analiza aiqbench
This research introduces a two-stage inference framework for vision-language models that dynamically requests high-resolution details only when initial thumbnail analysis indicates uncertainty, significantly reducing bandwidth usage without sacrificing accuracy. By leveraging the model's internal attention mechanisms to identify regions of interest, the approach effectively balances communication efficiency with performance in edge-to-server deployments. The method offers a practical solution for optimizing VLM operations in bandwidth-constrained environments where transmitting full-resolution images is prohibitively expensive.
Collaborative Edge-to-Server Inference for Vision-Language Models
arXiv:2512.16349v2 Announce Type: replace-cross Abstract: We propose a collaborative edge-to-server inference framework for vision-language models (VLMs) that reduces communication cost while maintaining inference accuracy. In typical deployments, visual data captured at edge devices (clients) is transmitted to the server for VLM inference. However, transmitting full-resolution images incurs high communication cost. Conversely, aggressive downsizing or excessive compression to mitigate communication overhead can discard fine-grained details, leading to accuracy degradation. To overcome this limitation, we design a communication-efficient two-stage framework. In the first stage, the server performs inference on the downsized thumbnail (global image) and quantifies the min-entropy of the output tokens. If the min-entropy exceeds a predefined threshold, the server identifies a region of interest (RoI) using the VLM's internal attention and requests the edge device to send a detail-preserved local image of the RoI. The server then refines its inference by jointly leveraging the global and local images. This selective retransmission strategy ensures that only essential visual content is additionally transmitted. Experimental results consistently confirm that the proposed framework substantially reduces communication overhead while maintaining inference accuracy across diverse VQA benchmarks.
Abstract:We propose a collaborative edge-to-server inference framework for vision-language models (VLMs) that reduces communication cost while maintaining inference accuracy. In typical deployments, visual data captured at edge devices (clients) is transmitted to the server for VLM inference. However, transmitting full-resolution images incurs high communication cost. Conversely, aggressive downsizing or excessive compression to mitigate communication overhead can discard fine-grained details, leading to accuracy degradation. To overcome this limitation, we design a communication-efficient two-stage framework. In the first stage, the server performs inference on the downsized thumbnail (global image) and quantifies the min-entropy of the output tokens. If the min-entropy exceeds a predefined threshold, the server identifies a region of interest (RoI) using the VLM's internal attention and requests the edge device to send a detail-preserved local image of the RoI. The server then refines its inference by jointly leveraging the global and local images. This selective retransmission strategy ensures that only essential visual content is additionally transmitted. Experimental results consistently confirm that the proposed framework substantially reduces communication overhead while maintaining inference accuracy across diverse VQA benchmarks.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
arXiv:2512.16349 [cs.CV]
(or arXiv:2512.16349v2 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2512.16349 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 274 cuvinte