Alibaba has introduced a new artificial intelligence model that surpasses its previous products in comprehending images and engaging in more intricate conversations, intensifying the global competition for technological dominance.
The Chinese tech giant unveiled two novel models, Qwen-VL and Qwen-VL-Chat, which will be available as open source resources.
This means that individuals, educational institutions, and businesses globally can utilize these models to build their own AI applications, eliminating the need to train their own systems and saving time and resources.
The Qwen-VL model is capable of addressing open-ended queries related to various images and generating descriptive captions for them. On the other hand, Qwen-VL-Chat caters to more intricate interactions, such as analyzing multiple image inputs and responding to multiple rounds of questions. Some of the tasks Qwen-VL-Chat can accomplish include crafting stories and producing images based on user-inputted photos, as well as solving mathematical equations depicted in pictures.
One illustrative instance provided by Alibaba involves an image of a hospital sign in Chinese. The AI can interpret the image of the sign to answer inquiries about the locations of specific hospital departments.
While generative AI, where technology generates responses based on human input, has predominantly centered on textual content, the most recent version of OpenAI’s ChatGPT can also grasp images and reply in text, resembling Qwen-VL-Chat’s capabilities.
Alibaba’s latest models build upon its extensive language model named Tongyi Qianwen, which was launched earlier in the year. Large language models (LLMs) are AI models trained on vast datasets and form the foundation for chatbot applications.
In addition to these developments, Alibaba recently made two other AI models open source.
While this move doesn’t yield licensing fees for Alibaba, the distribution of these open-source models will expand the user base for its AI technology, a strategic move at a time when the company’s cloud division is seeking renewed growth as it approaches its public listing.
