xAI, Elon Musk’s artificial intelligence venture, has unveiled an improved version of its Grok 1.5 model called Grok 1.5 Vision. This improved model now includes computer vision capabilities to understand and answer questions about images. This update comes shortly after OpenAI introduced the GPT-4 model, which also offers computer vision capabilities.
The announcement of this improvement was made via X’s official xAI account (formerly Twitter), where details about the model’s recent features were shared in a blog post. While the core features of Grok 1.5 remain unchanged in this updated version, the addition of vision capabilities is expected to expand its capabilities to interact with the real world.
xAI conducted benchmarking to evaluate Grok 1.5 Vision’s performance on a variety of metrics, including its proprietary RealWorldQA benchmark, which assesses the model’s understanding of real-world spatial concepts. Additionally, the model was evaluated in other tests such as MMMU and ChartQA. It is worth noting that in the RealWorldQA test, Grok outperformed OpenAI’s GPT-4 with Vision and Google’s Gemini 1.5 Pro, although it showed worse performance in other evaluations.
Computer vision is an electrifying field of computer science that focuses on enabling computers, including AI models, to identify and understand real-world objects using images and videos. Its goal is to provide machines with vision capabilities similar to those of humans.
The largest technology companies are investing heavily in the development of artificial intelligence models equipped with vision functions. Google Gemini 1.5 Pro and OpenAI GPT-4 with Vision are the leading competitors in this field.
The potential applications of computer vision are huge and transformative. For example, Healthify, an Indian calorie and nutrition tracking platform, recently introduced a feature called “Snap.” Users can take photos of food items, and artificial intelligence suggests healthier recipe changes and exercise plans to balance calorie intake. Computer vision also shows promise in medical diagnostics, autonomous vehicles, and more.
Posted: Apr 15, 2024 8:11 pm EST
