The paper introduces Rift, a two‑stage system that reduces the computational load of vision‑language models on satellites by pruning image tiles that do not affect the answer and then applying elastic prefill to limit token usage. By exploiting answer‑invariant token redundancy, Rift cuts energy consumption by 78 % and latency by 69 % compared to exhaustive tiled inference, while boosting accuracy from 45 % to 73 % on LLaVA‑1.5 7B running on a Jetson AGX Orin.
arXiv:2606. 11576v1 Announce Type: cross Abstract: Modern Vision-Language Models (VLMs) benefit from chain-of-thought prompting and test-time scaling, but these gains often come with prohibitive inference cost due to large visual contexts and long decoding chains.
arXiv:2606. 12412v1 Announce Type: cross Abstract: Vision-language models (VLMs) project images into hundreds to thousands of visual tokens, making decoder inference expensive in both attention computation and KV-cache memory.
Multimodal foundation models are reshaping edge-cloud visual intelligence from task-specific feature pipelines into token-based interfaces, where edge devices encode visual inputs into tokens for a general-purpose cloud MLLM. However, dense visual-token sequences increase cloud-side inference costs.
arXiv:2609.16841v1 Announce Type: cross Abstract: Increasing image resolution produces ever-longer visual-token sequences in vision-language models (VLMs), substantially raising their inference cost....
PACE introduces a training‑free Condense‑and‑Extract framework that speeds up Vision‑Language Model inference by first adaptively downsampling visual inputs before encoding and then selectively retaining essential tokens during decoding. The Adaptive Pixel Compressor (APC) reduces encoder workload while preserving global context, and the Dynamic Dual‑Attention Extractor (DDAE) keeps task‑critical details by fusing visual and language signals. Applied to Qwen2.5‑VL‑7B, PACE maintains 93.8% of performance using only 10% of visual tokens, achieving a 3.1× speedup in time to first token.
The paper introduces Adaptive Visual Token Pruning (AVTP), a training‑free framework that dynamically selects pruning layers and ratios for large vision‑language models (LVLMs) when processing multiple image sequences. By analyzing visual attention distributions across different LVLM architectures, AVTP adapts token retention to image importance, enabling efficient inference without relying on attention‑based computations incompatible with FlashAttention. Experiments show significant speedups—up to 2× for Qwen3VL‑8B—while preserving or even improving accuracy on multi‑image benchmarks.
arXiv:2607. 09520v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) are the perceptual backbone of embodied AI, but their energy footprint on edge hardware remains poorly understood.
GeoNLI introduces a unified, modular pipeline that combines advanced SAM variants with multimodal large language models to perform satellite image captioning, visual question answering (VQA), and visual grounding. The EarthMind model achieves strong results on captioning and VQA, while multiple RemoteSAM-SAM and DiffuSAM pipelines are used for grounding, ultimately employing a majority‑voting ensemble across several models. The system reports 82% captioning accuracy, 83.32% VQA accuracy, and 64.94% grounding accuracy, demonstrating improved consistency over task‑specific approaches.
The paper shows that only a small subset of attention heads in vision-language models is responsible for selecting critical visual tokens. By pruning tokens based on similarity before LLM reasoning and then applying head‑aware pruning during reasoning, the proposed ProViP framework achieves high task performance with significant speedups. Experiments on LLaVA‑1.5‑7B demonstrate 95.9% performance retention and a 1.62× inference speedup at an 88.9% pruning ratio.
arXiv:2607. 15942v1 Announce Type: cross Abstract: Remote sensing vision-language models are increasingly expected to support open-ended reasoning over Earth Observation data and a variety of tasks.
arXiv:2609.15131v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) require substantial computation to process numerous visual tokens across all transformer layers. Most method...
The paper introduces RS-OPSD, a reliable privileged on-policy self-distillation framework designed for ultra‑high‑resolution remote sensing visual question answering. It leverages a new dataset, GeoEvidence‑6K, and a human‑feedback guided skill refinement process to provide explicit question‑relevant evidence. By incorporating context‑preserving visual privilege and correctness‑aligned distillation, RS‑OPSD achieves state‑of‑the‑art performance on several benchmarks without requiring additional visual search or tool calls at inference time.