9 stories tagged Self Improving
Researchers introduced EliteTable, a self-improving program-search algorithm guided by large language models (LLMs). In evaluations across six Gymnasium control tasks, median candidate scores rose between the first and final generations in all six tasks, demonstrating consistent improvement.
An Anthropic researcher demonstrated an AI system that improved its performance on all 10 misaligned-behavior benchmarks it was tested on, without any loss in overall capability and without human intervention.
Mirendil has partnered with Google Cloud in a $100 million-plus deal to expand its infrastructure for developing self-improving AI systems. These AI models aim to accelerate scientific research and AI advancements.
CERN's Genesis Mission aims to develop and deploy self-improving AI models that can autonomously analyze particle physics data, potentially accelerating breakthroughs in high-energy physics and beyond.
A new survey on arXiv frames self-improving AI agents as adaptive systems that convert experience into capability gains with minimal human input. The research defines modern agents as a configuration coupling a foundation model with an operational scaffold of prompts, memory, tools, and control logic.
Scientists studied how AI agents update their own tools and strategies. They found that not all AI models benefit equally from these self-improvements, even if they're good at their original tasks.
A $650 million startup wants to create AI that can research and improve itself. The founder insists this isn't just a vision but a reality in the making. This could revolutionize how AI develops and scales.
Researchers developed a new AI algorithm called FREIA that helps large language models improve their reasoning skills on their own. This could lead to smarter AI assistants that learn and adapt without constant human supervision.
Laimark is an 8B parameter LLM designed to self-improve, and it runs on consumer-grade GPUs. This development could democratize advanced AI capabilities for individual developers and small teams.