LG Uplus officially releases ixi-GEN, a small generative AI model that can be fine-tuned locally

Major mobile operators in South Korea LG Uplus On Tuesday, it officially launched its generative artificial intelligence model ixi-GEN, joins its two larger rivals, KT Corp. and SK Telecom Co., in the race to lead the B2B generative AI market.

LG Uplus officially releases ixi-GEN, a small generative AI model that can be fine-tuned locally

LG Uplus said its generative AI is a small large language model (sLLM) trained on its communications and platform data, using the ultra-large-scale multimodal AI Exaone developed by LG AI Research, an artificial intelligence lab under LG Group.

Unlike general-purpose LLMs such as OpenAI's ChatGPT, sLLMs are smaller in size and focus on specific industries or businesses. As a lighter LLM, ixi-GEN can be fine-tuned relatively quickly to better perform specific tasks or behaviors for specific industry purposes. It has high security because it can be installed on a local server instead of in the cloud.

According to the company, it is also less likely to face copyright infringement lawsuits because it is trained on clean data provided by Exaone. LG Uplus will apply ixi-GEN to eight of its services, including chat agent and mobile store advisor services, within this year and then make it available to the B2B market. Its generative AI is a telecom-specific AI model that can be quickly applied to telecom services.

In addition to generative AI tools, the company has also launched an AI development platform called ixi Solution, which enables ordinary people to use AI in the real world.

statement:The content is collected from various media platforms such as public websites. If the included content infringes on your rights, please contact us by email and we will deal with it as soon as possible.
Information

Apple launches all-around visual model 4M-21 that can handle 21 different modalities

2024-6-26 8:54:58

Information

Weimob launches AI application product WAI Pro to provide customized large-scale model application capabilities

2024-6-26 8:57:59

Search