Evaluation on a representative multimodal benchmark demonstrates that EMMI can reduce the communication payload by 32x while maintaining comparable downstream accuracy, resulting in up to a 3.4x reduction in estimated end-to-end inference latency under bandwidth-constrained edge conditions.
Abstract
Recent advances in multimodal large language mod- els (MLLMs) have opened new opportunities for edge intelligence by enabling reasoning across heterogeneous sensor modalities, such as vision, text, and telemetry data. However, deploying these capabilities on resource-constrained edge platforms remains challenging due to the substantial computational, memory, and communication demands of modern MLLMs. Rather than transmitting raw sensor observations or partitioning neural networks at intermediate layers, Edge Multi-Modal Intelligence (EMMI) communicates a compact representation between edge devices and server resources, enabling communication-efficient edge MLLM inference. To achieve this, EMMI performs modality-specific encoding, cross-modal representation fusion, and learned compression at the edge, transmitting only a compact latent representation to server-side resources for high-capacity MLLM reasoning. This representation-centric design reduces communication overhead, preserves local data privacy, and provides a fixed-size interface between heterogeneous edge devices and server-side MLLMs. Evaluation on a representative multimodal benchmark demonstrates that EMMI can reduce the communication payload by 32x while maintaining comparable downstream accuracy, resulting in up to a 3.4x reduction in estimated end-to-end inference latency under bandwidth-constrained edge conditions.
This review examines architectures, compression techniques, privacy mechanisms, and energy-efficient inference strategies for multimodal small language models operating on smartphones, wearables, vehicles, robots, cameras, and embedded Internet of Things platforms, and develops a deploymentcentric taxonomy.
Stanley Nwakamma· WORLD JOURNAL OF INNOVATION...· 1 citation
By jointly optimizing distillation and quantization within a unified teacher-anchored framework, this method ensures consistent training under quantization, suppressing hard negatives and enlarging decision margins, and designs a lightweight cross-attention adapter that enhances non-RGB representations through RGB-guid...
Jinwoo Jeon, GyuYeop Do, Yunkyu Lim et al.· 0 citations
Collaborative inference between the edge cloud and user equipment (UE) is a promising paradigm for deploying large deep neural networks (DNNs) in 6G networks. However, existing cloud-edge inference and distillation schemes often require the real-time transmission of high-dimensional intermediate features or soft probab...
Wan-Ling Luo, Jia-Yi Zhang, Bo-Kai Xu et al.· 0 citations
The rapid proliferation of Internet of Things (IoT) devices is accelerating the development of distributed machine learning paradigms, such as split learning (SL). However, as the number of participating devices increases, SL faces severe communication challenges due to the high-dimensional smashed data transmitted ove...
Han-Yu Li, Miao Yang, Liang Zhong et al.· 2026 6th International Confe...· 0 citations
This work proposes MoR-MLLM, a computation-sparse MLLM based on the recent Mixture-of-Recursions (MoR) framework, which introduces adaptive per-token recursion, allowing the model to dynamically adjust its recursive depth and allocate more computation to visually or linguistically challenging tokens while skipping redu...
Peng-Cheng Zheng, Chao-Ning Zhang, Jia-Xing Yan et al.· IEEE transactions on multime...· 1 citation
Collaborative Selective Transmission is presented, a main-directed query--response framework that retrieves only helper information complementary to the current main representation that learns sample-adaptive, helper-specific sparse retrieval supports while discouraging retrieval of semantics already covered by the mai...
Haixu Chi, Junrui Zhang, Rui Ning et al.· arXiv.org· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 17, 2026
By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.
Microsoft Research Blog· microsoft.comNov 16, 2026
AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduNov 13, 2026