[
    {
        "id": "osp-15146",
        "type": "article-journal",
        "title": "SafeInferCom: Safe Inference-Time Compute via Verifier-Guided Mid-Generation Intervention for Robotic Task Planning",
        "author": [
            {
                "family": "Xu",
                "given": "Weizhe"
            },
            {
                "family": "Fan",
                "given": "Jialiang"
            },
            {
                "family": "Liu",
                "given": "Mengyu"
            },
            {
                "family": "Kong",
                "given": "Fanxin"
            }
        ],
        "URL": "https://omanscience.com/en/articles/safeinfercom-safe-inference-time-compute-via-verifier-guided-mid-generation-intervention-for-robotic-task-planning",
        "language": "en",
        "issued": {
            "date-parts": [
                [
                    2026
                ]
            ]
        },
        "abstract": "Large Reasoning Language Models (LRLMs) enable multi-step reasoning for robotic task planning, but continued reasoning can overwrite valid intermediate plans or leave constraint violations unresolved, reducing planning reliability and wasting inference-time computation. We develop an inference-time monitor that exposes and verifies intermediate plans without disrupting the original decoding trajectory. Building on this monitor, we propose SafeInferCom, a formal verifier-guided framework that preserves valid intermediate plans and directs error correction during generation. Experiments across multiple LRLMs and planning domains reveal reasoning-response inconsistency and limited self-correction under one-shot inference. SafeInferCom improves planning success and accelerates error correction relative to one-shot inference. When combined with iterative refinement, it further improves success while reducing token usage compared with refinement alone. We additionally evaluate SafeInferCom in VirtualHome and provide a real-world robotic-arm demonstration."
    }
]