3 rising AI/ML startups from South Korea
VietReader 19-11-2020, 09:36

The South Korean government announced a renewed focus on artificial intelligence early this year as part of the country’s plans for economic recovery post-Covid-19.

AI is already a key part of South Korea’s national strategy, but the nation wants to be one of the region’s top AI powerhouses by 2030. That might be a good idea, as AI and machine learning (ML) could address many of the challenges faced by Korean society, creating opportunities for startups to plug the gaps along the way.

We hear from three rising AI/ML startups in South Korea on their journey to tap into this burgeoning market.

The following responses have been edited for brevity and clarity.

1. Scatter Lab

3 rising AI/ML startups from South Korea

Jongyoun Kim, founder of Scatter Lab / Photo credit: Scatter Lab

Founded in 2011 by Jongyoun Kim, Scatter Lab is a developer of mobile application software that aims to humanize interactions between people and AI.

What is your core business or tech use case for which you use an AI or ML-based solution?

Scatter Lab is a startup that develops open-domain conversation technology, which allows AI to communicate freely with people just like a close friend. In the next five years, we aim to create a conversational AI that people prefer as a conversation partner over a person.

Open-domain chatbots are one of the most difficult and challenging areas in natural language processing (NLP) research, and it has not been studied much. However, rapid advancements in NLP in recent years have enabled developers to create an open-domain chatbot at a significant level. For example, big tech companies like Google and Facebook have unveiled open-domain chatbots such as Meena and Blender, respectively.

Scatter Lab has the most one-on-one open-domain conversation data between people, with an exclusive database of 10 billion Korean KakaoTalk chat messages and 1 billion Japanese Line chat messages. Based on the strength of this data retention, we are developing a state-of-the-art open-domain conversation technology. Our open-domain chatbot “Luda,” which will be released at the end of 2020, will highlight how far NLP and open-domain technology has come in the age of AI.

How has being on the cloud and AWS helped your startup scale fast?

Creating a great open-domain conversational AI is an extremely difficult challenge. You need to run many kinds of experiments efficiently and often at the same time and train large models. We had to do a large-scale distributed neural network training, but it was really difficult to configure the infrastructure. AWS provided the necessary infrastructure for distributed learning that we could use easily.

What are your plans for the future?

Luda will finish its beta tests soon and will be released later this year. Our goal is to make Luda the AI that has the most conversations with people in the world, and we’re pursuing our tech and product development with that goal in mind.

We don’t think it’s too far away for AI to have human-level communication skills. That way, AI – along with humans and pets – will be another great option to meet people’s social needs.

2. SmartMind

The SmartMind team / Photo credit: SmartMind

SmartMind, which was founded by Brandon Sangsoo Lee, is developing a campaign management platform that allows advertisers to analyze the performance of digital marketing campaigns more effectively.

What is your core business or tech use case for which you use an AI or ML-based solution?

SmartMind provides a full-stack digital transformation service for companies of any size to amplify revenue with digital marketing. With our advertising management platform, Lighthouse, we are laser-focused on building algorithms for advanced and accurate audience targeting and optimized media planning.

How has being on the cloud and AWS helped your startup scale fast?

As our services grew, we had to focus on so many things. But with the cloud, we could easily expand and manage the infrastructure to scale up, which helped us focus on actual development without worrying about management and maintenance. It has been very helpful for rapid growth. Plus, Amazon Web Services has plenty of support, which creates an open-source and friendly cloud development environment that gives us full flexibility to try out any algorithm.

What are your plans for the future?

On the product side, we will continue to tailor algorithms to maximize sales revenue with two to three more product pilots. Soon, the platform will provide easy-to-use digital advertisement management dashboards and will come with business intelligence to give useful insights.

We started with fashion retailers. Now, we are providing digital marketing solutions to Korea’s no. 1 energy infrastructure company. In the coming months, we will aggressively add more verticals in our pipeline.

3. Anipen

Photo credit: Anipen

Anipen is a 3D content development platform that enables enterprises to develop video communications and entertainment platforms.

What is your core business or tech use case for which you use an AI or ML-based solution?

We apply AI/ML to our own augmented reality (AR) motion sequence engine, which has its foundations in a sketch-based content authoring technology, where you can easily add motion to characters by simply drawing lines.

We apply AI/ML to automate the content creation process. We analyze the content that users create and convert them to another type of graphic. For example, if you upload content like images using our AR motion sequence engine, the AI/ML engine will analyze the objects and identify features such as a human face, body, etc. in the contents, and you can easily switch the objects into AR objects.

What inspired you to start an AI/ML startup?

We already had the motion sequence engine. So we thought that we could develop a totally different sequence-authoring and motion-rendering engine in combination with deep learning, which can analyze and use the objects in the pictures and videos.

We believe AI/ML will be one of the essential components in AR, virtual reality, and other extended reality (XR) developments. Even AR/XR glasses in the near future may have an image and video-based motion sequence engine.

How has being on the cloud and AWS helped your startup scale fast?

We’re using Amazon SageMaker to help label images for bounding box object detection and use Amazon Mechanical Turk to help us outsource work with public datasets, such as dataset labelling for segmentation and feature extraction.

The APAC AI Conclave is a free, virtual conference presented by Tech in Asia in partnership with Amazon Web Services and will be happening from November 25 to 26. Hear from these companies and more at the event, which will focus on helping AI/ML leaders and practitioners from the region’s startups build smart, customer-centric, and scalable solutions in the cloud using the latest, broadest, and deepest set of machine learning and AI services.

Get your ticket for free here.

This content was produced by Tech in Asia Studios, which connects brands with Asia’s tech community. Learn more about partnering with Tech in Asia Studios.

TAG: