default author photo

Deng Xiong

Stevens Institute of Technology

Hoboken, NJ 07030

United States

SCHOLARLY PAPERS

1

DOWNLOADS

13

TOTAL CITATIONS

0

Scholarly Papers (1)

1.

A Composite Path and Answer Self-Scoring Reward Mechanism for Test-Time Reinforcement Learning

Number of pages: 19 Posted: 06 Feb 2026
Sichuan University, Sichuan University, Sichuan University, Stevens Institute of Technology, Sichuan University, Sichuan University, Sichuan University and Hong Kong University of Science & Technology (HKUST)
Downloads 13 (1,548,648)

Abstract:

Loading...

Reinforcement Learning, Test time adaption, pseudo label, calibration, Unsupervised Learning