Over the last decade, deep learning has shown great success at performing computer vision tasks, including classification, super-resolution, and style transfer. Now, we apply it to data compression to help build the next generation of multimedia codecs. This thesis provides three primary contributions to this new field of learned compression. First, we present an efficient low-complexity entropy model that dynamically adapts the encoding distribution to a specific input by compressing and transmitting the encoding distribution itself as side information. Secondly, we propose a novel lightweight low-complexity point cloud codec that is highly specialized for classification, attaining significant reductions in bitrate compared to non-specialized codecs. Lastly, we explore how motion within the input domain between consecutive video frames is manifested in the corresponding convolutionally-derived latent space.
Motivation: Burrows-Wheeler Transform (BWT) is a common component in full-text indices. Initially developed for data compression, it is particularly powerful for encoding redundant sequences such as pangenome data. However, BWT construction is resource intensive and hard to be parallelized, and many methods for querying large full-text indices only report exact matches or their simple extensions. These limitations have hampered the biological applications of full-text indices. Results: We developed ropebwt3 for efficient BWT construction and query. Ropebwt3 indexed 320 assembled human genomes in 65 hours and indexed 7.3 terabases of commonly studied bacterial assemblies in 26 days. This was achieved using up to 170 gigabytes of memory at the peak without working disk space. Ropebwt3 can find maximal exact matches and inexact alignments under affine-gap penalties, and can retrieve similar local haplotypes matching a query sequence. It demonstrates the feasibility of full-text indexing at the terabase scale. Availability and implementation: https://github.com/lh3/ropebwt3
In hierarchical cognitive radio networks, edge or cloud servers utilize the data collected by edge devices for modulation classification, which, however, is faced with problems of the computation load, transmission overhead, and data privacy. In this article, an edge learning (EL) based framework jointly mobilizing the edge device and the edge server for intelligent co-inference is proposed to realize the collaborative automatic modulation classification (C-AMC) between them. A spectrum semantic compression neural network is designed for the edge device to compress the collected raw data into a compact semantic embedding that is then sent to the edge server via the wireless channel. On the edge server side, a modulation classification neural network combining the bidirectional long-short term memory and attention structures is elaborated to determine the modulation type from the noisy semantic embedding. The C-AMC framework decently balances the computation resources of both sides while avoiding the high transmission overhead and data privacy leakage. Both the offline and online training procedures of the C-AMC framework are elaborated. The compression strategy of the C-AMC framework is also developed to further facilitate the deployment, especially for the resource-constrained edge device. Simulation results show the superiority of the EL-based C-AMC framework in terms of the classification accuracy, computational complexity, and the data compression rate as well as reveal useful insights paving the practical implementation.
Empirically-determined scaling laws have been broadly successful in predicting the evolution of large machine learning models with training data and number of parameters. As a consequence, they have been useful for optimizing the allocation of limited resources, most notably compute time. In certain applications, storage space is an important constraint, and data format needs to be chosen carefully as a consequence. Computer vision is a prominent example: images are inherently analog, but are always stored in a digital format using a finite number of bits. Given a dataset of digital images, the number of bits $L$ to store each of them can be further reduced using lossy data compression. This, however, can degrade the quality of the model trained on such images, since each example has lower resolution. In order to capture this trade-off and optimize storage of training data, we propose a `storage scaling law' that describes the joint evolution of test error with sample size and number of bits per image. We prove that this law holds within a stylized model for image compression, and verify it empirically on two computer vision tasks, extracting the relevant parameters. We then show that this law can be used to optimize the lossy compression level. At given storage, models trained on optimally compressed images present a significantly smaller test error with respect to models trained on the original data. Finally, we investigate the potential benefits of randomizing the compression level.
To fill the gap of traditional GS compression method, in this paper, we first propose a simple and effective GS data compression anchor called Graph-based GS Compression (GGSC). GGSC is inspired by graph signal processing theory and uses two branches to compress the primitive center and attributes. We split the whole GS sample via KDTree and clip the high-frequency components after the graph Fourier transform. Followed by quantization, G-PCC and adaptive arithmetic coding are used to compress the primitive center and attribute residual matrix to generate the bitrate file. GGSS is the first work to explore traditional GS compression, with advantages that can reveal the GS distortion characteristics corresponding to typical compression operation, such as high-frequency clipping and quantization. Second, based on GGSC, we create a GS Quality Assessment dataset (GSQA) with 120 samples. A subjective experiment is conducted in a laboratory environment to collect subjective scores after rendering GS into Processed Video Sequences (PVS). We analyze the characteristics of different GS distortions based on Mean Opinion Scores (MOS), demonstrating the sensitivity of different attributes distortion to visual quality. The GGSC code and the dataset, including GS samples, MOS, and PVS, are made publicly available at https://github.com/Qi-Yangsjtu/GGSC.
Modern data compression methods are slowly reaching their limits after 80 years of research, millions of papers, and wide range of applications. Yet, the extravagant 6G communication speed requirement raises a major open question for revolutionary new ideas of data compression. We have previously shown all understanding or learning are compression, under reasonable assumptions. Large language models (LLMs) understand data better than ever before. Can they help us to compress data? The LLMs may be seen to approximate the uncomputable Solomonoff induction. Therefore, under this new uncomputable paradigm, we present LMCompress. LMCompress shatters all previous lossless compression algorithms, doubling the lossless compression ratios of JPEG-XL for images, FLAC for audios, and H.264 for videos, and quadrupling the compression ratio of bz2 for texts. The better a large model understands the data, the better LMCompress compresses.
Data is the cornerstone of large language models (LLMs), but not all data is useful for model learning. Carefully selected data can better elicit the capabilities of LLMs with much less computational overhead. Most methods concentrate on evaluating the quality of individual samples in data selection, while the combinatorial effects among samples are neglected. Even if each sample is of perfect quality, their combinations may be suboptimal in teaching LLMs due to their intrinsic homogeneity or contradiction. In this paper, we aim to uncover the underlying relationships between LLM performance and data selection. Inspired by the information compression nature of LLMs, we uncover an ``entropy law'' that connects LLM performance with data compression ratio and first-epoch training loss, which reflect the information redundancy of a dataset and the mastery of inherent knowledge encoded in this dataset, respectively. Through both theoretical deduction and empirical evaluation, we find that model performance is negatively correlated to the compression ratio of training data, which usually yields a lower training loss. Based on the findings of the entropy law, we propose a quite efficient and universal data selection method named \textbf{ZIP} for training LLMs, which aim to prioritize data subsets exhibiting a low compression ratio. Based on a multi-stage algorithm that selects diverse data in a greedy manner, we can obtain a good data subset with satisfactory diversity. Extensive experiments have been conducted to validate the entropy law and the superiority of ZIP across different LLM backbones and alignment stages. We also present an interesting application of entropy law that can detect potential performance risks at the beginning of model training.
Large language models (LLMs) may exhibit unintended or undesirable behaviors. Recent works have concentrated on aligning LLMs to mitigate harmful outputs. Despite these efforts, some anomalies indicate that even a well-conducted alignment process can be easily circumvented, whether intentionally or accidentally. Does alignment fine-tuning yield have robust effects on models, or are its impacts merely superficial? In this work, we make the first exploration of this phenomenon from both theoretical and empirical perspectives. Empirically, we demonstrate the elasticity of post-alignment models, i.e., the tendency to revert to the behavior distribution formed during the pre-training phase upon further fine-tuning. Leveraging compression theory, we formally deduce that fine-tuning disproportionately undermines alignment relative to pre-training, potentially by orders of magnitude. We validate the presence of elasticity through experiments on models of varying types and scales. Specifically, we find that model performance declines rapidly before reverting to the pre-training distribution, after which the rate of decline drops significantly. Furthermore, we further reveal that elasticity positively correlates with the increased model size and the expansion of pre-training data. Our findings underscore the need to address the inherent elasticity of LLMs to mitigate their resistance to alignment.
Although deep learning based models for underwater image enhancement have achieved good performance, they face limitations in both lightweight and effectiveness, which prevents their deployment and application on resource-constrained platforms. Moreover, most existing deep learning based models use data compression to get high-level semantic information in latent space instead of using the original information. Therefore, they require decoder blocks to generate the details of the output. This requires additional computational cost. In this paper, a lightweight network named lightweight selective attention network (LSNet) based on the top-k selective attention and transmission maps mechanism is proposed. The proposed model achieves a PSNR of 97\% with only 7K parameters compared to a similar attention-based model. Extensive experiments show that the proposed LSNet achieves excellent performance in state-of-the-art models with significantly fewer parameters and computational resources. The code is available at https://github.com/FuhengZhou/LSNet}{https://github.com/FuhengZhou/LSNet.
Learning discrete representations with vector quantization (VQ) has emerged as a powerful approach in various generative models. However, most VQ-based models rely on a single, fixed-rate codebook, requiring extensive retraining for new bitrates or efficiency requirements. We introduce Rate-Adaptive Quantization (RAQ), a multi-rate codebook adaptation framework for VQ-based generative models. RAQ applies a data-driven approach to generate variable-rate codebooks from a single baseline VQ model, enabling flexible tradeoffs between compression and reconstruction fidelity. Additionally, we provide a simple clustering-based procedure for pre-trained VQ models, offering an alternative when retraining is infeasible. Our experiments show that RAQ performs effectively across multiple rates, often outperforming conventional fixed-rate VQ baselines. By enabling a single system to seamlessly handle diverse bitrate requirements, RAQ extends the adaptability of VQ-based generative models and broadens their applicability to data compression, reconstruction, and generation tasks.
Inductive biases are crucial in disentangled representation learning for narrowing down an underspecified solution set. In this work, we consider endowing a neural network autoencoder with three select inductive biases from the literature: data compression into a grid-like latent space via quantization, collective independence amongst latents, and minimal functional influence of any latent on how other latents determine data generation. In principle, these inductive biases are deeply complementary: they most directly specify properties of the latent space, encoder, and decoder, respectively. In practice, however, naively combining existing techniques instantiating these inductive biases fails to yield significant benefits. To address this, we propose adaptations to the three techniques that simplify the learning problem, equip key regularization terms with stabilizing invariances, and quash degenerate incentives. The resulting model, Tripod, achieves state-of-the-art results on a suite of four image disentanglement benchmarks. We also verify that Tripod significantly improves upon its naive incarnation and that all three of its "legs" are necessary for best performance.
We propose SketchINR, to advance the representation of vector sketches with implicit neural models. A variable length vector sketch is compressed into a latent space of fixed dimension that implicitly encodes the underlying shape as a function of time and strokes. The learned function predicts the $xy$ point coordinates in a sketch at each time and stroke. Despite its simplicity, SketchINR outperforms existing representations at multiple tasks: (i) Encoding an entire sketch dataset into a fixed size latent vector, SketchINR gives $60\times$ and $10\times$ data compression over raster and vector sketches, respectively. (ii) SketchINR's auto-decoder provides a much higher-fidelity representation than other learned vector sketch representations, and is uniquely able to scale to complex vector sketches such as FS-COCO. (iii) SketchINR supports parallelisation that can decode/render $\sim$$100\times$ faster than other learned vector representations such as SketchRNN. (iv) SketchINR, for the first time, emulates the human ability to reproduce a sketch with varying abstraction in terms of number and complexity of strokes. As a first look at implicit sketches, SketchINR's compact high-fidelity representation will support future work in modelling long and complex sketches.
In Internet of Things (IoT) networks, the amount of data sensed by user devices may be huge, resulting in the serious network congestion. To solve this problem, intelligent data compression is critical. The variational information bottleneck (VIB) approach, combined with machine learning, can be employed to train the encoder and decoder, so that the required transmission data size can be reduced significantly. However, VIB suffers from the computing burden and network insecurity. In this paper, we propose a blockchain-enabled VIB (BVIB) approach to relieve the computing burden while guaranteeing network security. Extensive simulations conducted by Python and C++ demonstrate that BVIB outperforms VIB by 36%, 22% and 57% in terms of time and CPU cycles cost, mutual information, and accuracy under attack, respectively.
Tokenization is a foundational step in natural language processing (NLP) tasks, bridging raw text and language models. Existing tokenization approaches like Byte-Pair Encoding (BPE) originate from the field of data compression, and it has been suggested that the effectiveness of BPE stems from its ability to condense text into a relatively small number of tokens. We test the hypothesis that fewer tokens lead to better downstream performance by introducing PathPiece, a new tokenizer that segments a document's text into the minimum number of tokens for a given vocabulary. Through extensive experimentation we find this hypothesis not to be the case, casting doubt on the understanding of the reasons for effective tokenization. To examine which other factors play a role, we evaluate design decisions across all three phases of tokenization: pre-tokenization, vocabulary construction, and segmentation, offering new insights into the design of effective tokenizers. Specifically, we illustrate the importance of pre-tokenization and the benefits of using BPE to initialize vocabulary construction. We train 64 language models with varying tokenization, ranging in size from 350M to 2.4B parameters, all of which are made publicly available.
Storing and transmitting LiDAR point cloud data is essential for many AV applications, such as training data collection, remote control, cloud services or SLAM. However, due to the sparsity and unordered structure of the data, it is difficult to compress point cloud data to a low volume. Transforming the raw point cloud data into a dense 2D matrix structure is a promising way for applying compression algorithms. We propose a new lossless and calibrated 3D-to-2D transformation which allows compression algorithms to efficiently exploit spatial correlations within the 2D representation. To compress the structured representation, we use common image compression methods and also a self-supervised deep compression approach using a recurrent neural network. We also rearrange the LiDAR's intensity measurements to a dense 2D representation and propose a new metric to evaluate the compression performance of the intensity. Compared to approaches that are based on generic octree point cloud compression or based on raw point cloud data compression, our approach achieves the best quantitative and visual performance. Source code and dataset are available at https://github.com/ika-rwth-aachen/Point-Cloud-Compression.
With the exponential growth in data volume and the emergence of data-intensive applications, particularly in the field of machine learning, concerns related to resource utilization, privacy, and fairness have become paramount. This paper focuses on the textual domain of data and addresses challenges regarding encoding sentences to their optimized representations through the lens of information-theory. In particular, we use empirical estimates of mutual information, using the Donsker-Varadhan definition of Kullback-Leibler divergence. Our approach leverages this estimation to train an information-theoretic sentence embedding, called TexShape, for (task-based) data compression or for filtering out sensitive information, enhancing privacy and fairness. In this study, we employ a benchmark language model for initial text representation, complemented by neural networks for information-theoretic compression and mutual information estimations. Our experiments demonstrate significant advancements in preserving maximal targeted information and minimal sensitive information over adverse compression ratios, in terms of predictive accuracy of downstream models that are trained using the compressed data.
WiFi-enabled Internet-of-Things (IoT) devices are evolving from mere communication devices to sensing instruments, leveraging Channel State Information (CSI) extraction capabilities. Nevertheless, resource-constrained IoT devices and the intricacies of deep neural networks necessitate transmitting CSI to cloud servers for sensing. Although feasible, this leads to considerable communication overhead. In this context, this paper develops a novel Real-time Sensing and Compression Network (RSCNet) which enables sensing with compressed CSI; thereby reducing the communication overheads. RSCNet facilitates optimization across CSI windows composed of a few CSI frames. Once transmitted to cloud servers, it employs Long Short-Term Memory (LSTM) units to harness data from prior windows, thus bolstering both the sensing accuracy and CSI reconstruction. RSCNet adeptly balances the trade-off between CSI compression and sensing precision, thus streamlining real-time cloud-based WiFi sensing with reduced communication costs. Numerical findings demonstrate the gains of RSCNet over the existing benchmarks like SenseFi, showcasing a sensing accuracy of 97.4% with minimal CSI reconstruction error. Numerical results also show a computational analysis of the proposed RSCNet as a function of the number of CSI frames.
Information Bottleneck (IB) is a widely used framework that enables the extraction of information related to a target random variable from a source random variable. In the objective function, IB controls the trade-off between data compression and predictiveness through the Lagrange multiplier $\beta$. Traditionally, to find the trade-off to be learned, IB requires a search for $\beta$ through multiple training cycles, which is computationally expensive. In this study, we introduce Flexible Variational Information Bottleneck (FVIB), an innovative framework for classification task that can obtain optimal models for all values of $\beta$ with single, computationally efficient training. We theoretically demonstrate that across all values of reasonable $\beta$, FVIB can simultaneously maximize an approximation of the objective function for Variational Information Bottleneck (VIB), the conventional IB method. Then we empirically show that FVIB can learn the VIB objective as effectively as VIB. Furthermore, in terms of calibration performance, FVIB outperforms other IB and calibration methods by enabling continuous optimization of $\beta$. Our codes are available at https://github.com/sotakudo/fvib.
Existing methods for evaluating large language models face challenges such as data contamination, sensitivity to prompts, and the high cost of benchmark creation. To address this, we propose a lossless data compression based evaluation approach that tests how models' predictive abilities generalize after their training cutoff. Specifically, we collect comprehensive test data spanning 83 months from 2017 to 2023 and split the data into training and testing periods according to models' training data cutoff. We measure: 1) the compression performance on the testing period as a measure of generalization on unseen data; and 2) the performance gap between the training and testing period as a measure of robustness. Our experiments test 14 representative large language models with various sizes on sources including Wikipedia, news articles, code, arXiv papers, and multi-modal data. We find that the compression rate of many models reduces significantly after their cutoff date, but models such as Mistral and Llama-2 demonstrate a good balance between performance and robustness. Results also suggest that models struggle to generalize on news and code data, but work especially well on arXiv papers. We also find the context size and tokenization implementation have a big impact of on the overall compression performance.
Large Language Models (LLMs) have transformed natural language processing and extended their powerful capabilities to multi-modal domains. As LLMs continue to advance, it is crucial to develop diverse and appropriate metrics for their evaluation. In this paper, we introduce a novel rank-based metric, Diff-eRank, grounded in information theory and geometry principles. Diff-eRank assesses LLMs by analyzing their hidden representations, providing a quantitative measure of how efficiently they eliminate redundant information during training. We demonstrate the applicability of Diff-eRank in both single-modal (e.g., language) and multi-modal settings. For language models, our results show that Diff-eRank increases with model size and correlates well with conventional metrics such as loss and accuracy. In the multi-modal context, we propose an alignment evaluation method based on the eRank, and verify that contemporary multi-modal LLMs exhibit strong alignment performance based on our method. Our code is publicly available at https://github.com/waltonfuture/Diff-eRank.