PyTorchKR Papers

최신 논문

총 1,512편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2211.14241공식 코드 · Papers with Code

Look Around and Refer: 2D Synthetic Semantics Knowledge Distillation for 3D Visual Grounding

The 3D visual grounding task has been explored with visual and language streams comprehending referential language to identify target objects in 3D scenes. However, most existing methods devote the visual stream to capturing the 3D visual clues using off-the-shelf point clouds encoders. The main question we address in this paper is "can we consolidate the 3D visual stream by 2D clues synthesized from point clouds and efficiently utilize them in training and testing?". The main idea is to assist the 3D encoder by incorporating rich 2D object representations without requiring extra 2D inputs. To this end, we leverage 2D clues, synthetically generated from 3D point clouds, and empirically show their aptitude to boost the quality of the learned visual representations. We validate our approach through comprehensive experiments on Nr3D, Sr3D, and ScanRefer datasets and show consistent performance gains compared to existing methods. Our proposed module, dubbed as Look Around and Refer (LAR), significantly outperforms the state-of-the-art 3D visual grounding techniques on three benchmarks, i.e., Nr3D, Sr3D, and ScanRefer. The code is available at https://eslambakr.github.io/LAR.github.io/.

arXiv:2211.13929공식 코드 · Papers with Code

XKD: Cross-modal Knowledge Distillation with Domain Alignment for Video Representation Learning

We present XKD, a novel self-supervised framework to learn meaningful representations from unlabelled videos. XKD is trained with two pseudo objectives. First, masked data reconstruction is performed to learn modality-specific representations from audio and visual streams. Next, self-supervised cross-modal knowledge distillation is performed between the two modalities through a teacher-student setup to learn complementary information. We introduce a novel domain alignment strategy to tackle domain discrepancy between audio and visual modalities enabling effective cross-modal knowledge distillation. Additionally, to develop a general-purpose network capable of handling both audio and visual streams, modality-agnostic variants of XKD are introduced, which use the same pretrained backbone for different audio and visual tasks. Our proposed cross-modal knowledge distillation improves video action classification by $8\%$ to $14\%$ on UCF101, HMDB51, and Kinetics400. Additionally, XKD improves multimodal action classification by $5.5\%$ on Kinetics-Sound. XKD shows state-of-the-art performance in sound classification on ESC50, achieving top-1 accuracy of $96.5\%$.

arXiv:2211.13955공식 코드 · Papers with Code

MPCViT: Searching for Accurate and Efficient MPC-Friendly Vision Transformer with Heterogeneous Attention

Secure multi-party computation (MPC) enables computation directly on encrypted data and protects both data and model privacy in deep learning inference. However, existing neural network architectures, including Vision Transformers (ViTs), are not designed or optimized for MPC and incur significant latency overhead. We observe Softmax accounts for the major latency bottleneck due to a high communication complexity, but can be selectively replaced or linearized without compromising the model accuracy. Hence, in this paper, we propose an MPC-friendly ViT, dubbed MPCViT, to enable accurate yet efficient ViT inference in MPC. Based on a systematic latency and accuracy evaluation of the Softmax attention and other attention variants, we propose a heterogeneous attention optimization space. We also develop a simple yet effective MPC-aware neural architecture search algorithm for fast Pareto optimization. To further boost the inference efficiency, we propose MPCViT+, to jointly optimize the Softmax attention and other network components, including GeLU, matrix multiplication, etc. With extensive experiments, we demonstrate that MPCViT achieves 1.9%, 1.3% and 3.6% higher accuracy with 6.2x, 2.9x and 1.9x latency reduction compared with baseline ViT, MPCFormer and THE-X on the Tiny-ImageNet dataset, respectively. MPCViT+ further achieves a better Pareto front compared with MPCViT. The code and models for evaluation are available at https://github.com/PKU-SEC-Lab/mpcvit.

arXiv:2211.13125공식 코드 · Papers with Code

EEG aided boosting of single-lead ECG based sleep staging with Deep Knowledge Distillation

An electroencephalogram (EEG) signal is currently accepted as a standard for automatic sleep staging. Lately, Near-human accuracy in automated sleep staging has been achievable by Deep Learning (DL) based approaches, enabling multi-fold progress in this area. However, An extensive and expensive clinical setup is required for EEG based sleep staging. Additionally, the EEG setup being obtrusive in nature and requiring an expert for setup adds to the inconvenience of the subject under study, making it adverse in the point of care setting. An unobtrusive and more suitable alternative to EEG is Electrocardiogram (ECG). Unsurprisingly, compared to EEG in sleep staging, its performance remains sub-par. In order to take advantage of both the modalities, transferring knowledge from EEG to ECG is a reasonable approach, ultimately boosting the performance of ECG based sleep staging. Knowledge Distillation (KD) is a promising notion in DL that shares knowledge from a superior performing but usually more complex teacher model to an inferior but compact student model. Building upon this concept, a cross-modality KD framework assisting features learned through models trained on EEG to improve ECG-based sleep staging performance is proposed. Additionally, to better understand the distillation approach, extensive experimentation on the independent modules of the proposed model was conducted. Montreal Archive of Sleep Studies (MASS) dataset consisting of 200 subjects was utilized for this study. The results from the proposed model for weighted-F1-score in 3-class and 4-class sleep staging showed a 13.40 \% and 14.30 \% improvement, respectively. This study demonstrates the feasibility of KD for single-channel ECG based sleep staging's performance enhancement in 3-class (W-R-N) and 4-class (W-R-L-D) classification.

arXiv:2211.12933공식 코드 · Papers with Code

Join the High Accuracy Club on ImageNet with A Binary Neural Network Ticket

Binary neural networks are the extreme case of network quantization, which has long been thought of as a potential edge machine learning solution. However, the significant accuracy gap to the full-precision counterparts restricts their creative potential for mobile applications. In this work, we revisit the potential of binary neural networks and focus on a compelling but unanswered problem: how can a binary neural network achieve the crucial accuracy level (e.g., 80%) on ILSVRC-2012 ImageNet? We achieve this goal by enhancing the optimization process from three complementary perspectives: (1) We design a novel binary architecture BNext based on a comprehensive study of binary architectures and their optimization process. (2) We propose a novel knowledge-distillation technique to alleviate the counter-intuitive overfitting problem observed when attempting to train extremely accurate binary models. (3) We analyze the data augmentation pipeline for binary networks and modernize it with up-to-date techniques from full-precision models. The evaluation results on ImageNet show that BNext, for the first time, pushes the binary model accuracy boundary to 80.57% and significantly outperforms all the existing binary networks. Code and trained models are available at: https://github.com/hpi-xnor/BNext.git.

arXiv:2211.12881공식 코드 · Papers with Code

DGEKT: A Dual Graph Ensemble Learning Method for Knowledge Tracing

Knowledge tracing aims to trace students' evolving knowledge states by predicting their future performance on concept-related exercises. Recently, some graph-based models have been developed to incorporate the relationships between exercises to improve knowledge tracing, but only a single type of relationship information is generally explored. In this paper, we present a novel Dual Graph Ensemble learning method for Knowledge Tracing (DGEKT), which establishes a dual graph structure of students' learning interactions to capture the heterogeneous exercise-concept associations and interaction transitions by hypergraph modeling and directed graph modeling, respectively. To ensemble the dual graph models, we introduce the technique of online knowledge distillation, due to the fact that although the knowledge tracing model is expected to predict students' responses to the exercises related to different concepts, it is optimized merely with respect to the prediction accuracy on a single exercise at each step. With online knowledge distillation, the dual graph models are adaptively combined to form a stronger teacher model, which in turn provides its predictions on all exercises as extra supervision for better modeling ability. In the experiments, we compare DGEKT against eight knowledge tracing baselines on three benchmark datasets, and the results demonstrate that DGEKT achieves state-of-the-art performance.

arXiv:2211.12044공식 코드 · Papers with Code

Backdoor Cleansing with Unlabeled Data

Due to the increasing computational demand of Deep Neural Networks (DNNs), companies and organizations have begun to outsource the training process. However, the externally trained DNNs can potentially be backdoor attacked. It is crucial to defend against such attacks, i.e., to postprocess a suspicious model so that its backdoor behavior is mitigated while its normal prediction power on clean inputs remain uncompromised. To remove the abnormal backdoor behavior, existing methods mostly rely on additional labeled clean samples. However, such requirement may be unrealistic as the training data are often unavailable to end users. In this paper, we investigate the possibility of circumventing such barrier. We propose a novel defense method that does not require training labels. Through a carefully designed layer-wise weight re-initialization and knowledge distillation, our method can effectively cleanse backdoor behaviors of a suspicious network with negligible compromise in its normal behavior. In experiments, we show that our method, trained without labels, is on-par with state-of-the-art defense methods trained using labels. We also observe promising defense results even on out-of-distribution data. This makes our method very practical. Code is available at: https://github.com/luluppang/BCU.

arXiv:2211.12494공식 코드 · Papers with Code

On the Transferability of Visual Features in Generalized Zero-Shot Learning

Generalized Zero-Shot Learning (GZSL) aims to train a classifier that can generalize to unseen classes, using a set of attributes as auxiliary information, and the visual features extracted from a pre-trained convolutional neural network. While recent GZSL methods have explored various techniques to leverage the capacity of these features, there has been an extensive growth of representation learning techniques that remain under-explored. In this work, we investigate the utility of different GZSL methods when using different feature extractors, and examine how these models' pre-training objectives, datasets, and architecture design affect their feature representation ability. Our results indicate that 1) methods using generative components for GZSL provide more advantages when using recent feature extractors; 2) feature extractors pre-trained using self-supervised learning objectives and knowledge distillation provide better feature representations, increasing up to 15% performance when used with recent GZSL techniques; 3) specific feature extractors pre-trained with larger datasets do not necessarily boost the performance of GZSL methods. In addition, we investigate how GZSL methods fare against CLIP, a more recent multi-modal pre-trained model with strong zero-shot performance. We found that GZSL tasks still benefit from generative-based GZSL methods along with CLIP's internet-scale pre-training to achieve state-of-the-art performance in fine-grained datasets. We release a modular framework for analyzing representation learning issues in GZSL here: https://github.com/uvavision/TV-GZSL

arXiv:2211.11434공식 코드 · Papers with Code

Privacy in Practice: Private COVID-19 Detection in X-Ray Images (Extended Version)

Machine learning (ML) can help fight pandemics like COVID-19 by enabling rapid screening of large volumes of images. To perform data analysis while maintaining patient privacy, we create ML models that satisfy Differential Privacy (DP). Previous works exploring private COVID-19 models are in part based on small datasets, provide weaker or unclear privacy guarantees, and do not investigate practical privacy. We suggest improvements to address these open gaps. We account for inherent class imbalances and evaluate the utility-privacy trade-off more extensively and over stricter privacy budgets. Our evaluation is supported by empirically estimating practical privacy through black-box Membership Inference Attacks (MIAs). The introduced DP should help limit leakage threats posed by MIAs, and our practical analysis is the first to test this hypothesis on the COVID-19 classification task. Our results indicate that needed privacy levels might differ based on the task-dependent practical threat from MIAs. The results further suggest that with increasing DP guarantees, empirical privacy leakage only improves marginally, and DP therefore appears to have a limited impact on practical MIA defense. Our findings identify possibilities for better utility-privacy trade-offs, and we believe that empirical attack-specific privacy estimation can play a vital role in tuning for practical privacy.

arXiv:2211.11355공식 코드 · Papers with Code

Blind Knowledge Distillation for Robust Image Classification

Optimizing neural networks with noisy labels is a challenging task, especially if the label set contains real-world noise. Networks tend to generalize to reasonable patterns in the early training stages and overfit to specific details of noisy samples in the latter ones. We introduce Blind Knowledge Distillation - a novel teacher-student approach for learning with noisy labels by masking the ground truth related teacher output to filter out potentially corrupted knowledge and to estimate the tipping point from generalizing to overfitting. Based on this, we enable the estimation of noise in the training data with Otsus algorithm. With this estimation, we train the network with a modified weighted cross-entropy loss function. We show in our experiments that Blind Knowledge Distillation detects overfitting effectively during training and improves the detection of clean and noisy labels on the recently published CIFAR-N dataset. Code is available at GitHub.

arXiv:2211.11159공식 코드 · Papers with Code

Directed Acyclic Graph Factorization Machines for CTR Prediction via Knowledge Distillation

With the growth of high-dimensional sparse data in web-scale recommender systems, the computational cost to learn high-order feature interaction in CTR prediction task largely increases, which limits the use of high-order interaction models in real industrial applications. Some recent knowledge distillation based methods transfer knowledge from complex teacher models to shallow student models for accelerating the online model inference. However, they suffer from the degradation of model accuracy in knowledge distillation process. It is challenging to balance the efficiency and effectiveness of the shallow student models. To address this problem, we propose a Directed Acyclic Graph Factorization Machine (KD-DAGFM) to learn the high-order feature interactions from existing complex interaction models for CTR prediction via Knowledge Distillation. The proposed lightweight student model DAGFM can learn arbitrary explicit feature interactions from teacher networks, which achieves approximately lossless performance and is proved by a dynamic programming algorithm. Besides, an improved general model KD-DAGFM+ is shown to be effective in distilling both explicit and implicit feature interactions from any complex teacher model. Extensive experiments are conducted on four real-world datasets, including a large-scale industrial dataset from WeChat platform with billions of feature dimensions. KD-DAGFM achieves the best performance with less than 21.5% FLOPs of the state-of-the-art method on both online and offline experiments, showing the superiority of DAGFM to deal with the industrial scale data in CTR prediction task. Our implementation code is available at: https://github.com/RUCAIBox/DAGFM.

arXiv:2211.11014공식 코드 · Papers with Code

Understanding and Improving Knowledge Distillation for Quantization-Aware Training of Large Transformer Encoders

Knowledge distillation (KD) has been a ubiquitous method for model compression to strengthen the capability of a lightweight model with the transferred knowledge from the teacher. In particular, KD has been employed in quantization-aware training (QAT) of Transformer encoders like BERT to improve the accuracy of the student model with the reduced-precision weight parameters. However, little is understood about which of the various KD approaches best fits the QAT of Transformers. In this work, we provide an in-depth analysis of the mechanism of KD on attention recovery of quantized large Transformers. In particular, we reveal that the previously adopted MSE loss on the attention score is insufficient for recovering the self-attention information. Therefore, we propose two KD methods; attention-map and attention-output losses. Furthermore, we explore the unification of both losses to address task-dependent preference between attention-map and output losses. The experimental results on various Transformer encoder models demonstrate that the proposed KD methods achieve state-of-the-art accuracy for QAT with sub-2-bit weight quantization.

arXiv:2211.11300공식 코드 · Papers with Code

Multi-Level Knowledge Distillation for Out-of-Distribution Detection in Text

Self-supervised representation learning has proved to be a valuable component for out-of-distribution (OoD) detection with only the texts of in-distribution (ID) examples. These approaches either train a language model from scratch or fine-tune a pre-trained language model using ID examples, and then take the perplexity output by the language model as OoD scores. In this paper, we analyze the complementary characteristics of both OoD detection methods and propose a multi-level knowledge distillation approach that integrates their strengths while mitigating their limitations. Specifically, we use a fine-tuned model as the teacher to teach a randomly initialized student model on the ID examples. Besides the prediction layer distillation, we present a similarity-based intermediate layer distillation method to thoroughly explore the representation space of the teacher model. In this way, the learned student can better represent the ID data manifold while gaining a stronger ability to map OoD examples outside the ID data manifold with the regularization inherited from pre-training. Besides, the student model sees only ID examples during parameter learning, further promoting more distinguishable features for OoD detection. We conduct extensive experiments over multiple benchmark datasets, i.e., CLINC150, SST, ROSTD, 20 NewsGroups, and AG News; showing that the proposed method yields new state-of-the-art performance. We also explore its application as an AIGC detector to distinguish between answers generated by ChatGPT and human experts. It is observed that our model exceeds human evaluators in the pair-expert task on the Human ChatGPT Comparison Corpus.

arXiv:2211.09949공식 코드 · Papers with Code

Is Smaller Always Faster? Tradeoffs in Compressing Self-Supervised Speech Transformers

Transformer-based self-supervised models have achieved remarkable success in speech processing, but their large size and high inference cost present significant challenges for real-world deployment. While numerous compression techniques have been proposed, inconsistent evaluation metrics make it difficult to compare their practical effectiveness. In this work, we conduct a comprehensive study of four common compression methods, including weight pruning, head pruning, low-rank approximation, and knowledge distillation on self-supervised speech Transformers. We evaluate each method under three key metrics: parameter count, multiply-accumulate operations, and real-time factor. Results show that each method offers distinct advantages. In addition, we contextualize recent compression techniques, comparing DistilHuBERT, FitHuBERT, LightHuBERT, ARMHuBERT, and STaRHuBERT under the same framework, offering practical guidance on compression for deployment.

arXiv:2211.09394공식 코드 · Papers with Code

ConNER: Consistency Training for Cross-lingual Named Entity Recognition

Cross-lingual named entity recognition (NER) suffers from data scarcity in the target languages, especially under zero-shot settings. Existing translate-train or knowledge distillation methods attempt to bridge the language gap, but often introduce a high level of noise. To solve this problem, consistency training methods regularize the model to be robust towards perturbations on data or hidden states. However, such methods are likely to violate the consistency hypothesis, or mainly focus on coarse-grain consistency. We propose ConNER as a novel consistency training framework for cross-lingual NER, which comprises of: (1) translation-based consistency training on unlabeled target-language data, and (2) dropoutbased consistency training on labeled source-language data. ConNER effectively leverages unlabeled target-language data and alleviates overfitting on the source language to enhance the cross-lingual adaptability. Experimental results show our ConNER achieves consistent improvement over various baseline methods.

arXiv:2211.09386공식 코드 · Papers with Code

BEVDistill: Cross-Modal BEV Distillation for Multi-View 3D Object Detection

3D object detection from multiple image views is a fundamental and challenging task for visual scene understanding. Owing to its low cost and high efficiency, multi-view 3D object detection has demonstrated promising application prospects. However, accurately detecting objects through perspective views is extremely difficult due to the lack of depth information. Current approaches tend to adopt heavy backbones for image encoders, making them inapplicable for real-world deployment. Different from the images, LiDAR points are superior in providing spatial cues, resulting in highly precise localization. In this paper, we explore the incorporation of LiDAR-based detectors for multi-view 3D object detection. Instead of directly training a depth prediction network, we unify the image and LiDAR features in the Bird-Eye-View (BEV) space and adaptively transfer knowledge across non-homogenous representations in a teacher-student paradigm. To this end, we propose \textbf{BEVDistill}, a cross-modal BEV knowledge distillation (KD) framework for multi-view 3D object detection. Extensive experiments demonstrate that the proposed method outperforms current KD approaches on a highly-competitive baseline, BEVFormer, without introducing any extra cost in the inference phase. Notably, our best model achieves 59.4 NDS on the nuScenes test leaderboard, achieving new state-of-the-art in comparison with various image-based detectors. Code will be available at https://github.com/zehuichen123/BEVDistill.

arXiv:2211.07715공식 코드 · Papers with Code

Fast DistilBERT on CPUs

Transformer-based language models have become the standard approach to solving natural language processing tasks. However, industry adoption usually requires the maximum throughput to comply with certain latency constraints that prevents Transformer models from being used in production. To address this gap, model compression techniques such as quantization and pruning may be used to improve inference efficiency. However, these compression techniques require specialized software to apply and deploy at scale. In this work, we propose a new pipeline for creating and running Fast Transformer models on CPUs, utilizing hardware-aware pruning, knowledge distillation, quantization, and our own Transformer inference runtime engine with optimized kernels for sparse and quantized operators. We demonstrate the efficiency of our pipeline by creating a Fast DistilBERT model showing minimal accuracy loss on the question-answering SQuADv1.1 benchmark, and throughput results under typical production constraints and environments. Our results outperform existing state-of-the-art Neural Magic's DeepSparse runtime performance by up to 50% and up to 4.1x performance speedup over ONNX Runtime. Source code is publicly available at https://github.com/intel/intel-extension-for-transformers.

arXiv:2211.08161공식 코드 · Papers with Code

An Investigation of the Combination of Rehearsal and Knowledge Distillation in Continual Learning for Spoken Language Understanding

Continual learning refers to a dynamical framework in which a model receives a stream of non-stationary data over time and must adapt to new data while preserving previously acquired knowledge. Unluckily, neural networks fail to meet these two desiderata, incurring the so-called catastrophic forgetting phenomenon. Whereas a vast array of strategies have been proposed to attenuate forgetting in the computer vision domain, for speech-related tasks, on the other hand, there is a dearth of works. In this paper, we consider the joint use of rehearsal and knowledge distillation (KD) approaches for spoken language understanding under a class-incremental learning scenario. We report on multiple KD combinations at different levels in the network, showing that combining feature-level and predictions-level KDs leads to the best results. Finally, we provide an ablation study on the effect of the size of the rehearsal memory that corroborates the efficacy of our approach for low-resource devices.

arXiv:2211.07248공식 코드 · Papers with Code

FedCL: Federated Multi-Phase Curriculum Learning to Synchronously Correlate User Heterogeneity

Federated Learning (FL) is a decentralized learning method used to train machine learning algorithms. In FL, a global model iteratively collects the parameters of local models without accessing their local data. However, a significant challenge in FL is handling the heterogeneity of local data distribution, which often results in a drifted global model that is difficult to converge. To address this issue, current methods employ different strategies such as knowledge distillation, weighted model aggregation, and multi-task learning. These approaches are referred to as asynchronous FL, as they align user models either locally or post-hoc, where model drift has already occurred or has been underestimated. In this paper, we propose an active and synchronous correlation approach to address the challenge of user heterogeneity in FL. Specifically, our approach aims to approximate FL as standard deep learning by actively and synchronously scheduling user learning pace in each round with a dynamic multi-phase curriculum. A global curriculum is formed by an auto-regressive auto-encoder that integrates all user curricula on the server. This global curriculum is then divided into multiple phases and broadcast to users to measure and align the domain-agnostic learning pace. Empirical studies demonstrate that our approach outperforms existing asynchronous approaches in terms of generalization performance, even in the presence of severe user heterogeneity.

arXiv:2211.07198공식 코드 · Papers with Code

Fcaformer: Forward Cross Attention in Hybrid Vision Transformer

Currently, one main research line in designing a more efficient vision transformer is reducing the computational cost of self attention modules by adopting sparse attention or using local attention windows. In contrast, we propose a different approach that aims to improve the performance of transformer-based architectures by densifying the attention pattern. Specifically, we proposed forward cross attention for hybrid vision transformer (FcaFormer), where tokens from previous blocks in the same stage are secondary used. To achieve this, the FcaFormer leverages two innovative components: learnable scale factors (LSFs) and a token merge and enhancement module (TME). The LSFs enable efficient processing of cross tokens, while the TME generates representative cross tokens. By integrating these components, the proposed FcaFormer enhances the interactions of tokens across blocks with potentially different semantics, and encourages more information flows to the lower levels. Based on the forward cross attention (Fca), we have designed a series of FcaFormer models that achieve the best trade-off between model size, computational cost, memory cost, and accuracy. For example, without the need for knowledge distillation to strengthen training, our FcaFormer achieves 83.1% top-1 accuracy on Imagenet with only 16.3 million parameters and about 3.6 billion MACs. This saves almost half of the parameters and a few computational costs while achieving 0.7% higher accuracy compared to distilled EfficientFormer.