arXiv subjects

Yu Gong

Publications and source records attributed to Yu Gong.

At least 37 records · Page 2Linked to original sources

arXiv2023

COMCAT: Towards Efficient Compression and Customization of Attention-Based Vision Models

Attention-based vision models, such as Vision Transformer (ViT) and its variants, have shown promising performance in various computer vision tasks. However, these emerging architectures suffer from large model sizes and high computational costs, calling for efficient model compression solutions. To date, pruning ViTs has been well studied, while other compression strategies that have been widely applied in CNN compression, e.g., model factorization, is little explored in the context of ViT compression. This paper explores an efficient method for compressing vision transformers to enrich the toolset for obtaining compact attention-based vision models. Based on the new insight on the multi-head attention layer, we develop a highly efficient ViT compression solution, which outperforms the state-of-the-art pruning methods. For compressing DeiT-small and DeiT-base models on ImageNet, our proposed approach can achieve 0.45% and 0.76% higher top-1 accuracy even with fewer parameters. Our finding can also be applied to improve the customization efficiency of text-to-image diffusion models, with much faster training (up to $2.6\times$ speedup) and lower extra storage cost (up to $1927.5\times$ reduction) than the existing works.

cs.CV↗

arXiv2023

Human-machine knowledge hybrid augmentation method for surface defect detection based few-data learning

Visual-based defect detection is a crucial but challenging task in industrial quality control. Most mainstream methods rely on large amounts of existing or related domain data as auxiliary information. However, in actual industrial production, there are often multi-batch, low-volume manufacturing scenarios with rapidly changing task demands, making it difficult to obtain sufficient and diverse defect data. This paper proposes a parallel solution that uses a human-machine knowledge hybrid augmentation method to help the model extract unknown important features. Specifically, by incorporating experts' knowledge of abnormality to create data with rich features, positions, sizes, and backgrounds, we can quickly accumulate an amount of data from scratch and provide it to the model as prior knowledge for few-data learning. The proposed method was evaluated on the magnetic tile dataset and achieved F1-scores of 60.73%, 70.82%, 77.09%, and 82.81% when using 2, 5, 10, and 15 training images, respectively. Compared to the traditional augmentation method's F1-score of 64.59%, the proposed method achieved an 18.22% increase in the best result, demonstrating its feasibility and effectiveness in few-data industrial defect detection.

cs.CV↗

arXiv2023

HALOC: Hardware-Aware Automatic Low-Rank Compression for Compact Neural Networks

Low-rank compression is an important model compression strategy for obtaining compact neural network models. In general, because the rank values directly determine the model complexity and model accuracy, proper selection of layer-wise rank is very critical and desired. To date, though many low-rank compression approaches, either selecting the ranks in a manual or automatic way, have been proposed, they suffer from costly manual trials or unsatisfied compression performance. In addition, all of the existing works are not designed in a hardware-aware way, limiting the practical performance of the compressed models on real-world hardware platforms. To address these challenges, in this paper we propose HALOC, a hardware-aware automatic low-rank compression framework. By interpreting automatic rank selection from an architecture search perspective, we develop an end-to-end solution to determine the suitable layer-wise ranks in a differentiable and hardware-aware way. We further propose design principles and mitigation strategy to efficiently explore the rank space and reduce the potential interference problem. Experimental results on different datasets and hardware platforms demonstrate the effectiveness of our proposed approach. On CIFAR-10 dataset, HALOC enables 0.07% and 0.38% accuracy increase over the uncompressed ResNet-20 and VGG-16 models with 72.20% and 86.44% fewer FLOPs, respectively. On ImageNet dataset, HALOC achieves 0.9% higher top-1 accuracy than the original ResNet-18 model with 66.16% fewer FLOPs. HALOC also shows 0.66% higher top-1 accuracy increase than the state-of-the-art automatic low-rank compression solution with fewer computational and memory costs. In addition, HALOC demonstrates the practical speedups on different hardware platforms, verified by the measurement results on desktop GPU, embedded GPU and ASIC accelerator.

cs.LG↗

arXiv2022

Algorithm and Hardware Co-Design of Energy-Efficient LSTM Networks for Video Recognition with Hierarchical Tucker Tensor Decomposition

Long short-term memory (LSTM) is a type of powerful deep neural network that has been widely used in many sequence analysis and modeling applications. However, the large model size problem of LSTM networks make their practical deployment still very challenging, especially for the video recognition tasks that require high-dimensional input data. Aiming to overcome this limitation and fully unlock the potentials of LSTM models, in this paper we propose to perform algorithm and hardware co-design towards high-performance energy-efficient LSTM networks. At algorithm level, we propose to develop fully decomposed hierarchical Tucker (FDHT) structure-based LSTM, namely FDHT-LSTM, which enjoys ultra-low model complexity while still achieving high accuracy. In order to fully reap such attractive algorithmic benefit, we further develop the corresponding customized hardware architecture to support the efficient execution of the proposed FDHT-LSTM model. With the delicate design of memory access scheme, the complicated matrix transformation can be efficiently supported by the underlying hardware without any access conflict in an on-the-fly way. Our evaluation results show that both the proposed ultra-compact FDHT-LSTM models and the corresponding hardware accelerator achieve very high performance. Compared with the state-of-the-art compressed LSTM models, FDHT-LSTM enjoys both order-of-magnitude reduction in model size and significant accuracy improvement across different video recognition datasets. Meanwhile, compared with the state-of-the-art tensor decomposed model-oriented hardware TIE, our proposed FDHT-LSTM architecture achieves better performance in throughput, area efficiency and energy efficiency, respectively on LSTM-Youtube workload. For LSTM-UCF workload, our proposed design also outperforms TIE with higher throughput, higher energy efficiency and comparable area efficiency.

cs.CV↗

arXiv2022

GIFT: Graph-guIded Feature Transfer for Cold-Start Video Click-Through Rate Prediction

Short video has witnessed rapid growth in the past few years in e-commerce platforms like Taobao. To ensure the freshness of the content, platforms need to release a large number of new videos every day, making conventional click-through rate (CTR) prediction methods suffer from the item cold-start problem. In this paper, we propose GIFT, an efficient Graph-guIded Feature Transfer system, to fully take advantages of the rich information of warmed-up videos to compensate for the cold-start ones. Specifically, we establish a heterogeneous graph that contains physical and semantic linkages to guide the feature transfer process from warmed-up video to cold-start videos. The physical linkages represent explicit relationships, while the semantic linkages measure the proximity of multi-modal representations of two videos. We elaborately design the feature transfer function to make aware of different types of transferred features (e.g., id representations and historical statistics) from different metapaths on the graph. We conduct extensive experiments on a large real-world dataset, and the results show that our GIFT system outperforms SOTA methods significantly and brings a 6.82% lift on CTR in the homepage of Taobao App.

cs.IR↗

arXiv2022

Quasi-uniaxial pressure induced superconductivity in stoichiometric compound UTe$_2$

The recent discovery of superconductivity in heavy Fermion compound UTe2, a candidate topological and triplet-paired superconductor, has aroused widespread interest. However, to date, there is no consensus on whether the stoichiometric sample of UTe2 is superconducting or not due to lack of reliable evidence to distinguish the difference between the nominal and real compositions of samples. Here, we are the first to clarify that the stoichiometric UT2 is non-superconducting at ambient pressure and under hydrostatic pressure up to 6 GPa, however we find that it can be compressed into superconductivity by application of quasi-uniaxial pressure. Measurements of resistivity, magnetoresistance and susceptibility reveal that the quasi-uniaxial pressure results in a suppression of the Kondo coherent state seen at ambient pressure, and then leads to a superconductivity initially emerged on the ab-plane at 1.5 GPa. At 4.8 GPa, the superconductivity is developed in three crystallographic directions. The superconducting state coexists with an exotic magnetic ordered state that develops just below the onset temperature of the superconducting transition. The discovery of the quasi-uniaxial-pressure-induced superconductivity with exotic magnetic state in the stoichiometric UTe2 not only provide new understandings on this compound, but also highlight the vital role of Te deficiency in developing the superconductivity at ambient pressures.

cond-mat.supr-con↗

arXiv2022

Braiding lateral morphotropic grain boundary in homogeneitic oxides

Interfaces formed by correlated oxides offer a critical avenue for discovering emergent phenomena and quantum states. However, the fabrication of oxide interfaces with variable crystallographic orientations and strain states integrated along a film plane is extremely challenge by conventional layer-by-layer stacking or self-assembling. Here, we report the creation of morphotropic grain boundaries (GBs) in laterally interconnected cobaltite homostructures. Single-crystalline substrates and suspended ultrathin freestanding membranes provide independent templates for coherent epitaxy and constraint on the growth orientation, resulting in seamless and atomically sharp GBs. Electronic states and magnetic behavior in hybrid structures are laterally modulated and isolated by GBs, enabling artificially engineered functionalities in the planar matrix. Our work offers a simple and scalable method for fabricating unprecedented innovative interfaces through controlled synthesis routes as well as provides a platform for exploring potential applications in neuromorphics, solid state batteries, and catalysis.

cond-mat.mtrl-sci↗

arXiv2022

RankSim: Ranking Similarity Regularization for Deep Imbalanced Regression

Data imbalance, in which a plurality of the data samples come from a small proportion of labels, poses a challenge in training deep neural networks. Unlike classification, in regression the labels are continuous, potentially boundless, and form a natural ordering. These distinct features of regression call for new techniques that leverage the additional information encoded in label-space relationships. This paper presents the RankSim (ranking similarity) regularizer for deep imbalanced regression, which encodes an inductive bias that samples that are closer in label space should also be closer in feature space. In contrast to recent distribution smoothing based approaches, RankSim captures both nearby and distant relationships: for a given data sample, RankSim encourages the sorted list of its neighbors in label space to match the sorted list of its neighbors in feature space. RankSim is complementary to conventional imbalanced learning techniques, including re-weighting, two-stage training, and distribution smoothing, and lifts the state-of-the-art performance on three imbalanced regression benchmarks: IMDB-WIKI-DIR, AgeDB-DIR, and STS-B-DIR.

cs.LG↗

arXiv2021

N3H-Core: Neuron-designed Neural Network Accelerator via FPGA-based Heterogeneous Computing Cores

Accelerating the neural network inference by FPGA has emerged as a popular option, since the reconfigurability and high performance computing capability of FPGA intrinsically satisfies the computation demand of the fast-evolving neural algorithms. However, the popular neural accelerators on FPGA (e.g., Xilinx DPU) mainly utilize the DSP resources for constructing their processing units, while the rich LUT resources are not well exploited. Via the software-hardware co-design approach, in this work, we develop an FPGA-based heterogeneous computing system for neural network acceleration. From the hardware perspective, the proposed accelerator consists of DSP- and LUT-based GEneral Matrix-Multiplication (GEMM) computing cores, which forms the entire computing system in a heterogeneous fashion. The DSP- and LUT-based GEMM cores are computed w.r.t a unified Instruction Set Architecture (ISA) and unified buffers. Along the data flow of the neural network inference path, the computation of the convolution/fully-connected layer is split into two portions, handled by the DSP- and LUT-based GEMM cores asynchronously. From the software perspective, we mathematically and systematically model the latency and resource utilization of the proposed heterogeneous accelerator, regarding varying system design configurations. Through leveraging the reinforcement learning technique, we construct a framework to achieve end-to-end selection and optimization of the design specification of target heterogeneous accelerator, including workload split strategy, mixed-precision quantization scheme, and resource allocation of DSP- and LUT-core. In virtue of the proposed design framework and heterogeneous computing system, our design outperforms the state-of-the-art Mix&Match design with latency reduced by 1.12-1.32x with higher inference accuracy. The N3H-core is open-sourced at: https://github.com/elliothe/N3H_Core.

cs.AR↗

arXiv2021

MUSE: Feature Self-Distillation with Mutual Information and Self-Information

We present a novel information-theoretic approach to introduce dependency among features of a deep convolutional neural network (CNN). The core idea of our proposed method, called MUSE, is to combine MUtual information and SElf-information to jointly improve the expressivity of all features extracted from different layers in a CNN. We present two variants of the realization of MUSE -- Additive Information and Multiplicative Information. Importantly, we argue and empirically demonstrate that MUSE, compared to other feature discrepancy functions, is a more functional proxy to introduce dependency and effectively improve the expressivity of all features in the knowledge distillation framework. MUSE achieves superior performance over a variety of popular architectures and feature discrepancy functions for self-distillation and online distillation, and performs competitively with the state-of-the-art methods for offline distillation. MUSE is also demonstrably versatile that enables it to be easily extended to CNN-based models on tasks other than image classification such as object detection.

cs.CV↗

arXiv2021

Superconductivity modulated by structural phase transitions in pressurized vanadium-based kagome metals

The interplay of superconductivity with electronic and structural instabilities on the kagome lattice provides a fertile ground for the emergence of unusual phenomena. The vanadium-based kagome metals $A$V$_3$Sb$_5$ ($A=$ K, Rb, Cs) exhibit superconductivity on an almost ideal kagome lattice, with the superconducting transition temperature $T_{\rm c}$ forming two domes upon pressure-tuning. The first dome arises from the competition between superconductivity and a charge-density-wave, whereas the origin for the second dome remains unclear. Herein, we show that the appearance of the second superconducting dome in KV$_3$Sb$_5$ and RbV$_3$Sb$_5$ is associated with transitions from hexagonal $P6$/$mmm$ to monoclinic $P2$/$m$ structures, evidenced by splitting of structural peaks from synchrotron powder X-ray diffraction experiments and imaginary phonon frequencies in first-principles calculations. In KV$_3$Sb$_5$, transition to an orthorhombic $Pmmm$ structure is further observed for pressure $p\gtrsim20$ GPa, and is correlated with the strong suppression of $T_{\rm c}$ in the second superconducting dome. Our findings indicate distortions of the crystal structure modulates superconductivity in $A$V$_3$Sb$_5$ under pressure, providing a platform to study the emergence of superconductivity in the presence of multiple structural instabilities.

cond-mat.supr-con↗

arXiv2021

Blind Image Quality Assessment for MRI with A Deep Three-dimensional content-adaptive Hyper-Network

Image Quality Assessment (IQA) is of great value in the workflow of Magnetic Resonance Imaging (MRI)-based analysis. Blind IQA (BIQA) methods are especially required since high-quality reference MRI images are usually not available. Recently, many efforts have been devoted to developing deep learning-based BIQA approaches. However, the performance of these methods is limited due to the utilization of simple content-non-adaptive network parameters and the waste of the important 3D spatial information of the medical images. To address these issues, we design a 3D content-adaptive hyper-network for MRI BIQA. The overall 3D configuration enables the exploration of comprehensive 3D spatial information from MRI images, while the developed content-adaptive hyper-network contributes to the self-adaptive capacity of network parameters and thus, facilitates better BIQA performance. The effectiveness of the proposed method is extensively evaluated on the open dataset, MRIQC. Promising performance is achieved compared with the corresponding baseline and 4 state-of-the-art BIQA methods. We make our code available at \url{https://git.openi.org.cn/SIAT_Wangshanshan/HyS-Net}.

eess.IV↗

arXiv2021

Universal quantum transition from superconducting to insulating states in pressurized Bi2Sr2CaCu2O8+δ superconductors

Copper oxide superconductors have continually fascinated the communities of condensed matter physics and material sciences because they host the highest ambient-pressure superconducting transition temperature (Tc) and mysterious physics. Searching for the universal correlation between the superconducting state and its normal state or neighboring ground state is believed to be an effective way for finding clues to elucidate the underlying mechanism of the superconductivity. One of the common pictures for the copper oxide superconductors is that a well-behaved metallic phase will present after the superconductivity is entirely suppressed by chemical doping or application of the magnetic field. Here, we report a different observation of universal quantum transition from superconducting state to insulating-like state under pressure in the under-, optimally- and over-doped Bi2212 superconductors with two CuO2 planes in a unit cell. The same phenomenon has been also found in the Bi2201 superconductor with one CuO2 plane and the Bi2223 superconductor with three CuO2 planes in a unit cell. These results not only provide fresh information but also pose a new challenge for achieving a unified understanding on the underlying physics of the high-Tc superconductivity.

cond-mat.supr-con↗

arXiv2021

Various magnetism of the compressed antiferromagnetic topological insulator EuSn2As2

We report a comprehensive high-pressure study on the antiferromagnetic topological insulator EuSn2As2 up to 21.1 GPa through measurements of synchrotron x-ray diffraction, electrical resistance, magnetic resistance, and Hall transports combined with first-principles calculations. No evident trace of a structural phase transition is detected. The Neel temperatures determined from resistance are increased from 24 to 77 K under pressure, which is resulted from the enhanced magnetic exchange couplings between Eu2+ ions yielded by our first-principles calculations. The negative magnetoresistance of EuSn2As2 persists to higher temperatures accordantly. However, the enhancement of the observed Néel temperatures deviates from the calculations obviously above 10.0 GPa. In addition, the magnitude of the magnetoresistance, the Hall coefficients, and the charge carrier densities show abrupt changes between 6.9 to 10.0 GPa. The abrupt changes probably originate from a pressure induced valence change of Eu ions from a divalent state to a divalent and trivalent mixed state. Our results provide insights into variation of the magnetism of EuSn2As2 and similar antiferromagnetic topological insulators under pressure.

cond-mat.str-el↗

arXiv2021

GRN: Generative Rerank Network for Context-wise Recommendation

Reranking is attracting incremental attention in the recommender systems, which rearranges the input ranking list into the final rank-ing list to better meet user demands. Most existing methods greedily rerank candidates through the rating scores from point-wise or list-wise models. Despite effectiveness, neglecting the mutual influence between each item and its contexts in the final ranking list often makes the greedy strategy based reranking methods sub-optimal. In this work, we propose a new context-wise reranking framework named Generative Rerank Network (GRN). Specifically, we first design the evaluator, which applies Bi-LSTM and self-attention mechanism to model the contextual information in the labeled final ranking list and predict the interaction probability of each item more precisely. Afterwards, we elaborate on the generator, equipped with GRU, attention mechanism and pointer network to select the item from the input ranking list step by step. Finally, we apply cross-entropy loss to train the evaluator and, subsequently, policy gradient to optimize the generator under the guidance of the evaluator. Empirical results show that GRN consistently and significantly outperforms state-of-the-art point-wise and list-wise methods. Moreover, GRN has achieved a performance improvement of 5.2% on PV and 6.1% on IPV metric after the successful deployment in one popular recommendation scenario of Taobao application.

cs.IR↗

arXiv2021

Revisit Recommender System in the Permutation Prospective

Recommender systems (RS) work effective at alleviating information overload and matching user interests in various web-scale applications. Most RS retrieve the user's favorite candidates and then rank them by the rating scores in the greedy manner. In the permutation prospective, however, current RS come to reveal the following two limitations: 1) They neglect addressing the permutation-variant influence within the recommended results; 2) Permutation consideration extends the latent solution space exponentially, and current RS lack the ability to evaluate the permutations. Both drive RS away from the permutation-optimal recommended results and better user experience. To approximate the permutation-optimal recommended results effectively and efficiently, we propose a novel permutation-wise framework PRS in the re-ranking stage of RS, which consists of Permutation-Matching (PMatch) and Permutation-Ranking (PRank) stages successively. Specifically, the PMatch stage is designed to obtain the candidate list set, where we propose the FPSA algorithm to generate multiple candidate lists via the permutation-wise and goal-oriented beam search algorithm. Afterwards, for the candidate list set, the PRank stage provides a unified permutation-wise ranking criterion named LR metric, which is calculated by the rating scores of elaborately designed permutation-wise model DPWN. Finally, the list with the highest LR score is recommended to the user. Empirical results show that PRS consistently and significantly outperforms state-of-the-art methods. Moreover, PRS has achieved a performance improvement of 11.0% on PV metric and 8.7% on IPV metric after the successful deployment in one popular recommendation scenario of Taobao application.

cs.IR↗

arXiv2021

Variational Selective Autoencoder: Learning from Partially-Observed Heterogeneous Data

Learning from heterogeneous data poses challenges such as combining data from various sources and of different types. Meanwhile, heterogeneous data are often associated with missingness in real-world applications due to heterogeneity and noise of input sources. In this work, we propose the variational selective autoencoder (VSAE), a general framework to learn representations from partially-observed heterogeneous data. VSAE learns the latent dependencies in heterogeneous data by modeling the joint distribution of observed data, unobserved data, and the imputation mask which represents how the data are missing. It results in a unified model for various downstream tasks including data generation and imputation. Evaluation on both low-dimensional and high-dimensional heterogeneous datasets for these two tasks shows improvement over state-of-the-art models.

cs.LG↗

arXiv2020

Personalized Adaptive Meta Learning for Cold-start User Preference Prediction

A common challenge in personalized user preference prediction is the cold-start problem. Due to the lack of user-item interactions, directly learning from the new users' log data causes serious over-fitting problem. Recently, many existing studies regard the cold-start personalized preference prediction as a few-shot learning problem, where each user is the task and recommended items are the classes, and the gradient-based meta learning method (MAML) is leveraged to address this challenge. However, in real-world application, the users are not uniformly distributed (i.e., different users may have different browsing history, recommended items, and user profiles. We define the major users as the users in the groups with large numbers of users sharing similar user information, and other users are the minor users), existing MAML approaches tend to fit the major users and ignore the minor users. To address this cold-start task-overfitting problem, we propose a novel personalized adaptive meta learning approach to consider both the major and the minor users with three key contributions: 1) We are the first to present a personalized adaptive learning rate meta-learning approach to improve the performance of MAML by focusing on both the major and minor users. 2) To provide better personalized learning rates for each user, we introduce a similarity-based method to find similar users as a reference and a tree-based method to store users' features for fast search. 3) To reduce the memory usage, we design a memory agnostic regularizer to further reduce the space complexity to constant while maintain the performance. Experiments on MovieLens, BookCrossing, and real-world production datasets reveal that our method outperforms the state-of-the-art methods dramatically for both the minor and major users.

cs.IR↗