Paper Reading by Sylvia from the AI Floor - Open Registration
Sylvia will discuss a paper her team wrote and submitted to NeurIPS 2025 workshop
https://arxiv.org/html/2509.03059v1
We've all seen how LLMs excel in math and coding, but struggle with complex reasoning in fields like physics or finance. Why? A lack of high-quality, verified training data. Human annotation is expensive, and basic synthetic data isn't reliable.
Our new paper introduces a multi-agent system to automatically generate and verify reliable synthetic data.
🔎 Here's how it works:
1️⃣ Seed Dataset: Start with a small, high-quality set of human-created problem-solution (code) pairs.
2️⃣ Synthesizer: Use a multi-agent system to automatically generate new problems and code solutions.
3️⃣ Verifier: We use a two-pronged approach. A code-based solution is only accepted if it's verified by both code execution and a language-based CoT verifier. This ensures correctness at scale.
4️⃣ Reinforcement Learning: Agents are rewarded only for generating semantically equivalent, correct answers.
This method allows us to scale high-quality data creation, tackling the core challenge of complex reasoning in LLMs.
This event is hosted at the Frontier Tower:
We are transforming a 16-floor tower in San Francisco into a self-governed vertical village—a hub for frontier technologies and creative arts. Tier-one labs presenting AI, Ethereum, biotech, neuroscience, longevity, robotics, human flourishing, and arts & music. These floors will house innovators and creators pushing the boundaries of human potential in a post-AI-singularity world.
Apply here for founding citizenship: https://frontiertower.io/apply
Why should I become a citizen?
Be part of creating the first self-governed vertical village
Connect with the most creative people in the city
Get access to all floors, free event space & movement floor
Website: https://frontiertower.io/
Need more reading? Visit https://frontiertower.notion.site/