From Critic to Confidence: PPO for Language-Based Quantitative Prediction with Confidence Estimation
LLMs can perform language-based quantitative prediction from unstructured inputs, but remain susceptible to hallucinations and overconfident errors, making it critical to know not only what a model predicts, but when its predictions can be trusted. We introduce CARE-PPO, a reinforcement learning framework that establishes a connection between loss prediction for uncertainty estimation and actor-critic PPO fine-tuning, enabling joint learning of accurate numerical estimates and reliable confidence signals in language-based quantitative prediction. CARE-PPO uses a Confidence-Aligned Reward for E
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 55%hscspring/rl-llm-nlp →
- PossiblePossibly related (embedding) · 48%Learning to cope with the unexpected: training AI to manage uncertainty | E-pi Project | Results in Brief | H2020 - CORDIS →
- PossiblePossibly related (embedding) · 48%rllm-org/rllm →
- PossiblePossibly related (embedding) · 46%RLHF →
- LinkedLinked via arxiv author · 85%Mehak Dhaliwal →
“From Critic to Confidence: PPO for Language-Based Quantitative Prediction with Confidence Estimation”
- LinkedLinked via arxiv author · 85%Rasta Tadayon →
“From Critic to Confidence: PPO for Language-Based Quantitative Prediction with Confidence Estimation”
- LinkedLinked via arxiv author · 85%Andong Hua →
“From Critic to Confidence: PPO for Language-Based Quantitative Prediction with Confidence Estimation”
- LinkedLinked via arxiv author · 85%Haewon Jeong →
“From Critic to Confidence: PPO for Language-Based Quantitative Prediction with Confidence Estimation”