Large Vision and Language Models have exhibited remarkable human-like intelligence in tasks such as natural language comprehension, problem-solving, logical reasoning, and knowledge retrieval. However, training and serving these models require substantial computational resources, posing a significant barrier to their widespread application and further research. To mitigate this challenge, various model compression techniques have been developed to reduce computational requirements. Nevertheless, existing methods often employ uniform quantization configurations, failing to account for the varying difficulties across different layers in quantizing large neural network models. This paper tackles this issue by leveraging layer-sensitivity features, such as activation sensitivity and weight distribution Kurtosis, to identify layers that are challenging to quantize accurately and allocate additional memory budget. The proposed methods, named SensiBoost and KurtBoost, respectively, demonstrate notable improvement in quantization accuracy, achieving up to 9% lower perplexity with only a 2% increase in memory budget on LLama models compared to the baseline.
Large Language Models (LLMs) have grown increasingly expensive to deploy, driving the need for effective model compression techniques. While block pruning offers a straightforward approach to reducing model size, existing methods often struggle to maintain performance or require substantial computational resources for recovery. We present IteRABRe, a simple yet effective iterative pruning method that achieves superior compression results while requiring minimal computational resources. Using only 2.5M tokens for recovery, our method outperforms baseline approaches by ~3% on average when compressing the Llama3.1-8B and Qwen2.5-7B models. IteRABRe demonstrates particular strength in the preservation of linguistic capabilities, showing an improvement 5% over the baselines in language-related tasks. Our analysis reveals distinct pruning characteristics between these models, while also demonstrating preservation of multilingual capabilities.
The rapid advancement of artificial intelligence (AI) technologies has led to an increasing deployment of AI models on edge and terminal devices, driven by the proliferation of the Internet of Things (IoT) and the need for real-time data processing. This survey comprehensively explores the current state, technical challenges, and future trends of on-device AI models. We define on-device AI models as those designed to perform local data processing and inference, emphasizing their characteristics such as real-time performance, resource constraints, and enhanced data privacy. The survey is structured around key themes, including the fundamental concepts of AI models, application scenarios across various domains, and the technical challenges faced in edge environments. We also discuss optimization and implementation strategies, such as data preprocessing, model compression, and hardware acceleration, which are essential for effective deployment. Furthermore, we examine the impact of emerging technologies, including edge computing and foundation models, on the evolution of on-device AI models. By providing a structured overview of the challenges, solutions, and future directions, this survey aims to facilitate further research and application of on-device AI, ultimately contributing to the advancement of intelligent systems in everyday life.
The widespread deployment of InfRared Small-Target Detection(IRSTD) algorithms on edge devices necessitates the exploration of model compression techniques. Binary neural networks (BNNs) are distinguished by their exceptional efficiency in model compression. However, the small size of infrared targets introduces stringent precision requirements for the IRSTD task, while the inherent precision loss during binarization presents a significant challenge. To address this, we propose the Binarized Infrared Small-Target Detection Network (BiisNet), which preserves the core operations of binarized convolutions while integrating full-precision features into the network's information flow. Specifically, we propose the Dot-Binary Convolution, which retains fine-grained semantic information in feature maps while still leveraging the binarized convolution operations. In addition, we introduce a smooth and adaptive Dynamic Softsign function, which provides more comprehensive and progressively finer gradient during back-propagation, enhancing model stability and promoting an optimal weight distribution.Experimental results demonstrate that BiisNet not only significantly outperforms other binary architectures but also demonstrates strong competitiveness among state-of-the-art full-precision models.
In recent years, vision transformers (ViTs) have emerged as powerful and promising techniques for computer vision tasks such as image classification, object detection, and segmentation. Unlike convolutional neural networks (CNNs), which rely on hierarchical feature extraction, ViTs treat images as sequences of patches and leverage self-attention mechanisms. However, their high computational complexity and memory demands pose significant challenges for deployment on resource-constrained edge devices. To address these limitations, extensive research has focused on model compression techniques and hardware-aware acceleration strategies. Nonetheless, a comprehensive review that systematically categorizes these techniques and their trade-offs in accuracy, efficiency, and hardware adaptability for edge deployment remains lacking. This survey bridges this gap by providing a structured analysis of model compression techniques, software tools for inference on edge, and hardware acceleration strategies for ViTs. We discuss their impact on accuracy, efficiency, and hardware adaptability, highlighting key challenges and emerging research directions to advance ViT deployment on edge platforms, including graphics processing units (GPUs), application-specific integrated circuit (ASICs), and field-programmable gate arrays (FPGAs). The goal is to inspire further research with a contemporary guide on optimizing ViTs for efficient deployment on edge devices.
Modern large language models (LLMs) exhibit critical vulnerabilities to poison pill attacks: localized data poisoning that alters specific factual knowledge while preserving overall model utility. We systematically demonstrate these attacks exploit inherent architectural properties of LLMs, achieving 54.6% increased retrieval inaccuracy on long-tail knowledge versus dominant topics and up to 25.5% increase retrieval inaccuracy on compressed models versus original architectures. Through controlled mutations (e.g., temporal/spatial/entity alterations) and, our method induces localized memorization deterioration with negligible impact on models' performance on regular standard benchmarks (e.g., <2% performance drop on MMLU/GPQA), leading to potential detection evasion. Our findings suggest: (1) Disproportionate vulnerability in long-tail knowledge may result from reduced parameter redundancy; (2) Model compression may increase attack surfaces, with pruned/distilled models requiring 30% fewer poison samples for equivalent damage; (3) Associative memory enables both spread of collateral damage to related concepts and amplification of damage from simultaneous attack, particularly for dominant topics. These findings raise concerns over current scaling paradigms since attack costs are lowering while defense complexity is rising. Our work establishes poison pills as both a security threat and diagnostic tool, revealing critical security-efficiency trade-offs in language model compression that challenges prevailing safety assumptions.
Small language models (SLMs) have become increasingly prominent in the deployment on edge devices due to their high efficiency and low computational cost. While researchers continue to advance the capabilities of SLMs through innovative training strategies and model compression techniques, the security risks of SLMs have received considerably less attention compared to large language models (LLMs).To fill this gap, we provide a comprehensive empirical study to evaluate the security performance of 13 state-of-the-art SLMs under various jailbreak attacks. Our experiments demonstrate that most SLMs are quite susceptible to existing jailbreak attacks, while some of them are even vulnerable to direct harmful prompts.To address the safety concerns, we evaluate several representative defense methods and demonstrate their effectiveness in enhancing the security of SLMs. We further analyze the potential security degradation caused by different SLM techniques including architecture compression, quantization, knowledge distillation, and so on. We expect that our research can highlight the security challenges of SLMs and provide valuable insights to future work in developing more robust and secure SLMs.
Language model compression through knowledge distillation has emerged as a promising approach for deploying large language models in resource-constrained environments. However, existing methods often struggle to maintain performance when distilling multilingual models, especially for low-resource languages. In this paper, we present a novel hybrid distillation approach that combines traditional knowledge distillation with a simplified attention matching mechanism, specifically designed for multilingual contexts. Our method introduces an extremely compact student model architecture, significantly smaller than conventional multilingual models. We evaluate our approach on five African languages: Kinyarwanda, Swahili, Hausa, Igbo, and Yoruba. The distilled student model; AfroXLMR-Comet successfully captures both the output distribution and internal attention patterns of a larger teacher model (AfroXLMR-Large) while reducing the model size by over 85%. Experimental results demonstrate that our hybrid approach achieves competitive performance compared to the teacher model, maintaining an accuracy within 85% of the original model's performance while requiring substantially fewer computational resources. Our work provides a practical framework for deploying efficient multilingual models in resource-constrained environments, particularly benefiting applications involving African languages.
Motivated by reducing the computational and storage costs of LLMs, model compression and KV cache compression have attracted much attention from researchers. However, current methods predominantly emphasize maintaining the performance of compressed LLMs, as measured by perplexity or simple accuracy on tasks of common sense knowledge QA and basic arithmetic reasoning. In this blog, we present a brief review of recent advancements in LLMs related to retrieval-augmented generation, multi-step reasoning, external tools, and computational expressivity, all of which substantially enhance LLM performance. Then, we propose a lottery LLM hypothesis suggesting that for a given LLM and task, there exists a smaller lottery LLM capable of producing the same performance as the original LLM with the assistance of multi-step reasoning and external tools. Based on the review of current progress in LLMs, we discuss and summarize the essential capabilities that the lottery LLM and KV cache compression must possess, which are currently overlooked in existing methods.
Outliers have been widely observed in Large Language Models (LLMs), significantly impacting model performance and posing challenges for model compression. Understanding the functionality and formation mechanisms of these outliers is critically important. Existing works, however, largely focus on reducing the impact of outliers from an algorithmic perspective, lacking an in-depth investigation into their causes and roles. In this work, we provide a detailed analysis of the formation process, underlying causes, and functions of outliers in LLMs. We define and categorize three types of outliers-activation outliers, weight outliers, and attention outliers-and analyze their distributions across different dimensions, uncovering inherent connections between their occurrences and their ultimate influence on the attention mechanism. Based on these observations, we hypothesize and explore the mechanisms by which these outliers arise and function, demonstrating through theoretical derivations and experiments that they emerge due to the self-attention mechanism's softmax operation. These outliers act as implicit context-aware scaling factors within the attention mechanism. As these outliers stem from systematic influences, we term them systematic outliers. Our study not only enhances the understanding of Transformer-based LLMs but also shows that structurally eliminating outliers can accelerate convergence and improve model compression. The code is avilable at https://github.com/an-yongqi/systematic-outliers.
Large language models (LLMs) exhibit excellent performance in various tasks. However, the memory requirements of LLMs present a great challenge when deploying on memory-limited devices, even for quantized LLMs. This paper introduces a framework to compress LLM after quantization further, achieving about 2.2x compression ratio. A compression-aware quantization is first proposed to enhance model weight compressibility by re-scaling the model parameters before quantization, followed by a pruning method to improve further. Upon this, we notice that decompression can be a bottleneck during practical scenarios. We then give a detailed analysis of the trade-off between memory usage and latency brought by the proposed method. A speed-adaptive method is proposed to overcome it. The experimental results show inference with the compressed model can achieve a 40% reduction in memory size with negligible loss in accuracy and inference speed.
Large language models (LLMs) have demonstrated remarkable capabilities, yet prohibitive parameter complexity often hinders their deployment. Existing singular value decomposition (SVD) based compression methods simply deem singular values as importance scores of decomposed components. However, this importance ordered by singular values does not necessarily correlate with the performance of a downstream task. In this work, we introduce SoCo (Singular spectrum optimization for large language model Compression), a novel compression framework that learns to rescale the decomposed components of SVD in a data-driven manner. Concretely, we employ a learnable diagonal matrix to assign importance scores for singular spectrum and develop a three-stage training process that progressively refines these scores from initial coarse compression to fine-grained sparsification-thereby striking an effective balance between aggressive model compression and performance preservation. Thanks to the learnable singular spectrum, SoCo adaptively prunes components according to the sparsified importance scores, rather than relying on the fixed order of singular values. More importantly, the remaining components with amplified importance scores can compensate for the loss of the pruned ones. Experimental evaluations across multiple LLMs and benchmarks demonstrate that SoCo surpasses the state-of-the-art methods in model compression.
Large language models (LLMs) have demonstrated remarkable performance across a wide range of industrial applications, from search and recommendations to generative tasks. Although scaling laws indicate that larger models generally yield better generalization and performance, their substantial computational requirements often render them impractical for many real-world scenarios at scale. In this paper, we present methods and insights for training small language models (SLMs) that deliver high performance and efficiency in deployment. We focus on two key techniques: (1) knowledge distillation and (2) model compression via quantization and pruning. These approaches enable SLMs to retain much of the quality of their larger counterparts while significantly reducing training, serving costs, and latency. We detail the impact of these techniques on a variety of use cases at a large professional social network platform and share deployment lessons - including hardware optimization strategies that enhance speed and throughput for both predictive and reasoning-based applications.
The remarkable performance of large language models (LLMs) in various language tasks has attracted considerable attention. However, the ever-increasing size of these models presents growing challenges for deployment and inference. Structured pruning, an effective model compression technique, is gaining increasing attention due to its ability to enhance inference efficiency. Nevertheless, most previous optimization-based structured pruning methods sacrifice the uniform structure across layers for greater flexibility to maintain performance. The heterogeneous structure hinders the effective utilization of off-the-shelf inference acceleration techniques and impedes efficient configuration for continued training. To address this issue, we propose a novel masking learning paradigm based on minimax optimization to obtain the uniform pruned structure by optimizing the masks under sparsity regularization. Extensive experimental results demonstrate that our method can maintain high performance while ensuring the uniformity of the pruned model structure, thereby outperforming existing SOTA methods.
The rapid development of Vision Foundation Models (VFMs), particularly Vision Transformers (ViT) and Segment Anything Model (SAM), has sparked significant advances in the field of medical image analysis. These models have demonstrated exceptional capabilities in capturing long-range dependencies and achieving high generalization in segmentation tasks. However, adapting these large models to medical image analysis presents several challenges, including domain differences between medical and natural images, the need for efficient model adaptation strategies, and the limitations of small-scale medical datasets. This paper reviews the state-of-the-art research on the adaptation of VFMs to medical image segmentation, focusing on the challenges of domain adaptation, model compression, and federated learning. We discuss the latest developments in adapter-based improvements, knowledge distillation techniques, and multi-scale contextual feature modeling, and propose future directions to overcome these bottlenecks. Our analysis highlights the potential of VFMs, along with emerging methodologies such as federated learning and model compression, to revolutionize medical image analysis and enhance clinical applications. The goal of this work is to provide a comprehensive overview of current approaches and suggest key areas for future research that can drive the next wave of innovation in medical image segmentation.
With the emergence of Mixture-of-Experts (MoE), the efficient scaling of model size has accelerated the development of large language models in recent years. However, their high memory requirements prevent their use in resource-constrained environments. While knowledge distillation (KD) has been a proven method for model compression, its application to MoE teacher models remains underexplored. Through our investigation, we discover that non-activated experts in MoE models possess valuable knowledge that benefits student models. We further demonstrate that existing KD methods are not optimal for compressing MoE models, as they fail to leverage this knowledge effectively. To address this, we propose two intuitive MoE-specific KD methods for the first time: Knowledge Augmentation (KA) and Student-Aware Router (SAR), both designed to effectively extract knowledge from all experts. Specifically, KA augments knowledge by sampling experts multiple times, while SAR uses all experts and adjusts the expert weights through router training to provide optimal knowledge. Extensive experiments show that our methods outperform conventional KD methods, demonstrating their effectiveness for MoE teacher models.
We introduce model folding, a novel data-free model compression technique that merges structurally similar neurons across layers, significantly reducing the model size without the need for fine-tuning or access to training data. Unlike existing methods, model folding preserves data statistics during compression by leveraging k-means clustering, and using novel data-free techniques to prevent variance collapse or explosion. Our theoretical framework and experiments across standard benchmarks, including ResNet18 and LLaMA-7B, demonstrate that model folding achieves comparable performance to data-driven compression techniques and outperforms recently proposed data-free methods, especially at high sparsity levels. This approach is particularly effective for compressing large-scale models, making it suitable for deployment in resource-constrained environments.
Vision Large Language Models (VLMs) combine visual understanding with natural language processing, enabling tasks like image captioning, visual question answering, and video analysis. While VLMs show impressive capabilities across domains such as autonomous vehicles, smart surveillance, and healthcare, their deployment on resource-constrained edge devices remains challenging due to processing power, memory, and energy limitations. This survey explores recent advancements in optimizing VLMs for edge environments, focusing on model compression techniques, including pruning, quantization, knowledge distillation, and specialized hardware solutions that enhance efficiency. We provide a detailed discussion of efficient training and fine-tuning methods, edge deployment challenges, and privacy considerations. Additionally, we discuss the diverse applications of lightweight VLMs across healthcare, environmental monitoring, and autonomous systems, illustrating their growing impact. By highlighting key design strategies, current challenges, and offering recommendations for future directions, this survey aims to inspire further research into the practical deployment of VLMs, ultimately making advanced AI accessible in resource-limited settings.
Knowledge Distillation (KD) has emerged as a prominent technique for model compression. However, conventional KD approaches primarily focus on homogeneous architectures with identical tokenizers, constraining their applicability in cross-architecture scenarios. As for the cross-tokenizer KD, the differences in the tokenizers give rise to two fundamental challenges: (1) sequence misalignment caused by divergent tokenization strategies, and (2) mismatched vocabulary size and composition. While existing probability-matching methods attempt to address these issues, their efficacy remains limited due to suboptimal alignment in both the sequence and vocabulary aspects. To overcome these limitations, we propose Contextual Dynamic Mapping (CDM), a novel cross-tokenizer distillation framework that employs contextual information to enhance sequence alignment precision and dynamically improves vocabulary mapping. We evaluated the effectiveness of our approach across five advanced and widely-used model families (i.e, LLama3, Phi3, Gemma2, OPT and Qwen2), which were configured into three distinct teacher-student pairs. Our method shows significant advantages over existing cross-tokenizer distillation baselines across diverse benchmarks, including instruction-following, code generation and math. Notably, our analysis reveals that combining conventional same-tokenizer distillation and cross-tokenizer distillation through CDM yields further performance improvements. The code is available at https://github.com/pppa2019/ContexualDynamicMapping
Post-training pruning has emerged as a crucial optimization technique as large language models (LLMs) continue to grow rapidly. However, the significant variations in weight distributions across different LLMs make fixed pruning strategies inadequate for multiple models. In this paper, we introduce \textbf{\textsc{OptiShear}}, an efficient evolutionary optimization framework for adaptive LLM pruning. Our framework features two key innovations: an effective search space built on our Meta pruning metric to handle diverse weight distributions, and a model-wise reconstruction error for rapid evaluation during search trials. We employ Non-dominated Sorting Genetic Algorithm III (NSGA-III) to optimize both pruning metrics and layerwise sparsity ratios. Through extensive evaluation on LLaMA-1/2/3 and Mistral models (7B-70B) across multiple benchmarks, we demonstrate that our adaptive pruning metrics consistently outperform existing methods. Additionally, our discovered layerwise sparsity ratios enhance the effectiveness of other pruning metrics. The framework exhibits strong cross-task and cross-model generalizability, providing a cost-effective solution for model compression.