Data is virtualized, integrated, and utilized no matter where it is. The advantage is clear. We can utilize data very quickly and change quickly based on it,” emphasized Shanmuga Sundar, director of Munni & Di.
In the fourth expert session, Lee Joo-hwan, co-founder and CEO of the global collaboration platform Sweet, began his presentation on Phone Number List the topic of ‘Work AI that opens with data connection and flow.
CEO Lee Joo-hwan said that leading AI companies such as Microsoft and Google have recently been integrating LLM into their productivity products, and that they should pay attention to AI that innovates customer workflows. In other words, a new market is opening that uses AI to connect workflows and data scattered within a company, and the topic of Work AI is how much data workflow can be optimized.
CEO Joohwan Lee said, “Each team has different ways of working and different tools they use. Now is not the time to gather all employees in one place and manage them. However, it is also burdensome to introduce additional tools and solutions for connection. “How can we connect employees and data that are siled in multiple systems?” he said, emphasizing the need for a hub that can connect data flows.
In the fifth presentation, Oracle MySQL Manager Su-mi Ryu talked about how to easily apply machine learning to applications under the topic of 'MySQL HeatWave's new machine learning (ML) function.' According to Manager Ryu Su-mi, MySQL is a popular open source relational database that ranks unrivaled in the cloud sector. MySQL HeatWave, which is based on this, enables real-time data analysis without ETL, especially when performing analysis work in the cloud, said Manager Ryu Su-mi. In fact, one Brazilian education company is said to have achieved more than 300 times faster performance after migrating to MySQL HeatWave.
Next, Manager Ryu Soo-mi said, “Generally speaking, when using a database, there are usually separate databases and machine learning solutions. HeatWave also provides machine learning solutions. You can learn by applying the algorithm directly where there is actual data. Of course, other models may be suitable if there are data scientists and sufficient human resources, but there are many companies that do not have data scientists. “In these places, machine learning can be applied with automated models to improve business,” he explained.
In the expert session that followed, Lee Dong-hoon, head of AI the Nutrigene's AI division, evaluated ChatGPT as “the most plausible AI released to the public in the history of artificial intelligence,” and noted that it was the dawn of a superscale AI war.
So what choices should companies make and what role should they play? Lee Dong-hoon, head of the division, said, “New technologies such as big data, cloud, artificial intelligence, and large-scale language models hit like a carpet bomb every two to three years. Companies are experiencing various problems when introducing new technologies, such as skyrocketing costs, failure to achieve return on investment, or limitations in internal absorption capacity. “It can be said to be a limitation of technology-centered strategy,” he pointed out.
To overcome these AI barriers, division director Lee Dong-hoon advised that AI tasks that are aligned with business goals must first be selected and AI technology must be platformed to internalize capabilities. “The preparation, construction, deployment, and monitoring stages must be seamlessly connected like a Möbius strip. Only when this structure is in place can proper innovation occur. Enterprise AI must build an easy and intuitive user environment, in other words, an ‘AI analysis platform for everyone.’ Rather than a project that looks cool on the outside, you should do an AI project that is connected to business. In addition, it is important to look at the mid- to long-term, not just one year, but 2 to 3 years.”
Ingent senior manager Lee Yong-woo gave a presentation on the topic of ‘Flexible data architecture design and success stories using open source DBMS.’ Senior Manager Lee Yong-woo said, “Artificial intelligence and the cloud both have an intersection called data. And DBMS, which handles data, is a key element of infrastructure. Among them, open source DBMS is doing well. “Starting in 2021, open source DBMS is overtaking commercial DBMS.”
However, there are still many companies that are hesitant about open source. Senior Manager Lee Yong-woo said, “It is true that there are concerns about open source due to the burden of having to solve everything on one’s own, including installation, development, and response to failures, lack of technical documentation, lack of convenience functions necessary for operation management, and lack of experts. “This is also why we need a professional technical support company that provides enhanced stability, management convenience, and performance optimization.”
Senior Manager Lee Yong-woo continued, “Also, when you think of open source, you think it is suitable for small and light work, but that is not true. Ingent handled 120TB of data while building a patient data processing system for domestic medical institutions. “If you cooperate with a professional company, you can enjoy many advantages by utilizing the open source DBMS that many companies are already using for a wide range of tasks,” he added.
In the eighth expert session, HPE Director Kim Il-geun began his presentation with the topic of 'Creating digital competitive advantage through cloud-native data analysis.' “There are problems companies face in data processing,” he said. This is data integration. Today, data exists in a variety of places, from on-premises to the cloud and the edge. In fact, it was found that 80% of enterprise data resides outside the data center. “How can we integrate and process this scattered data?”
Director Kim Il-geun suggested that future data and analysis platforms should be built in a 'hybrid native' manner, and mentioned the HPE Esmeral platform, which optimizes and processes data and provides a consistent analysis experience. He added, “With a hybrid-native platform, we can optimize data, support enterprise-scale analytics teams, and unlock the value of data.”
Jaemin Park, manager of Sangpo Technology, gave a presentation on the topic of ‘New cloud service model for cloud subscription companies.’ Manager Jaemin Park cited ▲technology gap, ▲cost optimization, ▲management complexity, and ▲new security as the problems that companies face when adopting the cloud, adding, “60% of companies have adopted a cloud-first strategy by 2023 to respond to these problems. “We will transition to a cloud smart strategy,” he said.
According to Manager Park Jae-min, the cloud-first strategy is to use the cloud as a core platform. A cloud smart strategy optimizes the use of the cloud and leverages it with other technologies and platforms to support business. “Therefore, subscription cloud platforms that offer distributed cloud data, dedicated resources, data security, full-stack security protection, new security accountability models, online and on-site services, and one-on-one expert support are noteworthy,” he explained.
Daniel Wu, head of AI & ML at JPMorgan Chase, gave a presentation on the topic of ‘Responsible AI (RAI).’ ⓒFoundry
In the last session, Daniel Woo, head of AI & ML at JPMorgan Chase, began his talk on the topic of ‘Responsible AI (RAI).’ Daniel Wu, who mentioned the accelerating artificial intelligence at Cloud & AI Innovation 2022, said that now, a year later, the AI operation aspect is being discussed a lot. What is needed to practically apply and operate AI is being discussed.
Among them, with an emphasis on governance, “56% of companies were found to be struggling with governance security and auditing issues. 95% of companies said they were concerned about the potential ethical risks of AI. This is because artificial intelligence can be applied to automated decision-making on a large scale. “It must be explainable to be trustworthy, and trust must be established for users to truly accept AI,” explained Daniel Wu.
Daniel Wu said, “Responsible AI must be integrated into the end-to-end process from artificial intelligence development to deployment. “In particular, a company-wide effort is needed regarding responsible AI,” he said. “Initially, we were only interested in the performance of artificial intelligence models. “Now, we must carefully review whether AI will cause harm to users or society in the process of operating it,” he recommended.