Federated learning (FL) has enabled training of multilingual large language models (LLMs) on diverse and decentralized multilingual data, especially on low-resource languages. To improve client-specific performance, personalization via the use of parameter-efficient fine-tuning (PEFT) modules such as LoRA is common. This involves a personalization strategy (PS), such as the design of the PEFT adapter structures (e.g., in which layers to add LoRAs and what ranks) and choice of hyperparameters (e.g., learning rates) for fine-tuning. Instead of manual PS configuration, we propose FedP²EFT, a federated learning-to-personalize method for multilingual LLMs in cross-device FL settings. Unlike most existing PEFT structure selection methods, which are prone to overfitting low-data regimes, FedP²EFT collaboratively learns the optimal personalized PEFT structure for each client via Bayesian sparse rank selection. Evaluations on both simulated and real-world multilingual FL benchmarks demonstrate that FedP²EFT largely outperforms existing personalized fine-tuning methods, while complementing other existing FL methods.

In spite of the great potential of large language models (LLMs) across various tasks, their deployment on resource-constrained devices remains challenging due to their excessive computational and memory demands. Quantization has emerged as an effective solution by storing weights in reduced precision. However, utilizing low precisions (i.e.~2/3-bit) to substantially alleviate the memory-boundedness of LLM decoding, still suffers from prohibitive performance drop. In this work, we argue that existing approaches fail to explore the diversity in computational patterns, redundancy, and sensitivity to approximations of the different phases of LLM inference, resorting to a uniform quantization policy throughout. Instead, we propose a novel phase-aware method that selectively allocates precision during different phases of LLM inference, achieving both strong context extraction during prefill and efficient memory bandwidth utilization during decoding. To further address the memory-boundedness of the decoding phase, we introduce Progressive Mixed-Precision Decoding (PMPD), a technique that enables the gradual lowering of precision deeper in the generated sequence, together with a spectrum of precision-switching schedulers that dynamically drive the precision-lowering decisions in either task-adaptive or prompt-adaptive manner. Extensive evaluation across diverse language tasks shows that when targeting Nvidia GPUs, PMPD achieves 1.4$-$12.2$\times$ speedup in matrix-vector multiplications over fp16 models, while when targeting an LLM-optimized NPU, our approach delivers a throughput gain of 3.8$-$8.0$\times$ over fp16 models and up to 1.54$\times$ over uniform quantization approaches while preserving the output quality.

Pretrained large language models (LLMs) have emerged as a cornerstone in modern natural language processing, with their utility expanding to various applications and languages. However, the fine-tuning of multilingual LLMs, particularly for low-resource languages, is fraught with challenges steming from data-sharing restrictions (the physical border) and from the inherent linguistic differences (the linguistic border). These barriers hinder users of various languages, especially those in low-resource regions, from fully benefiting from the advantages of LLMs. To overcome these challenges, we propose the Federated Prompt Tuning Paradigm for Multilingual Scenarios, which leverages parameter-efficient fine-tuning in a manner that preserves user privacy. We have designed a comprehensive set of experiments and introduced the concept of "language distance" to highlight the several strengths of this paradigm. Even under computational constraints, our method not only bolsters data efficiency but also facilitates mutual enhancements across languages, particularly benefiting low-resource ones. Compared to traditional local crosslingual transfer tuning methods, our approach achieves a 6.9\% higher accuracy, reduces the training parameters by over 99\%, and demonstrates stronger cross-lingual generalization. Such findings underscore the potential of our approach to promote social equality, ensure user privacy, and champion linguistic diversity.

Federated learning (FL) has enabled distributed learning of a model across multiple clients in a privacy-preserving manner. One of the main challenges of FL is to accommodate clients with varying hardware capacities; clients have differing compute and memory requirements. To tackle this challenge, recent state-of-the-art approaches leverage the use of early exits. Nonetheless, these approaches fall short of mitigating the challenges of joint learning multiple exit classifiers, often relying on hand-picked heuristic solutions for knowledge distillation among classifiers and/or utilizing additional layers for weaker classifiers. In this work, instead of utilizing multiple classifiers, we propose a recurrent early exit approach named ReeFL that fuses features from different sub-models into a single shared classifier. Specifically, we use a transformer-based early-exit module shared among sub-models to i) better exploit multi-layer feature representations for task-specific prediction and ii) modulate the feature representation of the backbone model for subsequent predictions. We additionally present a per-client self-distillation approach where the best sub-model is automatically selected as the teacher of the other sub-models at each client. Our experiments on standard image and speech classification benchmarks across various emerging federated fine-tuning baselines demonstrate ReeFL effectiveness over previous works.

Data point selection (DPS) is becoming a critical topic in deep learning due to the ease of acquiring uncurated training data compared to the difficulty of obtaining curated or processed data. Existing approaches to DPS are predominantly based on a bi-level optimisation (BLO) formulation, which is demanding in terms of memory and computation, and exhibits some theoretical defects regarding minibatches. Thus, we propose a novel Bayesian approach to DPS. We view the DPS problem as posterior inference in a novel Bayesian model where the posterior distributions of the instance-wise weights and the main neural network parameters are inferred under a reasonable prior and likelihood model. We employ stochastic gradient Langevin MCMC sampling to learn the main network and instance-wise weights jointly, ensuring convergence even with minibatches. Our update equation is comparable to the widely used SGD and much more efficient than existing BLO-based methods. Through controlled experiments in both the vision and language domains, we present the proof-of-concept. Additionally, we demonstrate that our method scales effectively to large language models and facilitates automated per-task optimization for instruction fine-tuning datasets.

Recent image degradation estimation methods have enabled single-image super-resolution (SR) approaches to better upsample real-world images. Among these methods, explicit kernel estimation approaches have demonstrated unprecedented performance at handling unknown degradations. Nonetheless, a number of limitations constrain their efficacy when used by downstream SR models. Specifically, this family of methods yields i) excessive inference time due to long per-image adaptation times and ii) inferior image fidelity due to kernel mismatch. In this work, we introduce a learning-to-learn approach that meta-learns from the information contained in a distribution of images, thereby enabling significantly faster adaptation to new images with substantially improved performance in both kernel estimation and image fidelity. Specifically, we meta-train a kernelgenerating GAN, named MetaKernelGAN, on a range of tasks, such that when a new image is presented, the generator starts from an informed kernel estimate and the discriminator starts with a strong capability to distinguish between patch distributions. Compared with state-of-the-art methods, our experiments show that MetaKernelGAN better estimates the magnitude and covariance of the kernel, leading to state-of-the-art blind SR results within a similar computational regime when combined with a non-blind SR model. Through supervised learning of an unsupervised learner, our method maintains the generalizability of the unsupervised learner, improves the optimization stability of kernel estimation, and hence image adaptation, and leads to a faster inference with a speedup between 14.24 to 102.1× over existing methods.0

In recent years, image and video delivery systems have begun integrating deep learning super-resolution (SR) approaches, leveraging their unprecedented visual enhancement capabilities while reducing reliance on networking conditions. Nevertheless, deploying these solutions on mobile devices still remains an active challenge as SR models are excessively demanding with respect to workload and memory footprint. Despite recent progress on on-device SR frameworks, existing systems either penalize visual quality, lead to excessive energy consumption or make inefficient use of the available resources. This work presents NAWQ-SR, a novel framework for the efficient on-device execution of SR models. Through a novel hybrid-precision quantization technique and a runtime neural image codec, NAWQ-SR exploits the multi-precision capabilities of modern mobile NPUs in order to minimize latency, while meeting user-specified quality constraints. Moreover, NAWQ-SR selectively adapts the arithmetic precision at run time to equip the SR DNN's layers with wider representational power, improving visual quality beyond what was previously possible on NPUs.Altogether, NAWQ-SR achieves an average speedup of 7.9×, 3× and 1.91× over the state-of-the-art on-device SR systems that use heterogeneous processors (MobiSR), CPU (SplitSR) and NPU (XLSR), respectively.Furthermore, NAWQ-SR delivers an average of 3.2× speedup and 0.39 dB higher PSNR over status-quo INT8 NPU designs, but most importantly mitigates the negative effects of quantization on visual quality, setting a new state-of-the-art in the attainable quality of NPU-based SR.

Federated learning (FL) research has made progress in developing algorithms for distributed learning of global models, as well as algorithms for local personalization of those common models to the specifics of each client’s local data distribution. However, different FL problems may require different personalization strategies, and it may not even be possible to define an effective one-size-fits-all personalization strategy for all clients: Depending on how similar each client’s optimal predictor is to that of the global model, different personalization strategies may be preferred. In this paper, we consider the federated meta-learning problem of learning personalization strategies. Specifically, we consider meta-nets that induce the batch-norm and learning rate parameters for each client given local data statistics. By learning these meta-nets through FL, we allow the whole FL network to collaborate in learning a customized personalization strategy for each client. Empirical results show that this framework improves on a range of standard hand-crafted personalization baselines in both label and feature shift situations.

Attention-based neural networks have become pervasive in many AI tasks. Despite their excellent algorithmic performance, the use of the attention mechanism and feedforward network (FFN) demands excessive computational and memory resources, which often compromises their hardware performance. Although various sparse variants have been introduced, most approaches only focus on mitigating the quadratic scaling of attention on the algorithm level, without explicitly considering the efficiency of mapping their methods on real hardware designs. Furthermore, most efforts only focus on either the attention mechanism or the FFNs but without jointly optimizing both parts, causing most of the current designs to lack scalability when dealing with different input lengths. This paper systematically considers the sparsity patterns in different variants from a hardware perspective. On the algorithmic level, we propose FABNet, a hardware-friendly variant that adopts a unified butterfly sparsity pattern to approximate both the attention mechanism and the FFNs. On the hardware level, a novel adaptable butterfly accelerator is proposed that can be configured at runtime via dedicated hardware control to accelerate different butterfly layers using a single unified hardware engine. On the Long-Range-Arena dataset, FABNet achieves the same accuracy as the vanilla Transformer while reducing the amount of computation by 10$\sim66\times$ and the number of parameters 2$\sim22\times$. By jointly optimizing the algorithm and hardware, our FPGA-based butterfly accelerator achieves 14.2$\sim23.2\times$ speedup over state-of-the-art accelerators normalized to the same computational budget. Compared with optimized CPU and GPU designs on Raspberry Pi 4 and Jetson Nano, our system is up to $273.8\times$ and $15.1\times$ faster under the same power budget

Neural architecture search (NAS) has been very successful at outperforming human-designed convolutional neural networks (CNN) in accuracy, and when hardware information is present, latency as well. However, NAS-designed CNNs typically have a complicated topology, therefore, it may be difficult to design a custom hardware (HW) accelerator for such CNNs. We automate HW-CNN codesign using NAS by including parameters from both the CNN model and the HW accelerator, and we jointly search for the best model-accelerator pair that boosts accuracy and efficiency. We call this Codesign-NAS. In this paper we focus on defining the Codesign-NAS multiobjective optimization problem, demonstrating its effectiveness, and exploring different ways of navigating the codesign search space. For CIFAR-10 image classification, we enumerate close to 4 billion model-accelerator pairs, and find the Pareto frontier within that large search space. This allows us to evaluate three different reinforcement-learning-based search strategies. Finally, compared to ResNet on its most optimal HW accelerator from within our HW design space, we improve on CIFAR-100 classification accuracy by 1.3% while simultaneously increasing performance/area by 41% in just ~1000 GPU-hours of running Codesign-NAS.

Field-programmable gate arrays (FPGAs) have become a popular compute platform for convolutional neural network (CNN) inference; however, the design of a CNN model and its FPGA accelerator has been inherently sequential. A CNN is first prototyped with no-or-little hardware awareness to attain high accuracy; subsequently, an FPGA accelerator is tuned to that specific CNN to maximize its efficiency. Instead, we formulate a neural architecture search (NAS) optimization problem that contains parameters from both the CNN model and the FPGA accelerator, and we jointly search for the best CNN model-accelerator pair that boosts accuracy and efficiency -we call this Codesign-NAS. In this paper we focus on defining the Codesign-NAS multiobjective optimization problem, demonstrating its effectiveness, and exploring different ways of navigating the codesign search space. For Cifar-10 image classification, we enumerate close to 4 billion model-accelerator pairs, and find the Pareto frontier within that large search space. Next we propose accelerator innovations that improve the entire Pareto frontier. Finally, we compare to ResNet on a highly-tuned accelerator, and show that using codesign, we can improve on Cifar-100 classification accuracy by 1.8% while simultaneously increasing performance/area by 41% in just 1000 GPU-hours of running Codesign-NAS, thus demonstrating that our automated codesign approach is superior to sequential design of a CNN model and accelerator.

Neural architecture search (NAS) enables researchers to automatically explore broad design spaces in order to improve efficiency of neural networks. This efficiency is especially important in the case of on-device deployment, where improvements in accuracy should be balanced out with computational demands of a model. In practice, performance metrics of model are computationally expensive to obtain. Previous work uses a proxy (e.g., number of operations) or a layer-wise measurement of neural network layers to estimate end-to-end hardware performance but the imprecise prediction diminishes the quality of NAS. To address this problem, we propose BRP-NAS, an efficient hardware-aware NAS enabled by an accurate performance predictor-based on graph convolutional network (GCN). What is more, we investigate prediction quality on different metrics and show that sample efficiency of the predictor-based NAS can be improved by considering binary relations of models and an iterative data selection strategy. We show that our proposed method outperforms all prior methods on NAS-Bench-101, NAS-Bench-201 and DARTS. Finally, to raise awareness of the fact that accurate latency estimation is not a trivial task, we release LatBench -- a latency dataset of NAS-Bench-201 models running on a broad range of devices.

In recent years, convolutional networks have demonstrated unprecedented performance in the image restoration task of super-resolution (SR). SR entails the upscaling of a single low-resolution image in order to meet application-specific image quality demands and plays a key role in mobile devices. To comply with privacy regulations and reduce the overhead of cloud computing, executing SR models locally on-device constitutes a key alternative approach. Nevertheless, the excessive compute and memory requirements of SR workloads pose a challenge in mapping SR networks on resource-constrained mobile platforms. This work presents MobiSR, a novel framework for performing efficient super-resolution on-device. Given a target mobile platform, the proposed framework considers popular model compression techniques and traverses the design space to reach the highest performing trade-off between image quality and processing speed. At run time, a novel scheduler dispatches incoming image patches to the appropriate model-engine pair based on the patch's estimated upscaling difficulty in order to meet the required image quality with minimum processing latency. Quantitative evaluation shows that the proposed framework yields on-device SR designs that achieve an average speedup of 2.13x over highly-optimized parallel difficulty-unaware mappings and 4.79x over highly-optimized single compute engine implementations.

In recent years, convolutional networks have demonstrated unprecedented performance in the image restoration task of super-resolution (SR). SR entails the upscaling of a single low-resolution image in order to meet application-specific image quality demands and plays a key role in mobile devices. To comply with privacy regulations and reduce the overhead of cloud computing, executing SR models locally on-device constitutes a key alternative approach. Nevertheless, the excessive compute and memory requirements of SR workloads pose a challenge in mapping SR networks on resource-constrained mobile platforms. This work presents MobiSR, a novel framework for performing efficient super-resolution on-device. Given a target mobile platform, the proposed framework considers popular model compression techniques and traverses the design space to reach the highest performing trade-off between image quality and processing speed. At run time, a novel scheduler dispatches incoming image patches to the appropriate model-engine pair based on the patch's estimated upscaling difficulty in order to meet the required image quality with minimum processing latency. Quantitative evaluation shows that the proposed framework yields on-device SR designs that achieve an average speedup of 2.13x over highly-optimized parallel difficulty-unaware mappings and 4.79x over highly-optimized single compute engine implementations.