A naukri.com initiative
VentureBeat
4w
268
Image Credit: VentureBeat
OpenScholar: The open-source A.I. that’s outperforming GPT-4o in scientific research
- A new AI system called OpenScholar, built by the Allen Institute for AI (Ai2) and the University of Washington, is helping researchers access, evaluate and synthesise scientific literature fast and with greater confidence.
- OpenScholar has a retrieval-augmented language model that uses a datastore of open-access academic papers to deliver citation-backed, comprehensive answers to complex research questions.
- It differentiates itself from models like OpenAI's GPT-4o by retrieving real papers and synthesising their findings to generate a grounded answer, outperforming GPT-4o with citation accuracy and factuality.
- OpenScholar could democratise access to AI tools for smaller institutions, underfunded labs, and researchers in developing countries, being 100 times cheaper to operate than concurrent proprietary systems.
- The system's datastore is restricted to open-access papers, leaving out paywalled research that dominates some fields, although the researchers hope future iterations can responsibly incorporate closed-access content.
- OpenScholar represents a watershed moment in scientific computing, demonstrating the capacity to process, understand, and synthesise scientific literature with near-human accuracy, ushering in a new era of AI-assisted research.
- The researchers have released everything—code, models, data, and tools—betting that openness will accelerate progress more than keeping their breakthroughs behind closed doors.
- They've answered one of the most pressing questions in AI development: Can open-source solutions compete with Big Tech’s black boxes?
- The answer, it seems, is hiding in plain sight among 45 million papers.
Read Full Article
16 Likes
For uninterrupted reading, download the app