In industrial and environmental monitoring, achieving real-time and precise fluid flow measurement remains a critical challenge. This study applies linear quantization in FPGA-based soft sensors for fluid flow estimation, significantly enhancing Neural Network model precision by overcoming the limitations of traditional fixed-point quantization. Our approach achieves up to a 10.10% reduction in Mean Squared Error and a notable 9.39% improvement in inference speed through targeted hardware optimizations. Validated across multiple data sets, our findings demonstrate that the optimized FPGA-based quantized models can provide efficient, accurate real-time inference, offering a viable alternative to cloud-based processing in pervasive autonomous systems.
This paper presents a novel federated learning solution, QHetFed, suitable for large-scale Internet of Things deployments, addressing the challenges of large geographic span, communication resource limitation, and data heterogeneity. QHetFed is based on hierarchical federated learning over multiple device sets, where the learning process and learning parameters take the necessary data quantization and the data heterogeneity into consideration to achieve high accuracy and fast convergence. Unlike conventional hierarchical federated learning algorithms, the proposed approach combines gradient aggregation in intra-set iterations with model aggregation in inter-set iterations. We offer a comprehensive analytical framework to evaluate its optimality gap and convergence rate, and give a closed form expression for the optimal learning parameters under a deadline, that accounts for communication and computation times. Our findings reveal that QHetFed consistently achieves high learning accuracy and significantly outperforms other hierarchical algorithms, particularly in scenarios with heterogeneous data distributions.
Next-generation cellular concepts rely on the processing of large quantities of radio-frequency (RF) samples. This includes Radio Access Networks (RAN) connecting the cellular front-end based on software defined radios (SDRs) and a framework for the AI processing of spectrum-related data. The RF data collected by the dense RAN radio units and spectrum sensors may need to be jointly processed for intelligent decision making. Moving large amounts of data to AI agents may result in significant bandwidth and latency costs. We propose a deep learned compression (DLC) model, HQARF, based on learned vector quantization (VQ), to compress the complex-valued samples of RF signals comprised of 6 modulation classes. We are assessing the effects of HQARF on the performance of an AI model trained to infer the modulation class of the RF signal. Compression of narrow-band RF samples for the training and off-the-site inference will allow for an efficient use of the bandwidth and storage for non-real-time analytics, and for a decreased delay in real-time applications. While exploring the effectiveness of the HQARF signal reconstructions in modulation classification tasks, we highlight the DLC optimization space and some open problems related to the training of the VQ embedded in HQARF.
This work proposes to augment the lifting steps of the conventional wavelet transform with additional neural network assisted lifting steps. These additional steps reduce residual redundancy (notably aliasing information) amongst the wavelet subbands, and also improve the visual quality of reconstructed images at reduced resolutions. The proposed approach involves two steps, a high-to-low step followed by a low-to-high step. The high-to-low step suppresses aliasing in the low-pass band by using the detail bands at the same resolution, while the low-to-high step aims to further remove redundancy from detail bands, so as to achieve higher energy compaction. The proposed two lifting steps are trained in an end-to-end fashion; we employ a backward annealing approach to overcome the non-differentiability of the quantization and cost functions during back-propagation. Importantly, the networks employed in this paper are compact and with limited non-linearities, allowing a fully scalable system; one pair of trained network parameters are applied for all levels of decomposition and for all bit-rates of interest. By employing the proposed approach within the JPEG 2000 image coding standard, our method can achieve up to 17.4% average BD bit-rate saving over a wide range of bit-rates, while retaining quality and resolution scalability features of JPEG 2000.
Deploying Large Language Models (LLMs) on edge or mobile devices offers significant benefits, such as enhanced data privacy and real-time processing capabilities. However, it also faces critical challenges due to the substantial memory requirement of LLMs. Quantization is an effective way of reducing the model size while maintaining good performance. However, even after quantization, LLMs may still be too big to fit entirely into the limited memory of edge or mobile devices and have to be partially loaded from the storage to complete the inference. In this case, the I/O latency of model loading becomes the bottleneck of the LLM inference latency. In this work, we take a preliminary step of studying applying data compression techniques to reduce data movement and thus speed up the inference of quantized LLM on memory-constrained devices. In particular, we discussed the compressibility of quantized LLMs, the trade-off between the compressibility and performance of quantized LLMs, and opportunities to optimize both of them jointly.
Large language models (LLMs) have proven to be very superior to conventional methods in various tasks. However, their expensive computations and high memory requirements are prohibitive for deployment. Model quantization is an effective method for reducing this overhead. The problem is that in most previous works, the quantized model was calibrated using few samples from the training data, which might affect the generalization of the quantized LLMs to unknown cases and tasks. Hence in this work, we explore an important question: Can we design a data-independent quantization method for LLMs to guarantee its generalization performance? In this work, we propose EasyQuant, a training-free and data-independent weight-only quantization algorithm for LLMs. Our observation indicates that two factors: outliers in the weight and quantization ranges, are essential for reducing the quantization error. Therefore, in EasyQuant, we leave the outliers (less than 1%) unchanged and optimize the quantization range to reduce the reconstruction error. With these methods, we surprisingly find that EasyQuant achieves comparable performance to the original model. Since EasyQuant does not depend on any training data, the generalization performance of quantized LLMs is safely guaranteed. Moreover, EasyQuant can be implemented in parallel so that the quantized model could be attained in a few minutes even for LLMs over 100B. To our best knowledge, we are the first work that achieves almost lossless quantization performance for LLMs under a data-independent setting and our algorithm runs over 10 times faster than the data-dependent methods.
Multivariate time series anomaly detection is a crucial problem in many industrial and research applications. Timely detection of anomalies allows, for instance, to prevent defects in manufacturing processes and failures in cyberphysical systems. Deep learning methods are preferred among others for their accuracy and robustness for the analysis of complex multivariate data. However, a key aspect is being able to extract predictions in a timely manner, to accommodate real-time requirements in different applications. In the case of deep learning models, model reduction is extremely important to achieve optimal results in real-time systems with limited time and memory constraints. In this paper, we address this issue by proposing a novel compression method for deep autoencoders that involves three key factors. First, pruning reduces the number of weights, while preventing catastrophic drops in accuracy by means of a fast search process that identifies high sparsity levels. Second, linear and non-linear quantization reduces model complexity by reducing the number of bits for every single weight. The combined contribution of these three aspects allow the model size to be reduced, by removing a subset of the weights (pruning), and decreasing their bit-width (quantization). As a result, the compressed model is faster and easier to adopt in highly constrained hardware environments. Experiments performed on popular multivariate anomaly detection benchmarks, show that our method is capable of achieving significant model compression ratio (between 80% and 95%) without a significant reduction in the anomaly detection performance.
Low precision training can significantly reduce the computational overhead of training deep neural networks (DNNs). Though many such techniques exist, cyclic precision training (CPT), which dynamically adjusts precision throughout training according to a cyclic schedule, achieves particularly impressive improvements in training efficiency, while actually improving DNN performance. Existing CPT implementations take common learning rate schedules (e.g., cyclical cosine schedules) and use them for low precision training without adequate comparisons to alternative scheduling options. We define a diverse suite of CPT schedules and analyze their performance across a variety of DNN training regimes, some of which are unexplored in the low precision training literature (e.g., node classification with graph neural networks). From these experiments, we discover alternative CPT schedules that offer further improvements in training efficiency and model performance, as well as derive a set of best practices for choosing CPT schedules. Going further, we find that a correlation exists between model performance and training cost, and that changing the underlying CPT schedule can control the tradeoff between these two variables. To explain the direct correlation between model performance and training cost, we draw a connection between quantized training and critical learning periods, suggesting that aggressive quantization is a form of learning impairment that can permanently damage model performance.
Generative modeling of complex behaviors from labeled datasets has been a longstanding problem in decision making. Unlike language or image generation, decision making requires modeling actions - continuous-valued vectors that are multimodal in their distribution, potentially drawn from uncurated sources, where generation errors can compound in sequential prediction. A recent class of models called Behavior Transformers (BeT) addresses this by discretizing actions using k-means clustering to capture different modes. However, k-means struggles to scale for high-dimensional action spaces or long sequences, and lacks gradient information, and thus BeT suffers in modeling long-range actions. In this work, we present Vector-Quantized Behavior Transformer (VQ-BeT), a versatile model for behavior generation that handles multimodal action prediction, conditional generation, and partial observations. VQ-BeT augments BeT by tokenizing continuous actions with a hierarchical vector quantization module. Across seven environments including simulated manipulation, autonomous driving, and robotics, VQ-BeT improves on state-of-the-art models such as BeT and Diffusion Policies. Importantly, we demonstrate VQ-BeT's improved ability to capture behavior modes while accelerating inference speed 5x over Diffusion Policies. Videos and code can be found https://sjlee.cc/vq-bet
While recent large-scale text-to-speech (TTS) models have achieved significant progress, they still fall short in speech quality, similarity, and prosody. Considering speech intricately encompasses various attributes (e.g., content, prosody, timbre, and acoustic details) that pose significant challenges for generation, a natural idea is to factorize speech into individual subspaces representing different attributes and generate them individually. Motivated by it, we propose NaturalSpeech 3, a TTS system with novel factorized diffusion models to generate natural speech in a zero-shot way. Specifically, 1) we design a neural codec with factorized vector quantization (FVQ) to disentangle speech waveform into subspaces of content, prosody, timbre, and acoustic details; 2) we propose a factorized diffusion model to generate attributes in each subspace following its corresponding prompt. With this factorization design, NaturalSpeech 3 can effectively and efficiently model intricate speech with disentangled subspaces in a divide-and-conquer way. Experiments show that NaturalSpeech 3 outperforms the state-of-the-art TTS systems on quality, similarity, prosody, and intelligibility, and achieves on-par quality with human recordings. Furthermore, we achieve better performance by scaling to 1B parameters and 200K hours of training data.
The pursuit of optimizing cancer therapies is significantly advanced by the accurate prediction of drug synergy. Traditional methods, such as clinical trials, are reliable yet encumbered by extensive time and financial demands. The emergence of high-throughput screening and computational innovations has heralded a shift towards more efficient methodologies for exploring drug interactions. In this study, we present VQSynergy, a novel framework that employs the Vector Quantization (VQ) mechanism, integrated with gated residuals and a tailored attention mechanism, to enhance the precision and generalizability of drug synergy predictions. Our findings demonstrate that VQSynergy surpasses existing models in terms of robustness, particularly under Gaussian noise conditions, highlighting its superior performance and utility in the complex and often noisy domain of drug synergy research. This study underscores the potential of VQSynergy in revolutionizing the field through its advanced predictive capabilities, thereby contributing to the optimization of cancer treatment strategies.
BasedAI is a distributed network of machines which introduces decentralized infrastructure capable of integrating Fully Homomorphic Encryption (FHE) with any large language model (LLM) connected to its network. The proposed framework embeds a default mechanism, called "Cerberus Squeezing", into the mining process which enables the transformation of a standard LLMs into encrypted zero-knowledge LLMs, or "ZK-LLMs", leveraging insights from generative adversarial networks for data privacy. This novel quantization mechanism empowers BasedAI miners to process and respond to prompts derived from User interaction with LLMs without the need for decrypting ei- ther the queries or their corresponding responses. The introduction of Cerberus Squeezing significantly improves performance degradation caused by quantized functions in current FHE-compliant computing environments by proactively optimizing calls between users, miners, and validators.
Field-Programmable Gate Array (FPGA) accelerators have proven successful in handling latency- and resource-critical deep neural network (DNN) inference tasks. Among the most computationally intensive operations in a neural network (NN) is the dot product between the feature and weight vectors. Thus, some previous FPGA acceleration works have proposed mapping neurons with quantized inputs and outputs directly to lookup tables (LUTs) for hardware implementation. In these works, the boundaries of the neurons coincide with the boundaries of the LUTs. We propose relaxing these boundaries and mapping entire sub-networks to a single LUT. As the sub-networks are absorbed within the LUT, the NN topology and precision within a partition do not affect the size of the lookup tables generated. Therefore, we utilize fully connected layers with floating-point precision inside each partition, which benefit from being universal function approximators, but with rigid sparsity and quantization enforced between partitions, where the NN topology becomes exposed to the circuit topology. Although cheap to implement, this approach can lead to very deep NNs, and so to tackle challenges like vanishing gradients, we also introduce skip connections inside the partitions. The resulting methodology can be seen as training DNNs with a specific FPGA hardware-inspired sparsity pattern that allows them to be mapped to much shallower circuit-level networks, thereby significantly improving latency. We validate our proposed method on a known latency-critical task, jet substructure tagging, and on the classical computer vision task, digit classification using MNIST. Our approach allows for greater function expressivity within the LUTs compared to existing work, leading to up to $4.3\times$ lower latency NNs for the same accuracy.
Modern deep neural networks (DNNs) are extremely powerful; however, this comes at the price of increased depth and having more parameters per layer, making their training and inference more computationally challenging. In an attempt to address this key limitation, efforts have been devoted to the compression (e.g., sparsification and/or quantization) of these large-scale machine learning models, so that they can be deployed on low-power IoT devices. In this paper, building upon recent advances in neural tangent kernel (NTK) and random matrix theory (RMT), we provide a novel compression approach to wide and fully-connected \emph{deep} neural nets. Specifically, we demonstrate that in the high-dimensional regime where the number of data points $n$ and their dimension $p$ are both large, and under a Gaussian mixture model for the data, there exists \emph{asymptotic spectral equivalence} between the NTK matrices for a large family of DNN models. This theoretical result enables "lossless" compression of a given DNN to be performed, in the sense that the compressed network yields asymptotically the same NTK as the original (dense and unquantized) network, with its weights and activations taking values \emph{only} in $\{ 0, \pm 1 \}$ up to a scaling. Experiments on both synthetic and real-world data are conducted to support the advantages of the proposed compression scheme, with code available at \url{https://github.com/Model-Compression/Lossless_Compression}.
OOD detection has become more pertinent with advances in network design and increased task complexity. Identifying which parts of the data a given network is misclassifying has become as valuable as the network's overall performance. We can compress the model with quantization, but it suffers minor performance loss. The loss of performance further necessitates the need to derive the confidence estimate of the network's predictions. In line with this thinking, we introduce an Uncertainty Quantification(UQ) technique to quantify the uncertainty in the predictions from a pre-trained vision model. We subsequently leverage this information to extract valuable predictions while ignoring the non-confident predictions. We observe that our technique saves up to 80% of ignored samples from being misclassified. The code for the same is available here.
Large language models (LLMs) excel in natural language processing but demand intensive computation. To mitigate this, various quantization methods have been explored, yet they compromise LLM performance. This paper unveils a previously overlooked type of outliers in LLMs. Such outliers are found to allocate most of the attention scores on initial tokens of input, termed as pivot tokens, which are crucial to the performance of quantized LLMs. Given that, we propose IntactKV to generate the KV cache of pivot tokens losslessly from the full-precision model. The approach is simple and easy to combine with existing quantization solutions with no extra inference overhead. Besides, IntactKV can be calibrated as additional LLM parameters to boost the quantized LLMs further with minimal training costs. Mathematical analysis also proves that IntactKV effectively reduces the upper bound of quantization error. Empirical results show that IntactKV brings consistent improvement over various quantization methods across different LLMs and downstream tasks, leading to the new state-of-the-art for LLM quantization. The codes are available at https://github.com/ruikangliu/IntactKV.
Recent breakthroughs in Large-scale language models (LLMs) have demonstrated impressive performance on various tasks. The immense sizes of LLMs have led to very high resource demand and cost for running the models. Though the models are largely served using uniform high-caliber GPUs nowadays, utilizing a heterogeneous cluster with a mix of available high- and low-capacity GPUs can potentially substantially reduce the serving cost. There is a lack of designs to support efficient LLM serving using a heterogeneous cluster, while the current solutions focus on model partition and uniform compression among homogeneous devices. This paper proposes LLM-PQ, a system that advocates adaptive model quantization and phase-aware partition to improve LLM serving efficiency on heterogeneous GPU clusters. We carefully decide on mixed-precision model quantization together with phase-aware model partition and micro-batch sizing in distributed LLM serving with an efficient algorithm, to greatly enhance inference throughput while fulfilling user-specified model quality targets. Extensive experiments on production inference workloads in 11 different clusters demonstrate that LLM-PQ achieves up to 2.88x (2.26x on average) throughput improvement in inference, showing great advantages over state-of-the-art works.
Deep neural networks (DNNs) have achieved great breakthroughs in many fields such as image classification and natural language processing. However, the execution of DNNs needs to conduct massive numbers of multiply-accumulate (MAC) operations on hardware and thus incurs a large power consumption. To address this challenge, we propose a novel digital MAC design based on encoding. In this new design, the multipliers are replaced by simple logic gates to represent the results with a wide bit representation. The outputs of the new multipliers are added by bit-wise weighted accumulation and the accumulation results are compatible with existing computing platforms accelerating neural networks. Since the multiplication function is replaced by a simple logic representation, the critical paths in the resulting circuits become much shorter. Correspondingly, pipelining stages and intermediate registers used to store partial sums in the MAC array can be reduced, leading to a significantly smaller area as well as better power efficiency. The proposed design has been synthesized and verified by ResNet18- Cifar10, ResNet20-Cifar100, ResNet50-ImageNet, MobileNetV2-Cifar10, MobileNetV2-Cifar100, and EfficientNetB0-ImageNet. The experimental results confirmed the reduction of circuit area by up to 48.79% and the reduction of power consumption of executing DNNs by up to 64.41%, while the accuracy of the neural networks can still be well maintained. The open source code of this work can be found on GitHub with link https://github.com/Bo-Liu-TUM/EncodingNet/.
3D point cloud has been widely used in many mobile application scenarios, including autonomous driving and 3D sensing on mobile devices. However, existing 3D point cloud models tend to be large and cumbersome, making them hard to deploy on edged devices due to their high memory requirements and non-real-time latency. There has been a lack of research on how to compress 3D point cloud models into lightweight models. In this paper, we propose a method called T3DNet (Tiny 3D Network with augmEntation and disTillation) to address this issue. We find that the tiny model after network augmentation is much easier for a teacher to distill. Instead of gradually reducing the parameters through techniques such as pruning or quantization, we pre-define a tiny model and improve its performance through auxiliary supervision from augmented networks and the original model. We evaluate our method on several public datasets, including ModelNet40, ShapeNet, and ScanObjectNN. Our method can achieve high compression rates without significant accuracy sacrifice, achieving state-of-the-art performances on three datasets against existing methods. Amazingly, our T3DNet is 58 times smaller and 54 times faster than the original model yet with only 1.4% accuracy descent on the ModelNet40 dataset.
Achieving successful variable bitrate compression with computationally simple algorithms from a single end-to-end learned image or video compression model remains a challenge. Many approaches have been proposed, including conditional auto-encoders, channel-adaptive gains for the latent tensor or uniformly quantizing all elements of the latent tensor. This paper follows the traditional approach to vary a single quantization step size to perform uniform quantization of all latent tensor elements. However, three modifications are proposed to improve the variable rate compression performance. First, multi objective optimization is used for (post) training. Second, a quantization-reconstruction offset is introduced into the quantization operation. Third, variable rate quantization is used also for the hyper latent. All these modifications can be made on a pre-trained single-rate compression model by performing post training. The algorithms are implemented into three well-known image compression models and the achieved variable rate compression results indicate negligible or minimal compression performance loss compared to training multiple models. (Codes will be shared at https://github.com/InterDigitalInc/CompressAI)