Computer Use Agents
PulseAugur coverage of Computer Use Agents — every cluster mentioning Computer Use Agents across labs, papers, and developer communities, ranked by signal.
7 day(s) with sentiment data
-
Together and yutori_ai host CUA Night event on computer-use agents
Together and yutori_ai are co-hosting an event called CUA Night this Thursday. The event will focus on computer-use agents and cover topics such as research, models, data, RL environments, and inference.
-
AI agents prioritize explicit contracts and workflows over demos
The development of computer-use agents is advancing rapidly, with a focus shifting from visually appealing demonstrations to practical implementation. Success in this field will likely hinge on the ability to translate …
-
New 'PRAC' attack targets vision modality of AI agents
Researchers have developed a new attack called PRAC that targets the vision modality of multimodal foundation models, specifically Computer Use Agents (CUAs). Unlike previous attacks that directly manipulated model outp…
-
New benchmark tests AI agent safety against multi-step prompt injection attacks
Researchers have introduced StepJack, a new benchmark designed to test the safety of computer-use agents (CUAs) against multi-step indirect prompt injection attacks. These attacks involve distributing adversarial instru…
-
New AI Agent 'heuse+' Learns to Control Mouse and Keyboard
A new AI agent called heuse+ is being developed to control a computer's mouse and keyboard, aiming to automate tasks. This agent is part of a broader trend in AI development focused on creating more autonomous systems t…
-
New "Invisible Ink Threats" Expose Security Risks in AI Agents
Researchers have identified a new category of security threats targeting computer-use agents (CUAs), which are AI systems designed to operate operating systems and the web. These "Invisible Ink Threats" involve low-harm…
-
New research paper details design space for AI agent user experiences
A new research paper explores the user experience (UX) of computer use agents, which are generative AI systems designed to automate actions within user interfaces based on commands. The study identifies a design space f…
-
New RL framework uses vision-language models for GUI agent supervision
Researchers have developed a new reinforcement learning framework for Computer-Use Agents (CUAs) that leverages autonomous vision-language evaluation for supervision. This approach addresses the challenge of obtaining s…
-
Anthropic's Claude 3.5 Sonnet offers major speed gains; WeaveBench reveals agent limitations
Anthropic has released Claude 3.5 Sonnet, a new AI model that is twice as fast as its predecessor, Claude 3 Opus, while maintaining or improving performance. This advancement is significant for applications requiring ra…
-
New HiViG critic improves AI agents' GUI performance with history and vision
Researchers have developed HiViG, a novel framework designed to improve the performance of Computer Use Agents (CUAs) in complex graphical user interface environments. HiViG addresses limitations in existing critics by …
-
New research tackles safety and efficiency in computer-use agents · 6 sources tracked
Recent research is exploring the safety and efficiency of computer-use agents (CUAs). One paper introduces MisActBench and a guardrail called DeAction to detect and correct misaligned actions, significantly reducing att…
-
New frameworks aim to improve AI understanding of user intent
Two new research papers introduce computational frameworks for understanding and controlling user intent in AI interactions. The first, 'Intent Signal Theory,' formalizes the distinction between a user's latent intent a…
-
New benchmark CUActSpot targets complex interactions for AI agents
Researchers have introduced CUActSpot, a new benchmark designed to evaluate computer-use agents (CUAs) on complex and infrequent interactions across multiple modalities. The benchmark addresses the long-tail issue in GU…