MOEA), Digital Service Innovation Institute, Institute for Information Industry developed the “Misinformation Join-fighting Mechanism Enabled by AI Forensics Technology ”, which stood out from more than 1,000 entries from 17 countries and won the US R&D 100 Award. This technology has also been recognized by the Special Contribution Award for Corporate Social Responsibility. It is the only technology in Taiwan that has received double recognition, which is a new milestone for Taiwan’s software service industry.

Known as the Oscar of the science and technology industry, the R&D 100 Awards not only has not been affected by the epidemic this year but also has received the largest number of entries in recent years.

Starting from the needs of the industry, the Department of Industrial Technology (DoIT), the Ministry of Economic Affairs (MOEA), utilizes its own R&D capabilities accumulated from developing science and technology projects for many years to support the technology research and development of the Institute for Information Industry. This year, III lived up to the expectations of the public and came up with the “Misinformation Join-fighting Mechanism Enabled by AI Forensics Technology ”, which was recognized by nearly 50 judges worldwide in the Software/Services category, and won the U.S. R&D 100 Awards.

III Executive Vice President Ren-Dar Yang also delightfully pointed out that III participated in the competition as a representative of Taiwan and it not only has a technology winning a prize but also one being on the list of the finalists. This is the highest honor from the science and technology industry and also represents that Taiwan’s R&D achievements have been internationally recognized. The accomplishment can not be esteemed lightly. In the future, III expects to continuously use information systems to provide value-added software services so that it could help companies in the industry solve difficult problems and accelerate the process of digital transformation.

 “Misinformation Join-fighting Mechanism Enabled by AI Forensics Technology” Developed by Digital Service Innovation Institute, Institute for Information Industry Wins Two Grand Awards.

Misinformation is an important issue that people around the world must discuss. Digital Service Innovation Institute (DSI), Institute for Information Industry (III) finds that many fact-checking agencies are hindered by insufficient manpower and have a hard time accurately selecting topics among massive amounts of data. Therefore, DSI sets off to develop "Misinformation Quick filter Technology" and "AI Forensic Technology" to assist Taiwan FactCheck Center in successfully integrating multiple information channels regarding misinformation complaints as well as constructs a complete database of suspected rumors and misinformation quick filter dashboards, which are beneficial to monitoring rumors efficiently and conducting fact-checking.

 

Ping-Yi Chen, the group leader of DSI, further explains that rumors are different from general news reports. The biggest feature is that misinformation is designed to be swiftly and massively distributed by various forms with different combinations of text, pictures, videos, and links as well as with the same content slightly transforming over time. The development team utilizes the technologies such as in-depth Natural Language Processing (NLP), heterogeneous data cross-comparison, and social media data augmentation to automatically group rumors with different formats but similar content and compare them with fact-checking reports;  At the same time, the team develops "AI forensic technology" to construct AI risk models from multiple aspects such as content analysis (linguistic features characteristics), propagation analysis, and traceability analysis.  By human-machine collaboration, this technology helps inspectors to identify high-risk and high-impact issues for inspection so that they can accurately select topics by using objective data and quickly curb the spread of misinformation. 

 “If you go to convenience stores and conduct name registration at the same time with confirmed cases, you will be put in quarantine.” “Coronaviruses will live in the throat area for four days... Drink as much warm water, salt, or vinegar as much as possible can eliminate the viruses.” When these rumors caused by the epidemic spread from one place to another, it was difficult for the public to tell the truth from the false instantly. In May of this year (2021), the spread of rumors about the epidemic has increased four to five times, creating a new wave of peaks. III cooperates with the Taiwan FactCheck center and helps it quickly gather nearly 1.5 million suspicious messages. When there is a sudden increase in the number of rumors in a short period of time, the verification operation will be initiated immediately so that the spread of misinformation will be quickly eliminated through the multi-party verification process.

In addition to increasing the speed of fact-checking and clarification, the research and development of this technology can also be applied to the fields of news media, marketing, and information communication services, providing multiple services including industry expert verification, etc. “Misinformation Join-fighting Mechanism Enabled by AI Forensics Technology” not only is recognized by the judges of R&D 100 Awards and wins the R&D 100 Award but also is bestowed the Special Contribution Award for Corporate Social Responsibility.

Resilience Smart Maintenance System for Semiconductor Automation Equipment (RSMS), a System Assisting Manufacturing Industry in Making Intelligent Products, Shortlisted for R&D 100 Awards.

The overall turnover of the semiconductor industry in Taiwan has reached new highs year by year and the export volume and output value of Taiwan’s precision component-related products have continued to increase accordingly. Nevertheless, this is also becoming a major challenge for Taiwan’s manufacturing industry. Regional Industry Service Division (RISD), Institute for Information Industry. In order to help the industry stabilize production quality and expand production capacity, RISD has developed the "Resilience Smart Maintenance System for Semiconductor Automation Equipment (RSMS)”, which is equipped with a variety of sensors and has excellent equipment abnormality prediction and diagnosis functions, allowing professionals to timely understand the quality situation of machine processed products. With the advanced warnings issued by the AI ​​model, professionals can reduce unexpected downtime and improve production line efficiency.

Additionally, RISD Central Area Center Director Hsiao-yu Wang, further explains that the equipment-side immediate maintenance decision-making function is also a major feature of the system. The R&D team builds an industrial-level knowledge database as well as constructs maintenance documents and integrates the maintenance experience and knowledge of various events into the files so that when the equipment is abnormal, the system has the mechanism to diagnose key causes and clear troubles, which can solve the bottleneck of remote maintenance and repair when exporting equipment to foreign countries and provide equipment-side immediate maintenance decision-making services; the " Resilience Smart Maintenance System for Semiconductor Automation Equipment (RSMS)" has been empirically conducted by a number of domestic manufacturers, and the results have been excellent. It is a great honor that the system is recognized as a finalist of the US R&D 100 Awards. In the future, RISD will continue to assist the manufacturing industry in product intelligence upgrading.