Apple researchers have developed an AI system called ReALM that can understand screen context and ambiguous references, improving interactions with voice assistants.
ReALM reconstructs the screen using parsed on-screen entities to generate a textual representation, outperforming GPT-4.
Apple is investing in making Siri more conversant and context-aware through this research.
However, automated parsing of screens has limitations, especially with complex visual references.
Apple is catching up in AI research but faces stiff competition from tech rivals like Google, Microsoft, Amazon, and OpenAI.
Source: https://venturebeat.com/ai/apple-researchers-develop-ai-that-can-see-and-understand-screen-context/
[link] [comments]