Alibaba Launches Open Source AI Model That Can Analyze Human Emotions Through Video

VOV.VN - China's Alibaba Group has just launched a new artificial intelligence (AI) model, which is said to be capable of recognizing human emotions through videos, with the ambition of surpassing OpenAI's latest model.

Alibaba’s Tongyi Lab on Tuesday (March 11) released the open source R1-Omni, the first comprehensive multimodal language model that uses reinforcement learning with verifiable rewards (RLVR), a technology that promises to improve AI’s ability to reason and analyze sentiment.

The team said they used RLVR to optimize the open-source HumanOmni 0.5B model, significantly improving three key metrics: inference, emotion recognition accuracy, and generalization. According to the report, R1-Omni can better understand how visual and auditory information support emotion recognition, and clearly identify which factors play a more important role in assessing each emotion.

China's Alibaba Group has just launched a new artificial intelligence (AI) model - Photo: Bloomberg

In test videos , R1-Omni was able to identify a person's emotional state through the video, as well as describe the character's clothing and surroundings. This marks a major step forward in the field of computer vision.

AI emotion recognition technology is becoming an important trend in the technology industry. This technology is often seen in practical applications, such as customer support chatbots or monitoring systems that detect signs of driver fatigue when driving Tesla electric cars.

With the launch of the R1-Omni model, Alibaba is accelerating the AI ​​race and challenging OpenAI's position. While OpenAI launched GPT-4.5, which can recognize and respond to subtle signals in user questions, for a price of up to $200/month, Alibaba offers R1-Omni completely free on the Hugging Face platform.

Although R1-Omni currently only recognizes basic emotions like “happy” or “angry,” according to Alibaba CEO Wu Yongming, artificial general intelligence (AGI) is the company's top goal, and developing emotional intelligence for AI is an important step forward.