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

arXiv: https://arxiv.org/abs/2609.29029 · HTML: https://mlsanity.com/paper/2609.29029/

## 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.

## Placed in

- Multimodal LLMs / vision-language models

## Builds on (most cited first, 11 of 11)

- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models](https://mlsanity.com/paper/2403.06764/) (2024)
- [RemoteCLIP: A Vision Language Foundation Model for Remote Sensing](https://mlsanity.com/paper/2306.11029/) (2023)
- [VisionZip: Longer is Better but Not Necessary in Vision Language Models](https://mlsanity.com/paper/2412.04467/) (2024)
- [xView: Objects in Context in Overhead Imagery](https://mlsanity.com/paper/1802.07856/) (2018)
- [ATP-LLaVA: Adaptive Token Pruning for Large Vision Language Models](https://mlsanity.com/paper/2412.00447/) (2024)
- [When Large Vision-Language Model Meets Large Remote Sensing Imagery: Coarse-to-Fine Text-Guided Token Pruning](https://mlsanity.com/paper/2503.07588/) (2025)
- [VFlowOpt: A Token Pruning Framework for LMMs with Visual Information Flow-Guided Optimization](https://mlsanity.com/paper/2508.05211/) (2025)
- [CATP: Contextually Adaptive Token Pruning for Efficient and Enhanced Multimodal In-Context Learning](https://mlsanity.com/paper/2508.07871/) (2025)
- [Learning to Holistically Detect Bridges from Large-Size VHR Remote Sensing Imagery](https://mlsanity.com/paper/2312.02481/) (2023)
- [EarthSight: A Distributed Framework for Low-Latency Satellite Intelligence](https://mlsanity.com/paper/2511.10834/) (2025)

## 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%.