Deputy General Director Nguyen Tan Minh: AI Application – Technology is not a “Barrier”

Data and artificial intelligence play an extremely important role in implementing digital transformation as well as applying 4.0 technologies into real life. However, the key issue still lies in the mindset of whether businesses and users really want to “transform” or not. On this topic, Smile Express cites an interview with Mr. Nguyen Tan Minh – Deputy Director of the Company, conducted by Nhip Song So Magazine.

Interviewer: What is your assessment of the development and application of AI in Vietnam?

MinhNT: Artificial Intelligence (AI) exists in many areas of life, helping to improve human quality of life as well as the quality and efficiency of work. Following the global trend, Vietnam has been gradually developing products and services using AI technology. Since the early 1970s, AI research in Vietnam began with AI courses taught to computer science students at some universities. In the 1980s, Vietnam also had some initial AI-applied products, though not yet optimal or closely serving practical needs.

In recent years, with the advancement of technology, especially open-source libraries from Google, Microsoft, etc., along with the emergence of affordable high-performance computers (GPUs, etc.), AI research, development, and application in Vietnam has grown rapidly. Vietnamese companies are confident in their ability to create AI-based products of their own. Many AI products made in Vietnam have been applied in various fields such as education, socio-economics, transportation, healthcare… and have brought significant economic efficiency. I think these are positive signals that we can confidently believe in a future where AI leads the way.

Interviewer: In that context, what is GMO-Z.com RUNSYSTEM’s orientation for developing products to meet social needs?

MinhNT: It can be seen that AI has already been applied in many aspects of life and across industries, such as virtual assistants, smart homes, self-driving cars… along with many specialized products in specific sectors (such as finance, media…).

However, there are still some traditional industries that use a large amount of paper documents, such as banking, insurance, public administration… Moreover, these sectors have high demands for identity verification with personal documents (driver’s license, insurance card, ID card, citizen ID card…), making the application of technology to improve labor efficiency, reduce processes, time, and manual operations in storing and retrieving information extremely necessary.

Recognizing that potential demand, GMO-Z.com RUNSYSTEM has researched and developed the SmartOCR solution using AI technology for handwriting recognition, driver’s license recognition, insurance card recognition (applied in Japan), ID cards, citizen IDs, passports, etc.

Interviewer: Could you elaborate on the “stage” where AI technology demonstrates its value in such highly accurate tasks?

MinhNT: Let’s take SmartOCR as an example: by using machine learning (deep learning) combined with image processing technologies, the system improves character recognition quality. Thanks to that, AI can recognize whether the document is a driver’s license, ID card, or citizen ID in the image. The next step is to use image processing to adjust tilt, remove noise, segment text, smoothen, and then continue to use AI for text recognition. In deep learning, creating training data is very important; alongside human-labeled data, GMO-Z.com RUNSYSTEM also uses AI itself to generate additional training samples for machine learning.

For example, a seemingly difficult task is recognizing handwriting on forms (such as money transfer request forms, etc.) applied in Japan. Here, AI supports converting images (scanned handwritten forms) into text for input into Japanese banking systems. The AI has been trained on a large dataset of Japanese handwriting to identify and output accurate characters, then automatically input them into the system.

Testing, experience, and statistics have shown that applying AI for handwriting recognition with accuracy above 90% has reduced employee data entry time by more than 50%. For driver’s license, ID card, and citizen ID recognition, AI-based solutions shortened verification time by over 70%, increased processing speed, and improved customer service.

Interviewer: We know that GMO-Z.com RUNSYSTEM’s main clients for AI-related products are in Japan. Does the company have plans to develop similar AI applications in Vietnam?

MinhNT: Since research and development began in early 2017, GMO-Z.com RUNSYSTEM’s SmartOCR has already been trusted and “ordered” by several clients and is currently in the acceptance stage for practical deployment. For example: GMO Aozora Bank uses SmartOCR to recognize driver’s licenses and insurance cards (in Japanese), Japanese clients are testing and improving handwriting recognition on administrative forms. In Vietnam, a financial lending company is also evaluating its effectiveness.

In addition to SmartOCR, the company is also applying AI to other digital transformation solutions such as Checkman and AI Camera for the Vietnamese market.

Checkman is a solution that combines cameras with AI to inspect product quality through images, enhancing efficiency and reducing errors, resulting in higher productivity compared to manual inspection.

Normally, the accuracy of inspection depends on the sharpness of the human eye; when eyes get tired, inspection quality is affected. Checkman is the solution to ensure consistent quality inspection, meeting required standards.

AI Camera is a solution that uses cameras combined with AI to analyze customer information in stores, such as age, gender, lifestyle, fashion… and customer behavior such as frequency, time spent, interest in certain products, or which counters they linger at. Based on this information, the marketing team or store management can design campaigns or optimize product placement to better serve customers and increase store revenue.

In 2019, in line with digital transformation, GMO-Z.com RUNSYSTEM will continue to focus on investing in both human resources and technology to research and launch meaningful products that bring practical value to life.