- Flitto and Upstage sign MOU for AI language data construction.
- Focus on underrepresented Asian languages like Thai, Japanese, Lao, and Khmer.
- Initiatives include benchmark dataset creation, multilingual leaderboard operation, and LLM localization.
- Flitto brings expertise in language data scaling and copyright issue-free datasets.
- Upstage aims to expand language support for its LLM Solar, targeting Southeast Asian languages.
- CEOs of both companies emphasize the importance of quality language data for AI advancements.
Main AI News:
In a strategic move within the AI landscape, Flitto, a prominent AI data solution provider, has announced a momentous memorandum of understanding (MOU) with Upstage, an AI powerhouse specializing in expansive language models. This collaboration aims to spearhead the development of AI language data, particularly focusing on underrepresented languages in Asia.
The joint endeavor between these two entities signifies a pivotal step towards amplifying the capabilities of large language models (LLMs) by harnessing diverse datasets. With a primary focus on languages like Thai, Japanese, Lao, and Khmer, the partnership aims to bolster the performance metrics of LLMs, thereby advancing the realm of natural language processing.
Central to the agreement are several pivotal initiatives. Firstly, both companies will collaborate on crafting a benchmark dataset for Ko-LLM, fortifying the foundations for Korean LLMs. Additionally, they will jointly oversee the operation of a multilingual LLM leaderboard, fostering a culture of innovation and collaboration within the AI community. Moreover, the localization of LLMs through the utilization of low-resource language data will be a key focus area, aiming to democratize access to advanced language technologies across diverse linguistic landscapes. Lastly, the partnership will extend to data provision for the establishment of small language models (sLLMs), facilitating seamless integration into various business applications.
This strategic alignment between Flitto and Upstage underscores a shared commitment to driving advancements in language AI. Leveraging Flitto’s robust capacity for language data scaling and cutting-edge AI technology, the partnership is poised to deliver transformative outcomes in the near term. With Flitto’s proven track record in language data construction and provision of copyright issue-free datasets, the collaboration is primed to unlock new frontiers in AI-driven language innovation.
For Upstage, the collaboration represents a strategic imperative in expanding the language support of its preeminent LLM Solar. By securing high-quality, low-resource language data, Upstage aims to tailor Solar’s capabilities to cater to a diverse array of regional languages, particularly those prevalent in Southeast Asia. With an existing support base for Korean and English, Solar’s roadmap includes an expansion to encompass Japanese and Thai languages by year-end, further solidifying its position as a frontrunner in AI-driven language technologies.
Simon Lee, CEO of Flitto, expressed optimism about the partnership’s potential impact, stating, “Training on low-resource languages has emerged as a key factor in the performance of large language models.” He emphasized, “Through the partnership between the two companies, we aim to demonstrate how the synergy between high-quality AI data and advanced technologies can positively influence the domestic generative AI ecosystem.”
Echoing Lee’s sentiments, Sung Kim, CEO of Upstage, underscored the critical importance of quality language data in driving the generative AI revolution. Kim remarked, “Together with the generative AI boom driven by language models, securing quality language data is an essential task,” adding, “Through our partnership with Flitto, Upstage will endeavor to build advanced datasets, enabling a wider global audience to experience the innovation brought by generative AI technology.”
Conclusion:
The collaboration between Flitto and Upstage signifies a strategic push towards advancing language AI capabilities. By focusing on underrepresented languages and leveraging expertise in data construction and AI technology, the partnership is poised to drive significant innovation in the generative AI landscape. This alliance underscores the growing importance of diverse language datasets in fueling the next wave of AI-driven advancements, positioning both companies at the forefront of this transformative market.
