Shangguang Wang

Rongjie Yi, Liwei Guo, Shiyun Wei et al.

Large language models (LLMs) such as GPTs and Mixtral-8x7B have revolutionized machine intelligence due to their exceptional abilities in generic ML tasks. Transiting LLMs from datacenters to edge devices brings benefits like better privacy and availability, but is challenged by their massive parameter size and thus unbearable runtime costs. To this end, we present EdgeMoE, an on-device inference engine for mixture-of-expert (MoE) LLMs -- a popular form of sparse LLM that scales its parameter size with almost constant computing complexity. EdgeMoE achieves both memory- and compute-efficiency by partitioning the model into the storage hierarchy: non-expert weights are held in device memory; while expert weights are held on external storage and fetched to memory only when activated. This design is motivated by a key observation that expert weights are bulky but infrequently used due to sparse activation. To further reduce the expert I/O swapping overhead, EdgeMoE incorporates two novel techniques: (1) expert-wise bitwidth adaptation that reduces the expert sizes with tolerable accuracy loss; (2) expert preloading that predicts the activated experts ahead of time and preloads it with the compute-I/O pipeline. On popular MoE LLMs and edge devices, EdgeMoE showcase significant memory savings and speedup over competitive baselines. The code is available at https://github.com/UbiquitousLearning/mllm.

Chuntao Ding, Zhichao Lu, Shangguang Wang et al.

Multi-task learning (MTL) seeks to learn a single model to accomplish multiple tasks by leveraging shared information among the tasks. Existing MTL models, however, have been known to suffer from negative interference among tasks. Efforts to mitigate task interference have focused on either loss/gradient balancing or implicit parameter partitioning with partial overlaps among the tasks. In this paper, we propose ETR-NLP to mitigate task interference through a synergistic combination of non-learnable primitives (NLPs) and explicit task routing (ETR). Our key idea is to employ non-learnable primitives to extract a diverse set of task-agnostic features and recombine them into a shared branch common to all tasks and explicit task-specific branches reserved for each task. The non-learnable primitives and the explicit decoupling of learnable parameters into shared and task-specific ones afford the flexibility needed for minimizing task interference. We evaluate the efficacy of ETR-NLP networks for both image-level classification and pixel-level dense prediction MTL problems. Experimental results indicate that ETR-NLP significantly outperforms state-of-the-art baselines with fewer learnable parameters and similar FLOPs across all datasets. Code is available at this \href{https://github.com/zhichao-lu/etr-nlp-mtl}.

Jinliang Yuan, Chen Yang, Dongqi Cai et al. · cambridge

In today's landscape, smartphones have evolved into hubs for hosting a multitude of deep learning models aimed at local execution. A key realization driving this work is the notable fragmentation among these models, characterized by varied architectures, operators, and implementations. This fragmentation imposes a significant burden on the comprehensive optimization of hardware, system settings, and algorithms. Buoyed by the recent strides in large foundation models, this work introduces a pioneering paradigm for mobile AI: a collaborative management approach between the mobile OS and hardware, overseeing a foundational model capable of serving a broad spectrum of mobile AI tasks, if not all. This foundational model resides within the NPU and remains impervious to app or OS revisions, akin to firmware. Concurrently, each app contributes a concise, offline fine-tuned "adapter" tailored to distinct downstream tasks. From this concept emerges a concrete instantiation known as \sys. It amalgamates a curated selection of publicly available Large Language Models (LLMs) and facilitates dynamic data flow. This concept's viability is substantiated through the creation of an exhaustive benchmark encompassing 38 mobile AI tasks spanning 50 datasets, including domains such as Computer Vision (CV), Natural Language Processing (NLP), audio, sensing, and multimodal inputs. Spanning this benchmark, \sys unveils its impressive performance. It attains accuracy parity in 85\% of tasks, demonstrates improved scalability in terms of storage and memory, and offers satisfactory inference speed on Commercial Off-The-Shelf (COTS) mobile devices fortified with NPU support. This stands in stark contrast to task-specific models tailored for individual applications.

Dongqi Cai, Yaozong Wu, Shangguang Wang et al. · cambridge

Transformer-based pre-trained models have revolutionized NLP for superior performance and generality. Fine-tuning pre-trained models for downstream tasks often requires private data, for which federated learning is the de-facto approach (i.e., FedNLP). However, our measurements show that FedNLP is prohibitively slow due to the large model sizes and the resultant high network/computation cost. Towards practical FedNLP, we identify as the key building blocks adapters, small bottleneck modules inserted at a variety of model layers. A key challenge is to properly configure the depth and width of adapters, to which the training speed and efficiency is highly sensitive. No silver-bullet configuration exists: the optimal choice varies across downstream NLP tasks, desired model accuracy, and mobile resources. To automate adapter configuration, we propose FedAdapter, a framework that enhances the existing FedNLP with two novel designs. First, FedAdapter progressively upgrades the adapter configuration throughout a training session; the principle is to quickly learn shallow knowledge by only training fewer and smaller adapters at the model's top layers, and incrementally learn deep knowledge by incorporating deeper and larger adapters. Second, FedAdapter continuously profiles future adapter configurations by allocating participant devices to trial groups. Extensive experiments show that FedAdapter can reduce FedNLP's model convergence delay to no more than several hours, which is up to 155.5$\times$ faster compared to vanilla FedNLP and 48$\times$ faster compared to strong baselines.

Mengwei Xu, Dongqi Cai, Yaozong Wu et al. · cambridge

Large Language Models (LLMs) are transforming the landscape of mobile intelligence. Federated Learning (FL), a method to preserve user data privacy, is often employed in fine-tuning LLMs to downstream mobile tasks, an approach known as FedLLM. Though recent efforts have addressed the network issue induced by the vast model size, they have not practically mitigated vital challenges concerning integration with mobile devices, such as significant memory consumption and sluggish model convergence. In response to these challenges, this work introduces FwdLLM, an innovative FL protocol designed to enhance the FedLLM efficiency. The key idea of FwdLLM to employ backpropagation (BP)-free training methods, requiring devices only to execute ``perturbed inferences''. Consequently, FwdLLM delivers way better memory efficiency and time efficiency (expedited by mobile NPUs and an expanded array of participant devices). FwdLLM centers around three key designs: (1) it combines BP-free training with parameter-efficient training methods, an essential way to scale the approach to the LLM era; (2) it systematically and adaptively allocates computational loads across devices, striking a careful balance between convergence speed and accuracy; (3) it discriminatively samples perturbed predictions that are more valuable to model convergence. Comprehensive experiments with five LLMs and three NLP tasks illustrate FwdLLM's significant advantages over conventional methods, including up to three orders of magnitude faster convergence and a 14.6x reduction in memory footprint. Uniquely, FwdLLM paves the way for federated learning of billion-parameter LLMs such as LLaMA on COTS mobile devices -- a feat previously unattained.

Hanzi Mei, Dongqi Cai, Ao Zhou et al. · cambridge

As Large Language Models (LLMs) push the boundaries of AI capabilities, their demand for data is growing. Much of this data is private and distributed across edge devices, making Federated Learning (FL) a de-facto alternative for fine-tuning (i.e., FedLLM). However, it faces significant challenges due to the inherent heterogeneity among clients, including varying data distributions and diverse task types. Towards a versatile FedLLM, we replace traditional dense model with a sparsely-activated Mixture-of-Experts (MoE) architecture, whose parallel feed-forward networks enable greater flexibility. To make it more practical in resource-constrained environments, we present FedMoE, the efficient personalized FL framework to address data heterogeneity, constructing an optimal sub-MoE for each client and bringing the knowledge back to global MoE. FedMoE is composed of two fine-tuning stages. In the first stage, FedMoE simplifies the problem by conducting a heuristic search based on observed activation patterns, which identifies a suboptimal submodel for each client. In the second stage, these submodels are distributed to clients for further training and returned for server aggregating through a novel modular aggregation strategy. Meanwhile, FedMoE progressively adjusts the submodels to optimal through global expert recommendation. Experimental results demonstrate the superiority of our method over previous personalized FL methods.

Dongqi Cai, Shangguang Wang, Yaozong Wu et al. · cambridge

Natural language processing (NLP) sees rich mobile applications. To support various language understanding tasks, a foundation NLP model is often fine-tuned in a federated, privacy-preserving setting (FL). This process currently relies on at least hundreds of thousands of labeled training samples from mobile clients; yet mobile users often lack willingness or knowledge to label their data. Such an inadequacy of data labels is known as a few-shot scenario; it becomes the key blocker for mobile NLP applications. For the first time, this work investigates federated NLP in the few-shot scenario (FedFSL). By retrofitting algorithmic advances of pseudo labeling and prompt learning, we first establish a training pipeline that delivers competitive accuracy when only 0.05% (fewer than 100) of the training data is labeled and the remaining is unlabeled. To instantiate the workflow, we further present a system FeS, addressing the high execution cost with novel designs. (1) Curriculum pacing, which injects pseudo labels to the training workflow at a rate commensurate to the learning progress; (2) Representational diversity, a mechanism for selecting the most learnable data, only for which pseudo labels will be generated; (3) Co-planning of a model's training depth and layer capacity. Together, these designs reduce the training delay, client energy, and network traffic by up to 46.0$\times$, 41.2$\times$ and 3000.0$\times$, respectively. Through algorithm/system co-design, FFNLP demonstrates that FL can apply to challenging settings where most training samples are unlabeled.

Dongqi Cai, Yaozong Wu, Haitao Yuan et al. · cambridge

Transformer-based pre-trained models have emerged as the predominant solution for natural language processing (NLP). Fine-tuning such pre-trained models for downstream tasks often requires a considerable amount of labeled private data. In practice, private data is often distributed across heterogeneous mobile devices and may be prohibited from being uploaded. Moreover, well-curated labeled data is often scarce, presenting an additional challenge. To address these challenges, we first introduce a data generator for federated few-shot learning tasks, which encompasses the quantity and skewness of scarce labeled data in a realistic setting. Subsequently, we propose AUG-FedPrompt, a prompt-based federated learning system that exploits abundant unlabeled data for data augmentation. Our experiments indicate that AUG-FedPrompt can perform on par with full-set fine-tuning with a limited amount of labeled data. However, such competitive performance comes at a significant system cost.

Zhichao Lu, Chuntao Ding, Felix Juefei-Xu et al.

Deploying high-performance vision transformer (ViT) models on ubiquitous Internet of Things (IoT) devices to provide high-quality vision services will revolutionize the way we live, work, and interact with the world. Due to the contradiction between the limited resources of IoT devices and resource-intensive ViT models, the use of cloud servers to assist ViT model training has become mainstream. However, due to the larger number of parameters and floating-point operations (FLOPs) of the existing ViT models, the model parameters transmitted by cloud servers are large and difficult to run on resource-constrained IoT devices. To this end, this paper proposes a transmission-friendly ViT model, TFormer, for deployment on resource-constrained IoT devices with the assistance of a cloud server. The high performance and small number of model parameters and FLOPs of TFormer are attributed to the proposed hybrid layer and the proposed partially connected feed-forward network (PCS-FFN). The hybrid layer consists of nonlearnable modules and a pointwise convolution, which can obtain multitype and multiscale features with only a few parameters and FLOPs to improve the TFormer performance. The PCS-FFN adopts group convolution to reduce the number of parameters. The key idea of this paper is to propose TFormer with few model parameters and FLOPs to facilitate applications running on resource-constrained IoT devices to benefit from the high performance of the ViT models. Experimental results on the ImageNet-1K, MS COCO, and ADE20K datasets for image classification, object detection, and semantic segmentation tasks demonstrate that the proposed model outperforms other state-of-the-art models. Specifically, TFormer-S achieves 5% higher accuracy on ImageNet-1K than ResNet18 with 1.4$\times$ fewer parameters and FLOPs.

Dongqi Cai, Shangguang Wang, Chen Peng et al. · cambridge

Human memory is inherently prone to forgetting. To address this, multimodal embedding models have been introduced, which transform diverse real-world data into a unified embedding space. These embeddings can be retrieved efficiently, aiding mobile users in recalling past information. However, as model complexity grows, so do its resource demands, leading to reduced throughput and heavy computational requirements that limit mobile device implementation. In this paper, we introduce RECALL, a novel on-device multimodal embedding system optimized for resource-limited mobile environments. RECALL achieves high-throughput, accurate retrieval by generating coarse-grained embeddings and leveraging query-based filtering for refined retrieval. Experimental results demonstrate that RECALL delivers high-quality embeddings with superior throughput, all while operating unobtrusively with minimal memory and energy consumption.

Rongjie Yi, Ting Cao, Ao Zhou et al.

DNNs are ubiquitous on edge devices nowadays. With its increasing importance and use cases, it's not likely to pack all DNNs into device memory and expect that each inference has been warmed up. Therefore, cold inference, the process to read, initialize, and execute a DNN model, is becoming commonplace and its performance is urgently demanded to be optimized. To this end, we present NNV12, the first on-device inference engine that optimizes for cold inference NNV12 is built atop 3 novel optimization knobs: selecting a proper kernel (implementation) for each DNN operator, bypassing the weights transformation process by caching the post-transformed weights on disk, and pipelined execution of many kernels on asymmetric processors. To tackle with the huge search space, NNV12 employs a heuristic-based scheme to obtain a near-optimal kernel scheduling plan. We fully implement a prototype of NNV12 and evaluate its performance across extensive experiments. It shows that NNV12 achieves up to 15.2x and 401.5x compared to the state-of-the-art DNN engines on edge CPUs and GPUs, respectively.

Zhichao Lu, Chuntao Ding, Shangguang Wang et al.

Deploying high-performance convolutional neural network (CNN) models on low-earth orbit (LEO) satellites for rapid remote sensing image processing has attracted significant interest from industry and academia. However, the limited resources available on LEO satellites contrast with the demands of resource-intensive CNN models, necessitating the adoption of ground-station server assistance for training and updating these models. Existing approaches often require large floating-point operations (FLOPs) and substantial model parameter transmissions, presenting considerable challenges. To address these issues, this paper introduces a ground-station server-assisted framework. With the proposed framework, each layer of the CNN model contains only one learnable feature map (called the seed feature map) from which other feature maps are generated based on specific rules. The hyperparameters of these rules are randomly generated instead of being trained, thus enabling the generation of multiple feature maps from the seed feature map and significantly reducing FLOPs. Furthermore, since the random hyperparameters can be saved using a few random seeds, the ground station server assistance can be facilitated in updating the CNN model deployed on the LEO satellite. Experimental results on the ISPRS Vaihingen, ISPRS Potsdam, UAVid, and LoveDA datasets for semantic segmentation services demonstrate that the proposed framework outperforms existing state-of-the-art approaches. In particular, the SineFM-based model achieves a higher mIoU than the UNetFormer on the UAVid dataset, with 3.3x fewer parameters and 2.2x fewer FLOPs.

Tao Qi, Huili Wang, Yuanhong Huang et al.

The rapid advancement of diffusion-based image generation models has raised serious concerns regarding potential copyright and privacy infringements involving human-created data. Membership inference attacks (MIAs) have emerged as a promising tool for identifying unauthorized data usage during model training. Existing methods typically assess the ability of model to denoise perturbed suspect images as an indicator of membership status. However, the discriminative power of such features is highly dependent on the degree of model memorization and deteriorates significantly when applied to less exposed data (e.g., pre-training data). Although several methods attempt to enhance detection by leveraging internal model features, these features are generally inaccessible in mainstream closed-source image generation platforms, limiting their practicality. In this paper, we demonstrate that analyzing how a black-box diffusion model denoises a target image and corresponding perturbed textual instructions can reveal more distinctive membership cues. Based on this insight, we propose a black-box membership inference attack framework (named SD-MIA) that leverages a cross-modal data perturbation mechanism to detect pre-training data in diffusion models. We conduct extensive experiments on both a public benchmark dataset and a newly constructed dataset, each comprising pre-training membership and non-membership samples with identical distributions. Experimental results demonstrate that SD-MIA achieves superior performance compared to existing baselines, including those with the unfair advantage of accessing internal model features.

Peiru Yang, Haoran Zheng, Tong Ju et al.

Retrieval-augmented generation (RAG) is a widely adopted paradigm for enhancing LLMs in medical applications by incorporating expert multimodal knowledge during generation. However, the underlying retrieval databases may naturally contain, or be intentionally injected with, adversarial knowledge, which can perturb model outputs and undermine system reliability. To investigate this risk, prior studies have explored knowledge poisoning attacks in medical RAG systems. Nevertheless, most of them rely on the strong assumption that adversaries possess prior knowledge of user queries, which is unrealistic in deployments and substantially limits their practical applicability. In this paper, we propose M\textsuperscript{3}Att, a knowledge-poisoning framework designed for medical multimodal RAG systems, assuming only limited distribution knowledge of the underlying database. Our core idea is to inject covert misinformation into textual data while using paired visual data as a query-agnostic trigger to promote retrieval. We first propose a unified framework that introduces imperceptible perturbations to visual inputs to manipulate retrieval probabilities. Besides, due to the prior medical knowledge in LLMs, naively poisoned medical content with explicit factual errors can be corrected during generation. Thus, we leverage the inherent ambiguity of medical diagnosis and design a covert misinformation injection strategy that degrades diagnostic accuracy while evading model self-correction. Experiments on five LLMs and datasets demonstrate that M\textsuperscript{3}Att consistently produces clinically plausible yet incorrect generations. Codes: https://github.com/ypr17/M3Att.

Minxiao Chen, Haitao Yuan, Nan Jiang et al.

Traffic accidents pose a significant risk to human health and property safety. Therefore, to prevent traffic accidents, predicting their risks has garnered growing interest. We argue that a desired prediction solution should demonstrate resilience to the complexity of traffic accidents. In particular, it should adequately consider the regional background, accurately capture both spatial proximity and semantic similarity, and effectively address the sparsity of traffic accidents. However, these factors are often overlooked or difficult to incorporate. In this paper, we propose a novel multi-granularity hierarchical spatio-temporal network. Initially, we innovate by incorporating remote sensing data, facilitating the creation of hierarchical multi-granularity structure and the comprehension of regional background. We construct multiple high-level risk prediction tasks to enhance model's ability to cope with sparsity. Subsequently, to capture both spatial proximity and semantic similarity, region feature and multi-view graph undergo encoding processes to distill effective representations. Additionally, we propose message passing and adaptive temporal attention module that bridges different granularities and dynamically captures time correlations inherent in traffic accident patterns. At last, a multivariate hierarchical loss function is devised considering the complexity of the prediction purpose. Extensive experiments on two real datasets verify the superiority of our model against the state-of-the-art methods.

Jinhua Yin, Peiru Yang, Chen Yang et al.

Large vision-language models (LVLMs) derive their capabilities from extensive training on vast corpora of visual and textual data. Empowered by large-scale parameters, these models often exhibit strong memorization of their training data, rendering them susceptible to membership inference attacks (MIAs). Existing MIA methods for LVLMs typically operate under white- or gray-box assumptions, by extracting likelihood-based features for the suspected data samples based on the target LVLMs. However, mainstream LVLMs generally only expose generated outputs while concealing internal computational features during inference, limiting the applicability of these methods. In this work, we propose the first black-box MIA framework for LVLMs, based on a prior knowledge-calibrated memory probing mechanism. The core idea is to assess the model memorization of the private semantic information embedded within the suspected image data, which is unlikely to be inferred from general world knowledge alone. We conducted extensive experiments across four LVLMs and three datasets. Empirical results demonstrate that our method effectively identifies training data of LVLMs in a purely black-box setting and even achieves performance comparable to gray-box and white-box methods. Further analysis reveals the robustness of our method against potential adversarial manipulations, and the effectiveness of the methodology designs. Our code and data are available at https://github.com/spmede/KCMP.

Yunhe Yan, Shihe Wang, Jiajun Du et al.

(M)LLM-powered computer use agents (CUA) are emerging as a transformative technique to automate human-computer interaction. However, existing CUA benchmarks predominantly target GUI agents, whose evaluation methods are susceptible to UI changes and ignore function interactions exposed by application APIs, e.g., Model Context Protocol (MCP). To this end, we propose MCPWorld, the first automatic CUA testbed for API, GUI, and API-GUI hybrid agents. A key principle of MCPWorld is the use of "white-box apps", i.e., those with source code availability and can be revised/re-compiled as needed (e.g., adding MCP support), with two notable advantages: (1) It greatly broadens the design space of CUA, such as what and how the app features to be exposed/extracted as CUA-callable APIs. (2) It allows MCPWorld to programmatically verify task completion by directly monitoring application behavior through techniques like dynamic code instrumentation, offering robust, accurate CUA evaluation decoupled from specific agent implementations or UI states. Currently, MCPWorld includes 201 well curated and annotated user tasks, covering diversified use cases and difficulty levels. MCPWorld is also fully containerized with GPU acceleration support for flexible adoption on different OS/hardware environments. Our preliminary experiments, using a representative LLM-powered CUA framework, achieve 75.12% task completion accuracy, simultaneously providing initial evidence on the practical effectiveness of agent automation leveraging MCP. Overall, we anticipate MCPWorld to facilitate and standardize the benchmarking of next-generation computer use agents that can leverage rich external tools. Our code and dataset are publicly available at https://github.com/SAAgent/MCPWorld.

Duo Chai, Zizhen Liu, Shuhuai Wang et al.

Large language models (LLMs) are highly compute- and memory-intensive, posing significant demands on high-performance GPUs. At the same time, advances in GPU technology driven by shrinking transistor sizes and lower operating voltages have made these devices increasingly susceptible to soft errors. While prior work has examined GPU reliability, most studies have focused on general-purpose applications or conventional neural networks mostly used for vision tasks such as classification and detection. In contrast, systematic analysis of modern large-scale LLMs remains limited, despite their rapid adoption in diverse application scenarios. Given the unique characteristics of LLMs, their resilience to soft errors may differ substantially from earlier models. To bridge this gap, we conduct the first instruction-level fault injection study of LLM inference. Our approach reveals reliability characteristics from multiple perspectives, highlighting the effects of model architecture, parameter scale, and task complexity. These findings provide new insights into LLM reliability and inform the design of more effective fault tolerance mechanisms.

Rongjie Yi, Xiang Li, Weikai Xie et al.

The interest in developing small language models (SLM) for on-device deployment is fast growing. However, the existing SLM design hardly considers the device hardware characteristics. Instead, this work presents a simple yet effective principle for SLM design: architecture searching for (near-)optimal runtime efficiency before pre-training. Guided by this principle, we develop PhoneLM SLM family (currently with 0.5B and 1.5B versions), that acheive the state-of-the-art capability-efficiency tradeoff among those with similar parameter size. We fully open-source the code, weights, and training datasets of PhoneLM for reproducibility and transparency, including both base and instructed versions. We also release a finetuned version of PhoneLM capable of accurate Android Intent invocation, and an end-to-end Android demo. All materials are available at https://github.com/UbiquitousLearning/PhoneLM.

Alireza Furutanpey, Qiyang Zhang, Philipp Raith et al.

Nanosatellite constellations equipped with sensors capturing large geographic regions provide unprecedented opportunities for Earth observation. As constellation sizes increase, network contention poses a downlink bottleneck. Orbital Edge Computing (OEC) leverages limited onboard compute resources to reduce transfer costs by processing the raw captures at the source. However, current solutions have limited practicability due to reliance on crude filtering methods or over-prioritizing particular downstream tasks. This work presents FOOL, an OEC-native and task-agnostic feature compression method that preserves prediction performance. FOOL partitions high-resolution satellite imagery to maximize throughput. Further, it embeds context and leverages inter-tile dependencies to lower transfer costs with negligible overhead. While FOOL is a feature compressor, it can recover images with competitive scores on quality measures at lower bitrates. We extensively evaluate transfer cost reduction by including the peculiarity of intermittently available network connections in low earth orbit. Lastly, we test the feasibility of our system for standardized nanosatellite form factors. We demonstrate that FOOL permits downlinking over 100x the data volume without relying on prior information on the downstream tasks.

Zeling Zhang, Dongqi Cai, Yiran Zhang et al. · cambridge

Communication overhead is a significant bottleneck in federated learning (FL), which has been exaggerated with the increasing size of AI models. In this paper, we propose FedRDMA, a communication-efficient cross-silo FL system that integrates RDMA into the FL communication protocol. To overcome the limitations of RDMA in wide-area networks (WANs), FedRDMA divides the updated model into chunks and designs a series of optimization techniques to improve the efficiency and robustness of RDMA-based communication. We implement FedRDMA atop the industrial federated learning framework and evaluate it on a real-world cross-silo FL scenario. The experimental results show that \sys can achieve up to 3.8$\times$ speedup in communication efficiency compared to traditional TCP/IP-based FL systems.

Nan Jiang, Haitao Yuan, Jianing Si et al.

The next point-of-interest (POI) prediction is a significant task in location-based services, yet its complexity arises from the consolidation of spatial and semantic intent. This fusion is subject to the influences of historical preferences, prevailing location, and environmental factors, thereby posing significant challenges. In addition, the uneven POI distribution further complicates the next POI prediction procedure. To address these challenges, we enrich input features and propose an effective deep-learning method within a two-step prediction framework. Our method first incorporates remote sensing data, capturing pivotal environmental context to enhance input features regarding both location and semantics. Subsequently, we employ a region quad-tree structure to integrate urban remote sensing, road network, and POI distribution spaces, aiming to devise a more coherent graph representation method for urban spatial. Leveraging this method, we construct the QR-P graph for the user's historical trajectories to encapsulate historical travel knowledge, thereby augmenting input features with comprehensive spatial and semantic insights. We devise distinct embedding modules to encode these features and employ an attention mechanism to fuse diverse encodings. In the two-step prediction procedure, we initially identify potential spatial zones by predicting user-preferred tiles, followed by pinpointing specific POIs of a designated type within the projected tiles. Empirical findings from four real-world location-based social network datasets underscore the remarkable superiority of our proposed approach over competitive baseline methods.

Minxiao Chen, Haitao Yuan, Nan Jiang et al.

Online map matching is a fundamental problem in location-based services, aiming to incrementally match trajectory data step-by-step onto a road network. However, existing methods fail to meet the needs for efficiency, robustness, and accuracy required by large-scale online applications, making this task still challenging. This paper introduces a novel framework that achieves high accuracy and efficient matching while ensuring robustness in handling diverse scenarios. To improve efficiency, we begin by modeling the online map matching problem as an Online Markov Decision Process (OMDP) based on its inherent characteristics. This approach helps efficiently merge historical and real-time data, reducing unnecessary calculations. Next, to enhance robustness, we design a reinforcement learning method, enabling robust handling of real-time data from dynamically changing environments. In particular, we propose a novel model learning process and a comprehensive reward function, allowing the model to make reasonable current matches from a future-oriented perspective, and to continuously update and optimize during the decision-making process based on feedback. Lastly, to address the heterogeneity between trajectories and roads, we design distinct graph structures, facilitating efficient representation learning through graph and recurrent neural networks. To further align trajectory and road data, we introduce contrastive learning to decrease their distance in the latent space, thereby promoting effective integration of the two. Extensive evaluations on three real-world datasets confirm that our method significantly outperforms existing state-of-the-art solutions in terms of accuracy, efficiency and robustness.

Qian Li, Cheng Ji, Shu Guo et al.

Multi-modal relation extraction (MMRE) is a challenging task that aims to identify relations between entities in text leveraging image information. Existing methods are limited by their neglect of the multiple entity pairs in one sentence sharing very similar contextual information (ie, the same text and image), resulting in increased difficulty in the MMRE task. To address this limitation, we propose the Variational Multi-Modal Hypergraph Attention Network (VM-HAN) for multi-modal relation extraction. Specifically, we first construct a multi-modal hypergraph for each sentence with the corresponding image, to establish different high-order intra-/inter-modal correlations for different entity pairs in each sentence. We further design the Variational Hypergraph Attention Networks (V-HAN) to obtain representational diversity among different entity pairs using Gaussian distribution and learn a better hypergraph structure via variational attention. VM-HAN achieves state-of-the-art performance on the multi-modal relation extraction task, outperforming existing methods in terms of accuracy and efficiency.

Yanan Li, Fanxu Meng, Muhan Zhang et al.

As Large Language Models (LLMs) are frequently updated, LoRA weights trained on earlier versions quickly become obsolete. The conventional practice of retraining LoRA weights from scratch on the latest model is costly, time-consuming, and environmentally detrimental, particularly as the diversity of LLMs and downstream tasks expands. This motivates a critical question: "How can we efficiently leverage existing LoRA weights to adapt to newer model versions?" To address this, we propose LoRASuite, a modular approach tailored specifically to various types of LLM updates. First, we compute a transfer matrix utilizing known parameters from both old and new LLMs. Next, we allocate corresponding layers and attention heads based on centered kernel alignment and cosine similarity metrics, respectively. A subsequent small-scale, skillful fine-tuning step ensures numerical stability. Experimental evaluations demonstrate that LoRASuite consistently surpasses small-scale vanilla LoRA methods. Notably, on backbone LLMs such as MiniCPM and Qwen, LoRASuite even exceeds the performance of full-scale LoRA retraining, with average improvements of +1.4 and +6.6 points on math tasks, respectively. Additionally, LoRASuite significantly reduces memory consumption by 5.5 GB and computational time by 78.23%.

Peiru Yang, Xintian Li, Zhiyang Hu et al.

Retrieval-augmented generation (RAG) methods can enhance the performance of LLMs by incorporating retrieved knowledge chunks into the generation process. In general, the retrieval and generation steps usually have different requirements for these knowledge chunks. The retrieval step benefits from comprehensive information to improve retrieval accuracy, whereas excessively long chunks may introduce redundant contextual information, thereby diminishing both the effectiveness and efficiency of the generation process. However, existing RAG methods typically employ identical representations of knowledge chunks for both retrieval and generation, resulting in suboptimal performance. In this paper, we propose a heterogeneous RAG framework (\myname) that decouples the representations of knowledge chunks for retrieval and generation, thereby enhancing the LLMs in both effectiveness and efficiency. Specifically, we utilize short chunks to represent knowledge to adapt the generation step and utilize the corresponding chunk with its contextual information from multi-granular views to enhance retrieval accuracy. We further introduce an adaptive prompt tuning method for the retrieval model to adapt the heterogeneous retrieval augmented generation process. Extensive experiments demonstrate that \myname achieves significant improvements compared to baselines.

Zhihan Zheng, Haitao Yuan, Minxiao Chen et al.

En Route Travel Time Estimation (ER-TTE) aims to learn driving patterns from traveled routes to achieve rapid and accurate real-time predictions. However, existing methods ignore the complexity and dynamism of real-world traffic systems, resulting in significant gaps in efficiency and accuracy in real-time scenarios. Addressing this issue is a critical yet challenging task. This paper proposes a novel framework that redefines the implementation path of ER-TTE to achieve highly efficient and effective predictions. Firstly, we introduce a novel pipeline consisting of a Decision Maker and a Predictor to rectify the inefficient prediction strategies of current methods. The Decision Maker performs efficient real-time decisions to determine whether the high-complexity prediction model in the Predictor needs to be invoked, and the Predictor recalculates the travel time or infers from historical prediction results based on these decisions. Next, to tackle the dynamic and uncertain real-time scenarios, we model the online decision-making problem as a Markov decision process and design an intelligent agent based on reinforcement learning for autonomous decision-making. Moreover, to fully exploit the spatio-temporal correlation between online data and offline data, we meticulously design feature representation and encoding techniques based on the attention mechanism. Finally, to improve the flawed training and evaluation strategies of existing methods, we propose an end-to-end training and evaluation approach, incorporating curriculum learning strategies to manage spatio-temporal data for more advanced training algorithms. Extensive evaluations on three real-world datasets confirm that our method significantly outperforms state-of-the-art solutions in both accuracy and efficiency.

Zhenyan Lu, Daliang Xu, Dongqi Cai et al. · cambridge

Large language models (LLMs) are deployed on mobile devices to power killer applications such as intelligent assistants. LLMs pre-trained on general corpora often hallucinate when handling personalized or unseen queries, leading to incorrect or outdated responses. Knowledge editing addresses this by identifying and adjusting a small crucial portion of model weights, without compromising the general knowledge. However, prior knowledge editing methods are impractical to run on local devices due to the resource-heavy backpropagation (BP) needed for updates. We present MobiEdit, the first mobile knowledge editing framework that enables efficient LLM personalization on commercial off-the-shelf (COTS) mobile devices. MobiEdit replaces full-precision BP with quantized forward-only gradient estimation, thus compatible with the energy-efficient mobile neural processing units (NPUs). MobiEdit replaces full-precision backpropagation with quantized forward-only gradient estimation, making it compatible with energy-efficient mobile NPUs. To further improve gradient estimation efficiency, we introduce two optimizations: an early stoping mechanism that adaptively terminates editing upon success and a prefix cache that reuses computation across steps. Our approach enables real-time editing of a 3B-parameter model (Qwen2.5-3B-Instruct) on COTS mobile devices with 7.6$\times$ less memory, 14.7 $\times$ less energy and 3.6$\times$ less latency compared to previous knowledge editing methods.

Zhiyang Chen, Daliang Xu, Haiyang Shen et al.

Enhancing on-device large language models (LLMs) with contextual information from local data enables personalized and task-aware generation, powering use cases such as intelligent assistants and UI agents. While recent developments in neural processors have substantially improved the efficiency of prefill on mobile devices, the token-by-token generation process still suffers from high latency and limited hardware utilization due to its inherently memory-bound characteristics. This work presents sd.npu, a mobile inference framework that integrates speculative decoding with dynamic hardware scheduling to accelerate context-aware text generation on mobile devices. The framework introduces three synergistic components: (1) adaptive execution scheduling, which dynamically balances compute graphs between prefill and decoding phases; (2) context-aligned drafting, which improves speculative efficiency through lightweight online calibration to current tasks; and (3) hardware-efficient draft extension, which reuses and expands intermediate sequences to improve processing parallelism and reduce verification cost. Experiments on multiple smartphones and representative workloads show consistent improvements of up to 3.8x in generation speed and 4.7x in energy efficiency compared with existing mobile inference solutions. Component-level analysis further validates the contribution of each optimization.

Chuntao Ding, Jianhang Xie, Junna Zhang et al.

It has become mainstream to deploy Convolutional Neural Network (CNN) models on ubiquitous Internet of Things (IoT) devices with the help of the cloud to provide users with a variety of high-quality services. Most existing methods have two limitations: (i) low robustness in handling corrupted image data collected by IoT devices; and (ii) high consumption of computational and transmission resources. To this end, we propose the Grouped NonLinear transformation generation method (GroupNL), which generates diversified feature maps by utilizing data-agnostic Nonlinear Transformation Functions (NLFs) to improve the robustness of the CNN model. Specifically, partial convolution filters are designated as seed filters in a convolutional layer, and a small set of feature maps, i.e., seed feature maps, are first generated based on vanilla convolution operation. Then, we split seed feature maps into several groups, each with a set of different NLFs, to generate corresponding diverse feature maps with in-place nonlinear processing. Moreover, GroupNL effectively reduces the parameter transmission between multiple nodes during model training by setting the hyperparameters of NLFs to random initialization and not updating them during model training, and reduces the computing resources by using NLFs to generate feature maps instead of most feature maps generated based on sliding windows. Experimental results on CIFAR-10, GTSRB, CIFAR-10-C, Icons50, and ImageNet-1K datasets in NVIDIA RTX GPU platforms show that the proposed GroupNL outperforms other state-of-the-art methods in model robust and training acceleration. Specifically, on the Icons-50 dataset, the accuracy of GroupNL-ResNet-18 achieves approximately 2.86% higher than the vanilla ResNet-18. GroupNL improves training speed by about 53% compared to vanilla CNN when trained on a cluster of 8 NVIDIA RTX 4090 GPUs on the ImageNet-1K dataset.

Peiru Yang, Jinhua Yin, Haoran Zheng et al.

Multimodal retrieval-augmented generation (RAG) systems enhance large vision-language models by integrating cross-modal knowledge, enabling their increasing adoption across real-world multimodal tasks. These knowledge databases may contain sensitive information that requires privacy protection. However, multimodal RAG systems inherently grant external users indirect access to such data, making them potentially vulnerable to privacy attacks, particularly membership inference attacks (MIAs). % Existing MIA methods targeting RAG systems predominantly focus on the textual modality, while the visual modality remains relatively underexplored. To bridge this gap, we propose MrM, the first black-box MIA framework targeted at multimodal RAG systems. It utilizes a multi-object data perturbation framework constrained by counterfactual attacks, which can concurrently induce the RAG systems to retrieve the target data and generate information that leaks the membership information. Our method first employs an object-aware data perturbation method to constrain the perturbation to key semantics and ensure successful retrieval. Building on this, we design a counterfact-informed mask selection strategy to prioritize the most informative masked regions, aiming to eliminate the interference of model self-knowledge and amplify attack efficacy. Finally, we perform statistical membership inference by modeling query trials to extract features that reflect the reconstruction of masked semantics from response patterns. Experiments on two visual datasets and eight mainstream commercial visual-language models (e.g., GPT-4o, Gemini-2) demonstrate that MrM achieves consistently strong performance across both sample-level and set-level evaluations, and remains robust under adaptive defenses.

Mengwei Xu, Wangsong Yin, Dongqi Cai et al.

Large foundation models, including large language models (LLMs), vision transformers (ViTs), diffusion, and LLM-based multimodal models, are revolutionizing the entire machine learning lifecycle, from training to deployment. However, the substantial advancements in versatility and performance these models offer come at a significant cost in terms of hardware resources. To support the growth of these large models in a scalable and environmentally sustainable way, there has been a considerable focus on developing resource-efficient strategies. This survey delves into the critical importance of such research, examining both algorithmic and systemic aspects. It offers a comprehensive analysis and valuable insights gleaned from existing literature, encompassing a broad array of topics from cutting-edge model architectures and training/serving algorithms to practical system designs and implementations. The goal of this survey is to provide an overarching understanding of how current approaches are tackling the resource challenges posed by large foundation models and to potentially inspire future breakthroughs in this field.

Qiyang Zhang, Xiang Li, Xiangying Che et al.

Deploying deep learning (DL) on mobile devices has been a notable trend in recent years. To support fast inference of on-device DL, DL libraries play a critical role as algorithms and hardware do. Unfortunately, no prior work ever dives deep into the ecosystem of modern DL libs and provides quantitative results on their performance. In this paper, we first build a comprehensive benchmark that includes 6 representative DL libs and 15 diversified DL models. We then perform extensive experiments on 10 mobile devices, which help reveal a complete landscape of the current mobile DL libs ecosystem. For example, we find that the best-performing DL lib is severely fragmented across different models and hardware, and the gap between those DL libs can be rather huge. In fact, the impacts of DL libs can overwhelm the optimizations from algorithms or hardware, e.g., model quantization and GPU/DSP-based heterogeneous computing. Finally, atop the observations, we summarize practical implications to different roles in the DL lib ecosystem.

Jinliang Yuan, Mengwei Xu, Xiao Ma et al.

Federated learning (FL) was designed to enable mobile phones to collaboratively learn a global model without uploading their private data to a cloud server. However, exiting FL protocols has a critical communication bottleneck in a federated network coupled with privacy concerns, usually powered by a wide-area network (WAN). Such a WAN-driven FL design leads to significantly high cost and much slower model convergence. In this work, we propose an efficient FL protocol, which involves a hierarchical aggregation mechanism in the local-area network (LAN) due to its abundant bandwidth and almost negligible monetary cost than WAN. Our proposed FL can accelerate the learning process and reduce the monetary cost with frequent local aggregation in the same LAN and infrequent global aggregation on a cloud across WAN. We further design a concrete FL platform, namely LanFL, that incorporates several key techniques to handle those challenges introduced by LAN: cloud-device aggregation architecture, intra-LAN peer-to-peer (p2p) topology generation, inter-LAN bandwidth capacity heterogeneity. We evaluate LanFL on 2 typical Non-IID datasets, which reveals that LanFL can significantly accelerate FL training (1.5x-6.0x), save WAN traffic (18.3x-75.6x), and reduce monetary cost (3.8x-27.2x) while preserving the model accuracy.

Jinliang Yuan, Mengwei Xu, Yuxin Zhao et al.

To preserve user privacy while enabling mobile intelligence, techniques have been proposed to train deep neural networks on decentralized data. However, training over decentralized data makes the design of neural architecture quite difficult as it already was. Such difficulty is further amplified when designing and deploying different neural architectures for heterogeneous mobile platforms. In this work, we propose an automatic neural architecture search into the decentralized training, as a new DNN training paradigm called Federated Neural Architecture Search, namely federated NAS. To deal with the primary challenge of limited on-client computational and communication resources, we present FedNAS, a highly optimized framework for efficient federated NAS. FedNAS fully exploits the key opportunity of insufficient model candidate re-training during the architecture search process, and incorporates three key optimizations: parallel candidates training on partial clients, early dropping candidates with inferior performance, and dynamic round numbers. Tested on large-scale datasets and typical CNN architectures, FedNAS achieves comparable model accuracy as state-of-the-art NAS algorithm that trains models with centralized data, and also reduces the client cost by up to two orders of magnitude compared to a straightforward design of federated NAS.

Binbin Huang, Zhongjin Lia, Peng Tang et al.

It is a big challenge for resource-limited mobile devices (MDs) to execute various complex and energy-consumed mobile applications. Fortunately, as a novel computing paradigm, edge computing (MEC) can provide abundant computing resources to execute all or parts of the tasks of MDs and thereby can greatly reduce the energy of MD and improve the QoS of applications. However, offloading workflow tasks to the MEC servers are liable to external security threats (e.g., snooping, alteration). In this paper, we propose a security and energy efficient computation offloading (SEECO) strategy for service workflows in MEC environment, the goal of which is to optimize the energy consumption under the risk probability and deadline constraints. First, we build a security overhead model to measure the execution time of security services. Then, we formulate the computation offloading problem by incorporating the security, energy consumption and execution time of workflow application. Finally, based on the genetic algorithm (GA), the corresponding coding strategies of SEECO are devised by considering tasks execution order and location and security services selection. Extensive experiments with the variety of workflow parameters demonstrate that SEECO strategy can achieve the security and energy efficiency for the mobile applications.

Jinliang Xu, Shangguang Wang, Fangchun Yang et al.

Cognitive inference of user demographics, such as gender and age, plays an important role in creating user profiles for adjusting marketing strategies and generating personalized recommendations because user demographic data is usually not available due to data privacy concerns. At present, users can readily express feedback regarding products or services that they have purchased. During this process, user demographics are concealed, but the data has never yet been successfully utilized to contribute to the cognitive inference of user demographics. In this paper, we investigate the inference power of user ratings data, and propose a simple yet general cognitive inference model, called rating to profile (R2P), to infer user demographics from user provided ratings. In particular, the proposed R2P model can achieve the following: 1. Correctly integrate user ratings into model training. 2.Infer multiple demographic attributes of users simultaneously, capturing the underlying relevance between different demographic attributes. 3. Train its two components, i.e. feature extractor and classifier, in an integrated manner under a supervised learning paradigm, which effectively helps to discover useful hidden patterns from highly sparse ratings data. We introduce how to incorporate user ratings data into the research field of cognitive inference of user demographic data, and detail the model development and optimization process for the proposed R2P. Extensive experiments are conducted on two real-world ratings datasets against various compared state-of-the-art methods, and the results from multiple aspects demonstrate that our proposed R2P model can significantly improve on the cognitive inference performance of user demographic data.

Jinliang Xu, Shangguang Wang, Fangchun Yang et al.

Inference of user context information, including user's gender, age, marital status, location and so on, has been proven to be valuable for building context aware recommender system. However, prevalent existing studies on user context inference have two shortcommings: 1. focusing on only a single data source (e.g. Internet browsing logs, or mobile call records), and 2. ignoring the interdependence of multiple user contexts (e.g. interdependence between age and marital status), which have led to poor inference performance. To solve this problem, in this paper, we first exploit tensor outer product to fuse multiple data sources in the feature space to obtain an extensional user feature representation. Following this, by taking this extensional user feature representation as input, we propose a multiple attribute probabilistic model called MulAProM to infer user contexts that can take advantage of the interdependence between them. Our study is based on large telecommunication datasets from the local mobile operator of Shanghai, China, and consists of two data sources, 4.6 million call detail records and 7.5 million data traffic records of 8,000 mobile users, collected in the course of six months. The experimental results show that our model can outperform other models in terms of \emph{recall}, \emph{precision}, and the \emph{F1-measure}.