From inferring situations in photos to analyzing tables and graphs, and solving math geometry problems
Example related to understanding image charts of Naver ClovaX. Provided by Naver
, ‘[Financial News] Naver’s interactive artificial intelligence (AI) agent ClovaX will enhance its visual information processing capabilities through a service update on the 27th. Furthermore, Naver aims to enhance its generative AI-based voice synthesis technology by making its Hyper ClovaX model capable of processing images, voices, and texts simultaneously, thus improving its competitiveness in generative AI technology.’,
, ‘According to Naver on the 22nd, with the update of ClovaX’s image understanding capabilities on the 27th, users can engage in conversations with AI based on the information extracted from images uploaded to the ClovaX chat window and the queries entered.’,
, ‘ClovaX can perform various instructions, such as describing phenomena or inferring situations in photos. For example, ClovaX can understand and analyze tables and graphs presented in images or illustrations. Expectations are high that with the enhancement of image understanding capabilities, ClovaX will expand its range of applications beyond tasks such as logical writing, code creation, and translation, positioning itself as a tool for enhancing personal productivity. ‘,
Example image related to Naver ClovaX understanding codes and generating ideas. Provided by Naver
, ‘In particular, Naver’s proprietary giant language model (LLM), Hyper ClovaX, provides more accurate and reliable services by combining various features. According to Naver, when 1,480 exam questions for Korean elementary, middle, and high school exams were inputted in image form to the AI model, ClovaX achieved an accuracy rate of around 84%, surpassing OpenAI’s GPT-4o (78%).’,
, ‘Naver also introduced its Hyper ClovaX-based voice AI technology through a tech blog. With an advanced model that goes beyond conventional speech recognition and synthesis technologies, it can engage in natural conversations with improved language structure, pronunciation accuracy, and even emotional expressions using LLM’s exceptional contextual understanding and command interpretation capabilities.’,
, ‘Operating various voice AI services such as AI voice note ‘Clova Note,’ AI well-being call ‘Clova Care Call,’ and AI voice synthesis ‘Clova Dubbing,’ Naver plans to provide more convenient services with its voice multi-modal LLM technology.’,
, ‘Sung Nak-ho, head of Naver Cloud’s Hyper Scale AI technology, stated, “Starting as a giant language model, Hyper ClovaX has evolved into a large vision language model (Large Vision Language Model) augmented with image understanding capabilities, and further into a speech multi-modal language model.” He added, “The enhanced capabilities of Hyper ClovaX will create new user value by introducing them to various Naver services, offering them as enterprise AI solutions, and further expanding the ecosystem of Hyper ClovaX.”‘,
, ‘Furthermore, Naver plans to actively practice “AI safety” in the process of advancing Hyper ClovaX into a multi-modal LLM and applying it to services. Naver’s AI Safety Framework (ASF), announced in June, evaluates potential risks of AI systems and plans to conduct multidimensional reviews for providing safer voice AI services. ‘,
, ‘#Naver #Multi-Modal #ClovaX ‘,
,