Sources
Loading...
Additional media
Loading...
DeepNewz, mobile.
People-sourced. AI-powered. Unbiased News.
Researchers at Stanford University, including M Yuksekgonul, F Bianchi, J Boen, S Liu, and Z Huang, have introduced TEXTGRAD, a novel framework designed to optimize AI systems by backpropagating textual feedback provided by large language models (LLMs). This innovative approach, described as 'automatic differentiation via text,' has demonstrated significant improvements in performance metrics, achieving a 51% to 55% increase on GPQA and a 20% relative gain in LeetCode-Hard. TEXTGRAD, which has been likened to PyTorch-for-text, is poised to enhance various applications, including designing new molecules and improving medical treatments.
Researchers at Stanford Introduce TEXTGRAD: A Powerful AI Framework Performing Automatic “Differentiation” via Text https://t.co/FK6at2OnL1 #AI #AIframework #AIautomation #AIsolutions #Textgrad #ai #news #llm #ml #research #ainews #innovation #artificialintelligence #machinel… https://t.co/H0HkV9ly6r
With LLMs becoming a common tool today, we've innovated a way to enhance responses by introducing automated "differentiation" via text. Super excited to be part of this project to build PyTorch-for-text!🔥 https://t.co/k7fYsjwhno
⚡️This is the most fun project! We built PyTorch-for-text! 🔥 #TextGrad: automated "differentiation" via text to optimize AI systems by backpropagating LLM text feedback. TextGrad + GPT4o: 💻LeetCodeHard best score ❓GPQA sota 🧬Designs new molecules 🩺Improves treatments 🧵 https://t.co/eFLqVM4VH9