{"slug": "terraagent-demonstration-grounded-multi-agent-reasoning-for-long-context-remote", "title": "TerraAgent: Demonstration-Grounded Multi-Agent Reasoning for Long-Context Remote Sensing Visual Question Answering", "summary": "A training-free framework called AVA, presented in arXiv paper 2610.12517v1, wraps a frozen remote sensing multimodal large language model with three modules to improve remote sensing visual question answering. On RSVQA-LR/HR, VRSBench, and LHRS-Bench, AVA raised average accuracy by up to 2.40 points over the strongest same-setting baselines, lifted the visual-dependency score from 0.28 to 0.49, cut the prior-only answer rate from 46.2 percent to 21.7 percent, and lowered the hallucination rate from 17.1 percent to 9.8 percent, at 2.6x the cost of a single backbone forward pass and with no fine-tuning.", "body_md": "arXiv:2610.12517v1 Announce Type: new \nAbstract: Remote sensing visual question answering (RS-VQA) requires answers that truly depend on image evidence, yet current remote sensing multimodal large language models (RS MLLMs) are still evaluated mainly by accuracy, which cannot reveal whether a correct answer was produced without using the image. This setting introduces three challenges: demonstrations retrieved across different ground sample distances (GSD), sensors, and city textures may be poorly matched; real queries often involve long visual contexts from multi-scale tiles, multi-temporal frames, and retrieved knowledge, where evidence can be diluted and errors can propagate through multi-step reasoning; and prior-only answering remains invisible to standard metrics. We present AVA, a training-free framework that wraps a frozen RS MLLM with three modules. GDR performs GSD-bucketed, scene-graph-based demonstration retrieval and filters demonstrations whose answers are unsupported by their images. MACE decomposes each query into an evidence-typed sub-question DAG, routes sub-questions to suitable perception tools, and records verified provenance in a shared evidence ledger with bounded re-perception. CVA audits the final answer through counterfactual perturbations of cited evidence, including ablation and swap, and computes a visual-dependency score for confidence calibration and evidence-forcing re-query. On RSVQA-LR/HR, VRSBench, and LHRS-Bench, AVA improves average accuracy by up to 2.40 points over the strongest same-setting baselines, raises the visual-dependency score from 0.28 to 0.49, reduces the prior-only answer rate from 46.2 percent to 21.7 percent, and lowers the hallucination rate from 17.1 percent to 9.8 percent, with 2.6x the cost of a single backbone forward pass and no fine-tuning.", "url": "https://wpnews.pro/news/terraagent-demonstration-grounded-multi-agent-reasoning-for-long-context-remote", "canonical_source": "https://arxiv.org/abs/2610.12517", "published_at": "2026-10-12 04:00:00+00:00", "updated_at": "2026-10-12 04:42:25.934358+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "large-language-models", "ai-research", "ai-agents"], "entities": ["AVA", "RSVQA-LR/HR", "VRSBench", "LHRS-Bench", "arXiv"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/terraagent-demonstration-grounded-multi-agent-reasoning-for-long-context-remote", "markdown": "https://wpnews.pro/news/terraagent-demonstration-grounded-multi-agent-reasoning-for-long-context-remote.md", "text": "https://wpnews.pro/news/terraagent-demonstration-grounded-multi-agent-reasoning-for-long-context-remote.txt", "jsonld": "https://wpnews.pro/news/terraagent-demonstration-grounded-multi-agent-reasoning-for-long-context-remote.jsonld"}}