AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

RMultiplex200K: Toward Reliable Multimodal Process Supervision for Visual Language Models on Telecommunications

Sijia Chen, Bin Song

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Chain-of-thought multimodal reasoning · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
fine-tune on RMultiplex200KQwen-2-VL-72B · RMultiplex200K testbedaccuracy (%): 21.3 -> 36.8Improved
add TC-Navigator process reward verifier (best-of-100)fine-tuned Qwen-2-VL-72B · RMultiplex200K testbedaccuracy (%): 36.8 -> 46.8Improved
add TC-Navigator verifier with self-consistencyfine-tuned Llama-3.2-90B · RMultiplex200K testbedaccuracy (%): 39.2 -> 46.1Improved
use process reward model (TC) instead of outcome reward model (ablation)fine-tuned Qwen-2-VL-72B · RMultiplex200K testbedaccuracy (%): 44.5 -> 46.8Improved
use PRM instead of ORM on problems with more reasoning steps (ablation)PPO-reinforced Llama-3.2-11B · RMultiplex200K CSP categoryaccuracy (%): 31.9 -> 36.2Improved
use PRM instead of ORM on problems with fewer reasoning steps (ablation)PPO-reinforced Llama-3.2-11B · RMultiplex200K SC categoryaccuracy (%): 35.8 -> 37.9Improved
add RA1 to TC-Navigator verifier (ablation)TC-Navigator PPO · RMultiplex200K testbedaccuracy (%): 36.3 -> 37Improved
add RA2 to TC-Navigator verifier (ablation)TC-Navigator PPO · RMultiplex200K testbedaccuracy (%): 35.9 -> 37Improved
reinforce generator with step-by-step PPO using TC rewardLlama-3.2-11B · RMultiplex200K testbedaccuracy (%): 34 -> 37Improved
combine PPO-reinforced generator with TC+SC verificationLlama-3.2-11B · RMultiplex200K testbedaccuracy (%): 37 -> 43.3Improved
train verifier on ApPA labels instead of MATH-Shepherd labels (ablation)TC verifier Qwen-2-VL-72B · RMultiplex200K testbedaccuracy (%): over 6%Improved
use TC verifier trained on RMultiplex200K for external benchmarkQwen-2-VL-72B · TeleQnAaccuracy (%): 66.3 -> 76.8Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy