Quantization is a key method for deploying deep neural networks on edge devices with limited memory and computation resources. Recent improvements in Post-Training Quantization (PTQ) methods were achieved by an additional local optimization process for learning the weight quantization rounding policy. However, a gap exists when employing network-wise optimization with small representative datasets. In this paper, we propose a new method for enhanced PTQ (EPTQ) that employs a network-wise quantization optimization process, which benefits from considering cross-layer dependencies during optimization. EPTQ enables network-wise optimization with a small representative dataset using a novel sample-layer attention score based on a label-free Hessian matrix upper bound. The label-free approach makes our method suitable for the PTQ scheme. We give a theoretical analysis for the said bound and use it to construct a knowledge distillation loss that guides the optimization to focus on the more sensitive layers and samples. In addition, we leverage the Hessian upper bound to improve the weight quantization parameters selection by focusing on the more sensitive elements in the weight tensors. Empirically, by employing EPTQ we achieve state-of-the-art results on various models, tasks, and datasets, including ImageNet classification, COCO object detection, and Pascal-VOC for semantic segmentation.
We present Clinical Prediction with Large Language Models (CPLLM), a method that involves fine-tuning a pre-trained Large Language Model (LLM) for clinical disease and readmission prediction. We utilized quantization and fine-tuned the LLM using prompts. For diagnosis prediction, we predict whether patients will be diagnosed with a target disease during their next visit or in the subsequent diagnosis, leveraging their historical diagnosis records. We compared our results to various baselines, including RETAIN, and Med-BERT, the current state-of-the-art model for disease prediction using temporal structured EHR data. In addition, We also evaluated CPLLM for patient hospital readmission prediction and compared our method's performance with benchmark baselines. Our experiments have shown that our proposed method, CPLLM, surpasses all the tested models in terms of PR-AUC and ROC-AUC metrics, showing state-of-the-art results for diagnosis prediction and patient hospital readmission prediction. Such a method can be easily implemented and integrated into the clinical process to help care providers estimate the next steps of patients
Direction of arrival (DOA) estimation employing low-resolution analog-to-digital convertors (ADCs) has emerged as a challenging and intriguing problem, particularly with the rise in popularity of large-scale arrays. The substantial quantization distortion complicates the extraction of signal and noise subspaces from the quantized data. To address this issue, this paper introduces a novel approach that leverages the Transformer model to aid the subspace estimation. In this model, multiple snapshots are processed in parallel, enabling the capture of global correlations that span them. The learned subspace empowers us to construct the MUSIC spectrum and perform gridless DOA estimation using a neural network-based peak finder. Additionally, the acquired subspace encodes the vital information of model order, allowing us to determine the exact number of sources. These integrated components form a unified algorithmic framework referred to as TransMUSIC. Numerical results demonstrate the superiority of the TransMUSIC algorithm, even when dealing with one-bit quantized data. The results highlight the potential of Transformer-based techniques in DOA estimation.
The task of preserving privacy while ensuring efficient communication is a fundamental challenge in federated learning. In this work, we tackle this challenge in the trusted aggregator model, and propose a solution that achieves both objectives simultaneously. We show that employing a quantization scheme based on subtractive dithering at the clients can effectively replicate the normal noise addition process at the aggregator. This implies that we can guarantee the same level of differential privacy against other clients while substantially reducing the amount of communication required, as opposed to transmitting full precision gradients and using central noise addition. We also experimentally demonstrate that the accuracy of our proposed approach matches that of the full precision gradient method.
JPEG remains one of the most widespread lossy image coding methods. However, the non-differentiable nature of JPEG restricts the application in deep learning pipelines. Several differentiable approximations of JPEG have recently been proposed to address this issue. This paper conducts a comprehensive review of existing diff. JPEG approaches and identifies critical details that have been missed by previous methods. To this end, we propose a novel diff. JPEG approach, overcoming previous limitations. Our approach is differentiable w.r.t. the input image, the JPEG quality, the quantization tables, and the color conversion parameters. We evaluate the forward and backward performance of our diff. JPEG approach against existing methods. Additionally, extensive ablations are performed to evaluate crucial design choices. Our proposed diff. JPEG resembles the (non-diff.) reference implementation best, significantly surpassing the recent-best diff. approach by $3.47$dB (PSNR) on average. For strong compression rates, we can even improve PSNR by $9.51$dB. Strong adversarial attack results are yielded by our diff. JPEG, demonstrating the effective gradient approximation. Our code is available at https://github.com/necla-ml/Diff-JPEG.
Summary: Raw nanopore signals can be analyzed while they are being generated, a process known as real-time analysis. Real-time analysis of raw signals is essential to utilize the unique features that nanopore sequencing provides, enabling the early stopping of the sequencing of a read or the entire sequencing run based on the analysis. The state-of-the-art mechanism, RawHash, offers the first hash-based efficient and accurate similarity identification between raw signals and a reference genome by quickly matching their hash values. In this work, we introduce RawHash2, which provides major improvements over RawHash, including a more sensitive quantization and chaining implementation, weighted mapping decisions, frequency filters to reduce ambiguous seed hits, minimizers for hash-based sketching, and support for the R10.4 flow cell version and various data formats such as POD5 and SLOW5. Compared to RawHash, RawHash2 provides better F1 accuracy (on average by 10.57% and up to 20.25%) and better throughput (on average by 4.0x and up to 9.9x) than RawHash. Availability and Implementation: RawHash2 is available at https://github.com/CMU-SAFARI/RawHash. We also provide the scripts to fully reproduce our results on our GitHub page.
This work introduces a novel class of channel estimators tailored for coarse quantization systems. The proposed estimators are founded on conditionally Gaussian latent generative models, specifically Gaussian mixture models (GMMs), mixture of factor analyzers (MFAs), and variational autoencoders (VAEs). These models effectively learn the unknown channel distribution inherent in radio propagation scenarios, providing valuable prior information. Conditioning on the latent variable of these generative models yields a locally Gaussian channel distribution, thus enabling the application of the well-known Bussgang decomposition. By exploiting the resulting conditional Bussgang decomposition, we derive parameterized linear minimum mean square error (MMSE) estimators for the considered generative latent variable models. In this context, we explore leveraging model-based structural features to reduce memory and complexity overhead associated with the proposed estimators. Furthermore, we devise necessary training adaptations, enabling direct learning of the generative models from quantized pilot observations without requiring ground-truth channel samples during the training phase. Through extensive simulations, we demonstrate the superiority of our introduced estimators over existing state-of-the-art methods for coarsely quantized systems, as evidenced by significant improvements in mean square error (MSE) and achievable rate metrics.
In the context of kernel machines, polynomial and Fourier features are commonly used to provide a nonlinear extension to linear models by mapping the data to a higher-dimensional space. Unless one considers the dual formulation of the learning problem, which renders exact large-scale learning unfeasible, the exponential increase of model parameters in the dimensionality of the data caused by their tensor-product structure prohibits to tackle high-dimensional problems. One of the possible approaches to circumvent this exponential scaling is to exploit the tensor structure present in the features by constraining the model weights to be an underparametrized tensor network. In this paper we quantize, i.e. further tensorize, polynomial and Fourier features. Based on this feature quantization we propose to quantize the associated model weights, yielding quantized models. We show that, for the same number of model parameters, the resulting quantized models have a higher bound on the VC-dimension as opposed to their non-quantized counterparts, at no additional computational cost while learning from identical features. We verify experimentally how this additional tensorization regularizes the learning problem by prioritizing the most salient features in the data and how it provides models with increased generalization capabilities. We finally benchmark our approach on large regression task, achieving state-of-the-art results on a laptop computer.
Large Language Models (LLMs) have demonstrated exceptional proficiency in language-related tasks, but their deployment poses significant challenges due to substantial memory and storage requirements. Weight-only quantization has emerged as a promising solution, significantly reducing memory and storage needs without sacrificing too much performance. In this study, we introduce SignRound, a method that leverages signed gradient descent (SignSGD) to optimize rounding values and weight clipping in just 200 steps. SignRound integrates the advantages of Quantization-Aware Training (QAT) and Post-Training Quantization (PTQ), delivering exceptional results across 2 to 4 bits while minimizing tuning costs and avoiding additional inference overhead. For example, SignRound achieved absolute average accuracy improvements ranging from 6.91% to 33.22% at 2bits, as measured by the average zero-shot accuracy across 11 tasks. It also demonstrates strong generalization in recent models, achieving near-lossless 4-bit quantization in most scenarios. The source code is publicly available at https://github.com/intel/auto-round.
The quantization problem aims to find the best possible approximation of probability measures on ${\mathbb{R}}^d$ using finite, discrete measures. The Wasserstein distance is a typical choice to measure the quality of the approximation. This contribution investigates the properties and robustness of the entropy-regularized quantization problem, which relaxes the standard quantization problem. The proposed approximation technique naturally adopts the softmin function, which is well known for its robustness in terms of theoretical and practicability standpoints. Moreover, we use the entropy-regularized Wasserstein distance to evaluate the quality of the soft quantization problem's approximation, and we implement a stochastic gradient approach to achieve the optimal solutions. The control parameter in our proposed method allows for the adjustment of the optimization problem's difficulty level, providing significant advantages when dealing with exceptionally challenging problems of interest. As well, this contribution empirically illustrates the performance of the method in various expositions.
Vision transformer (ViT) and its variants have swept through visual learning leaderboards and offer state-of-the-art accuracy in tasks such as image classification, object detection, and semantic segmentation by attending to different parts of the visual input and capturing long-range spatial dependencies. However, these models are large and computation-heavy. For instance, the recently proposed ViT-B model has 86M parameters making it impractical for deployment on resource-constrained devices. As a result, their deployment on mobile and edge scenarios is limited. In our work, we aim to take a step toward bringing vision transformers to the edge by utilizing popular model compression techniques such as distillation, pruning, and quantization. Our chosen application environment is an unmanned aerial vehicle (UAV) that is battery-powered and memory-constrained, carrying a single-board computer on the scale of an NVIDIA Jetson Nano with 4GB of RAM. On the other hand, the UAV requires high accuracy close to that of state-of-the-art ViTs to ensure safe object avoidance in autonomous navigation, or correct localization of humans in search-and-rescue. Inference latency should also be minimized given the application requirements. Hence, our target is to enable rapid inference of a vision transformer on an NVIDIA Jetson Nano (4GB) with minimal accuracy loss. This allows us to deploy ViTs on resource-constrained devices, opening up new possibilities in surveillance, environmental monitoring, etc. Our implementation is made available at https://github.com/chensy7/efficient-vit.
Automatic few-shot font generation (AFFG), aiming at generating new fonts with only a few glyph references, reduces the labor cost of manually designing fonts. However, the traditional AFFG paradigm of style-content disentanglement cannot capture the diverse local details of different fonts. So, many component-based approaches are proposed to tackle this problem. The issue with component-based approaches is that they usually require special pre-defined glyph components, e.g., strokes and radicals, which is infeasible for AFFG of different languages. In this paper, we present a novel font generation approach by aggregating styles from character similarity-guided global features and stylized component-level representations. We calculate the similarity scores of the target character and the referenced samples by measuring the distance along the corresponding channels from the content features, and assigning them as the weights for aggregating the global style features. To better capture the local styles, a cross-attention-based style transfer module is adopted to transfer the styles of reference glyphs to the components, where the components are self-learned discrete latent codes through vector quantization without manual definition. With these designs, our AFFG method could obtain a complete set of component-level style representations, and also control the global glyph characteristics. The experimental results reflect the effectiveness and generalization of the proposed method on different linguistic scripts, and also show its superiority when compared with other state-of-the-art methods. The source code can be found at https://github.com/awei669/VQ-Font.
With recent rapid growth of large language models (LLMs), discrete speech tokenization has played an important role for injecting speech into LLMs. However, this discretization gives rise to a loss of information, consequently impairing overall performance. To improve the performance of these discrete speech tokens, we present RepCodec, a novel speech representation codec for semantic speech tokenization. In contrast to audio codecs which reconstruct the raw audio, RepCodec learns a vector quantization codebook through reconstructing speech representations from speech encoders like HuBERT or data2vec. Together, the speech encoder, the codec encoder and the vector quantization codebook form a pipeline for converting speech waveforms into semantic tokens. The extensive experiments illustrate that RepCodec, by virtue of its enhanced information retention capacity, significantly outperforms the widely used k-means clustering approach in both speech understanding and generation. Furthermore, this superiority extends across various speech encoders and languages, affirming the robustness of RepCodec. We believe our method can facilitate large language modeling research on speech processing.
Category information plays a crucial role in enhancing the quality and personalization of recommender systems. Nevertheless, the availability of item category information is not consistently present, particularly in the context of ID-based recommendations. In this work, we propose a novel approach to automatically learn and generate entity (i.e., user or item) category trees for ID-based recommendation. Specifically, we devise a differentiable vector quantization framework for automatic category tree generation, namely CAGE, which enables the simultaneous learning and refinement of categorical code representations and entity embeddings in an end-to-end manner, starting from the randomly initialized states. With its high adaptability, CAGE can be easily integrated into both sequential and non-sequential recommender systems. We validate the effectiveness of CAGE on various recommendation tasks including list completion, collaborative filtering, and click-through rate prediction, across different recommendation models. We release the code and data for others to reproduce the reported results.
Current speech large language models build upon discrete speech representations, which can be categorized into semantic tokens and acoustic tokens. However, existing speech tokens are not specifically designed for speech language modeling. To assess the suitability of speech tokens for building speech language models, we established the first benchmark, SLMTokBench. Our results indicate that neither semantic nor acoustic tokens are ideal for this purpose. Therefore, we propose SpeechTokenizer, a unified speech tokenizer for speech large language models. SpeechTokenizer adopts the Encoder-Decoder architecture with residual vector quantization (RVQ). Unifying semantic and acoustic tokens, SpeechTokenizer disentangles different aspects of speech information hierarchically across different RVQ layers. Furthermore, We construct a Unified Speech Language Model (USLM) leveraging SpeechTokenizer. Experiments show that SpeechTokenizer performs comparably to EnCodec in speech reconstruction and demonstrates strong performance on the SLMTokBench benchmark. Also, USLM outperforms VALL-E in zero-shot Text-to-Speech tasks. Code and models are available at https://github.com/ZhangXInFD/SpeechTokenizer/.
Generative retrieval (GR) directly predicts the identifiers of relevant documents (i.e., docids) based on a parametric model. It has achieved solid performance on many ad-hoc retrieval tasks. So far, these tasks have assumed a static document collection. In many practical scenarios, however, document collections are dynamic, where new documents are continuously added to the corpus. The ability to incrementally index new documents while preserving the ability to answer queries with both previously and newly indexed relevant documents is vital to applying GR models. In this paper, we address this practical continual learning problem for GR. We put forward a novel Continual-LEarner for generatiVE Retrieval (CLEVER) model and make two major contributions to continual learning for GR: (i) To encode new documents into docids with low computational cost, we present Incremental Product Quantization, which updates a partial quantization codebook according to two adaptive thresholds; and (ii) To memorize new documents for querying without forgetting previous knowledge, we propose a memory-augmented learning mechanism, to form meaningful connections between old and new documents. Empirical results demonstrate the effectiveness and efficiency of the proposed model.
Graph Neural Network (GNN) training and inference involve significant challenges of scalability with respect to both model sizes and number of layers, resulting in degradation of efficiency and accuracy for large and deep GNNs. We present an end-to-end solution that aims to address these challenges for efficient GNNs in resource constrained environments while avoiding the oversmoothing problem in deep GNNs. We introduce a quantization based approach for all stages of GNNs, from message passing in training to node classification, compressing the model and enabling efficient processing. The proposed GNN quantizer learns quantization ranges and reduces the model size with comparable accuracy even under low-bit quantization. To scale with the number of layers, we devise a message propagation mechanism in training that controls layer-wise changes of similarities between neighboring nodes. This objective is incorporated into a Lagrangian function with constraints and a differential multiplier method is utilized to iteratively find optimal embeddings. This mitigates oversmoothing and suppresses the quantization error to a bound. Significant improvements are demonstrated over state-of-the-art quantization methods and deep GNN approaches in both full-precision and quantized models. The proposed quantizer demonstrates superior performance in INT2 configurations across all stages of GNN, achieving a notable level of accuracy. In contrast, existing quantization approaches fail to generate satisfactory accuracy levels. Finally, the inference with INT2 and INT4 representations exhibits a speedup of 5.11 $\times$ and 4.70 $\times$ compared to full precision counterparts, respectively.
Deep Neural Networks (DNNs) have been a large driver for AI breakthroughs in recent years. However, these models have been getting increasingly large as they become more accurate and safe. This means that their training becomes increasingly costly and time-consuming and typically yields a single model to fit all targets. Various techniques have been proposed in the literature to mitigate this, including pruning, sparsification, or quantization of model weights and updates. While achieving high compression rates, they often incur significant computational overheads at training or lead to non-negligible accuracy penalty. Alternatively, factorization methods have been leveraged for low-rank compression of DNNs. Similarly, such techniques (e.g., SVD) frequently rely on heavy iterative decompositions of layers and are potentially sub-optimal for non-linear models, such as DNNs. We take a further step in designing efficient low-rank models and propose Maestro, a framework for trainable low-rank layers. Instead of iteratively applying a priori decompositions, the low-rank structure is baked into the training process through LoD, a low-rank ordered decomposition. Not only is this the first time importance ordering via sampling is applied on the decomposed DNN structure, but it also allows selecting ranks at a layer granularity. Our theoretical analysis demonstrates that in special cases LoD recovers the SVD decomposition and PCA. Applied to DNNs, Maestro enables the extraction of lower footprint models that preserve performance. Simultaneously, it enables the graceful trade-off between accuracy-latency for deployment to even more constrained devices without retraining.
In an age dominated by resource-intensive foundation models, the ability to efficiently adapt to downstream tasks is crucial. Visual Prompting (VP), drawing inspiration from the prompting techniques employed in Large Language Models (LLMs), has emerged as a pivotal method for transfer learning in the realm of computer vision. As the importance of efficiency continues to rise, research into model compression has become indispensable in alleviating the computational burdens associated with training and deploying over-parameterized neural networks. A primary objective in model compression is to develop sparse and/or quantized models capable of matching or even surpassing the performance of their over-parameterized, full-precision counterparts. Although previous studies have explored the effects of model compression on transfer learning, its impact on visual prompting-based transfer remains unclear. This study aims to bridge this gap, shedding light on the fact that model compression detrimentally impacts the performance of visual prompting-based transfer, particularly evident in scenarios with low data volume. Furthermore, our findings underscore the adverse influence of sparsity on the calibration of downstream visual-prompted models. However, intriguingly, we also illustrate that such negative effects on calibration are not present when models are compressed via quantization. This empirical investigation underscores the need for a nuanced understanding beyond mere accuracy in sparse and quantized settings, thereby paving the way for further exploration in Visual Prompting techniques tailored for sparse and quantized models.
Few-shot font generation is challenging, as it needs to capture the fine-grained stroke styles from a limited set of reference glyphs, and then transfer to other characters, which are expected to have similar styles. However, due to the diversity and complexity of Chinese font styles, the synthesized glyphs of existing methods usually exhibit visible artifacts, such as missing details and distorted strokes. In this paper, we propose a VQGAN-based framework (i.e., VQ-Font) to enhance glyph fidelity through token prior refinement and structure-aware enhancement. Specifically, we pre-train a VQGAN to encapsulate font token prior within a codebook. Subsequently, VQ-Font refines the synthesized glyphs with the codebook to eliminate the domain gap between synthesized and real-world strokes. Furthermore, our VQ-Font leverages the inherent design of Chinese characters, where structure components such as radicals and character components are combined in specific arrangements, to recalibrate fine-grained styles based on references. This process improves the matching and fusion of styles at the structure level. Both modules collaborate to enhance the fidelity of the generated fonts. Experiments on a collected font dataset show that our VQ-Font outperforms the competing methods both quantitatively and qualitatively, especially in generating challenging styles.