Jiawei Xu

most citedRethinking the Value of Multi-Agent Workflow: A Strong Single Agent Baseline

2 citations · 2 across the 4 of their papers we have counts for

collaborators

Showing cs.LGShow all

2 papers · 1 filter

cs.LG2026

$β$ -OPSD: Deriving with Policy Optimization, Training with Self-Distillation

Jiawei Xu, Minghui Liu, Juzheng Zhang +2

On-policy self-distillation (OPSD) is a promising approach to improve reasoning language models, but it remains brittle in practice: making it work reliably often requires substant…

cs.LG2026

Scheduling Thoughts: Learning the Order of Thought in Diffusion Language Models

Jiawei Xu, Minghui Liu, Aakriti Agrawal +2

Masked diffusion language models decode by iteratively unmasking tokens, where the unmasking order defines an "order of thought" that strongly influences generation quality yet is…

β-OPSD: Deriving with Policy Optimization, Training with Self-Distillation

Scheduling Thoughts: Learning the Order of Thought in Diffusion Language Models

$β$ -OPSD: Deriving with Policy Optimization, Training with Self-Distillation