mlsanityData to 14 Oct 2026

Exploiting answer-invariant redundancies in satellite imagery for efficient VLM inference on edge

arXiv 2609.29029 ↗First posted 24 Sep 2026cs.CV cs.LGcited by 0

New in Multimodal LLMs / vision-language models; no direction yet. Multimodal LLMs / vision-language models

Abstract (arXiv)

Onboard vision-language models could enable satellites to answer queries directly, but exhaustive tiled inference over high-resolution imagery is slow and energy-intensive. We identify answer-invariant token redundancy (AITR): image tiles and vision tokens that can be removed without changing the final answer. We present Rift, a two-stage system that performs query-conditioned tile pruning followed by elastic prefill to reduce token budget. We evaluate it on LLaVA-1.5 7B running on Jetson AGX Orin. Compared with exhaustive tiled inference, Rift reduces energy by 78% and latency by 69%, while increasing accuracy from 45% to 73%.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 11 of its 20 references that are in the map. First posted on arXiv 24 Sep 2026.

Builds on

11 of 20 references.

References · 11

  1. 2023Improved Baselines with Visual Instruction Tuningcited by 4,418
  2. 2024An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Modelscited by 642
  3. 2023RemoteCLIP: A Vision Language Foundation Model for Remote Sensingcited by 363
  4. 2024VisionZip: Longer is Better but Not Necessary in Vision Language Modelscited by 313
  5. 2018xView: Objects in Context in Overhead Imagerycited by 146
  6. 2024ATP-LLaVA: Adaptive Token Pruning for Large Vision Language Modelscited by 89
  7. 2025When Large Vision-Language Model Meets Large Remote Sensing Imagery: Coarse-to-Fine Text-Guided Token Pruningcited by 35
  8. 2025VFlowOpt: A Token Pruning Framework for LMMs with Visual Information Flow-Guided Optimizationcited by 22
  9. 2025CATP: Contextually Adaptive Token Pruning for Efficient and Enhanced Multimodal In-Context Learningcited by 18
  10. 2023Learning to Holistically Detect Bridges from Large-Size VHR Remote Sensing Imagerycited by 17
  11. 2025EarthSight: A Distributed Framework for Low-Latency Satellite Intelligencecited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026SA-GEM: Scale-Adaptive and Geospatial Evidence-Modulated Token Pruning for Efficient Remote Sensing Large Vision-Language Models
  2. 2026Summarize First, Download Later: Onboard VLMs for Bandwidth-Efficient Earth Observation
  3. 2025Enabling Near-realtime Remote Sensing via Satellite-Ground Collaboration of Large Vision-Language Models
  4. 2026A Unified Framework for Efficient Remote Sensing Visual Question Answering: Adapting Dual, Hybrid, and Encoder-Decoder Architectures
  5. 2025Few-Shot Vision-Language Reasoning for Satellite Imagery via Verifiable Rewards
  6. 2026RS-OPSD: Reliable Privileged On-Policy-Self-Distillation for Ultra-High-Resolution Remote Sensing VQA
  7. 2026TerraAgent: Demonstration-Grounded Multi-Agent Reasoning for Long-Context Remote Sensing Visual Question Answering
  8. 2026Open-SAT: LLM-Guided Query Embedding Refinement for Open-Vocabulary Object Retrieval in Satellite Imagery