Skip to yearly menu bar Skip to main content


Session

Deep Learning

Room 327 - 329

Moderator: Eitan Borgnia

Abstract:

Chat is not available.

Tue 19 July 7:30 - 7:35 PDT

Spotlight
Structural Entropy Guided Graph Hierarchical Pooling

Junran Wu · Xueyuan Chen · Ke Xu · Shangzhe Li

Following the success of convolution on non-Euclidean space, the corresponding pooling approaches have also been validated on various tasks regarding graphs. However, because of the fixed compression ratio and stepwise pooling design, these hierarchical pooling methods still suffer from local structure damage and suboptimal problem. In this work, inspired by structural entropy, we propose a hierarchical pooling approach, SEP, to tackle the two issues. Specifically, without assigning the layer-specific compression ratio, a global optimization algorithm is designed to generate the cluster assignment matrices for pooling at once. Then, we present an illustration of the local structure damage from previous methods in reconstruction of ring and grid synthetic graphs. In addition to SEP, we further design two classification models, SEP-G and SEP-N for graph classification and node classification, respectively. The results show that SEP outperforms state-of-the-art graph pooling methods on graph classification benchmarks and obtains superior performance on node classifications.

Tue 19 July 7:35 - 7:40 PDT

Spotlight
Self-Supervised Representation Learning via Latent Graph Prediction

Yaochen Xie · Zhao Xu · Shuiwang Ji

Self-supervised learning (SSL) of graph neural networks is emerging as a promising way of leveraging unlabeled data. Currently, most methods are based on contrastive learning adapted from the image domain, which requires view generation and a sufficient number of negative samples. In contrast, existing predictive models do not require negative sampling, but lack theoretical guidance on the design of pretext training tasks. In this work, we propose the LaGraph, a theoretically grounded predictive SSL framework based on latent graph prediction. Learning objectives of LaGraph are derived as self-supervised upper bounds to objectives for predicting unobserved latent graphs. In addition to its improved performance, LaGraph provides explanations for recent successes of predictive models that include invariance-based objectives. We provide theoretical analysis comparing LaGraph to related methods in different domains. Our experimental results demonstrate the superiority of LaGraph in performance and the robustness to decreasing of training sample size on both graph-level and node-level tasks.

Tue 19 July 7:40 - 7:45 PDT

Spotlight
DSTAGNN: Dynamic Spatial-Temporal Aware Graph Neural Network for Traffic Flow Forecasting

Shiyong Lan · Yitong Ma · Weikang Huang · Wenwu Wang · Hongyu Yang · pyang li

As a typical problem in time series analysis, traffic flow prediction is one of the most important application fields of machine learning. However, achieving highly accurate traffic flow prediction is a challenging task, due to the presence of complex dynamic spatial-temporal dependencies within a road network. This paper proposes a novel Dynamic Spatial-Temporal Aware Graph Neural Network (DSTAGNN) to model the complex spatial-temporal interaction in road network. First, considering the fact that historical data carries intrinsic dynamic information about the spatial structure of road networks, we propose a new dynamic spatial-temporal aware graph based on a data-driven strategy to replace the pre-defined static graph usually used in traditional graph convolution. Second, we design a novel graph neural network architecture, which can not only represent dynamic spatial relevance among nodes with an improved multi-head attention mechanism, but also acquire the wide range of dynamic temporal dependency from multi-receptive field features via multi-scale gated convolution. Extensive experiments on real-world data sets demonstrate that our proposed method significantly outperforms the state-of-the-art methods.

Tue 19 July 7:45 - 7:50 PDT

Spotlight
Coarsening the Granularity: Towards Structurally Sparse Lottery Tickets

Tianlong Chen · Xuxi Chen · Xiaolong Ma · Yanzhi Wang · Zhangyang “Atlas” Wang

The lottery ticket hypothesis (LTH) has shown that dense models contain highly sparse subnetworks (i.e., winning tickets) that can be trained in isolation to match full accuracy. Despite many exciting efforts being made, there is one "commonsense" rarely challenged: a winning ticket is found by iterative magnitude pruning (IMP) and hence the resultant pruned subnetworks have only unstructured sparsity. That gap limits the appeal of winning tickets in practice, since the highly irregular sparse patterns are challenging to accelerate on hardware. Meanwhile, directly substituting structured pruning for unstructured pruning in IMP damages performance more severely and is usually unable to locate winning tickets. In this paper, we demonstrate the first positive result that a structurally sparse winning ticket can be effectively found in general. The core idea is to append "post-processing techniques" after each round of (unstructured) IMP, to enforce the formation of structural sparsity. Specifically, we first "re-fill" pruned elements back in some channels deemed to be important, and then "re-group" non-zero elements to create flexible group-wise structural patterns. Both our identified channel- and group-wise structural subnetworks win the lottery, with substantial inference speedups readily supported by existing hardware. Extensive experiments, conducted on diverse datasets across multiple network backbones, consistently validate our proposal, showing that the hardware acceleration roadblock of LTH is now removed. Specifically, the structural winning tickets obtain up to {64.93%, 64.84%, 60.23%} running time savings at {36%~80%, 74%, 58%} sparsity on {CIFAR, Tiny-ImageNet, ImageNet}, while maintaining comparable accuracy. Code is at https://github.com/VITA-Group/Structure-LTH.

Tue 19 July 7:50 - 7:55 PDT

Spotlight
Omni-Granular Ego-Semantic Propagation for Self-Supervised Graph Representation Learning

Ling Yang · Shenda Hong

Unsupervised/self-supervised graph representation learning is critical for downstream node- and graph-level classification tasks. Global structure of graphs helps discriminating representations and existing methods mainly utilize the global structure by imposing additional supervisions. However, their global semantics are usually invariant for all nodes/graphs and they fail to explicitly embed the global semantics to enrich the representations. In this paper, we propose Omni-Granular Ego-Semantic Propagation for Self-Supervised Graph Representation Learning (OEPG). Specifically, we introduce instance-adaptive global-aware ego-semantic descriptors, leveraging the first- and second-order feature differences between each node/graph and hierarchical global clusters of the entire graph dataset. The descriptors can be explicitly integrated into local graph convolution as new neighbor nodes. Besides, we design an omni-granular normalization on the whole scales and hierarchies of the ego-semantic to assign attentional weight to each descriptor from an omni-granular perspective. Specialized pretext tasks and cross-iteration momentum update are further developed for local-global mutual adaptation. In downstream tasks, OEPG consistently achieves the best performance with a 2%~6% accuracy gain on multiple datasets cross scales and domains. Notably, OEPG also generalizes to quantity- and topology-imbalance scenarios.

Tue 19 July 7:55 - 8:00 PDT

Spotlight
Analyzing and Mitigating Interference in Neural Architecture Search

Jin Xu · Xu Tan · Kaitao Song · Renqian Luo · Yichong Leng · Tao Qin · Tie-Yan Liu · Jian Li

Weight sharing is a popular approach to reduce the training cost of neural architecture search (NAS) by reusing the weights of shared operators from previously trained child models. However, the rank correlation between the estimated accuracy and ground truth accuracy of those child models is low due to the interference among different child models caused by weight sharing. In this paper, we investigate the interference issue by sampling different child models and calculating the gradient similarity of shared operators, and observe that: 1) the interference on a shared operator between two child models is positively correlated with the number of different operators between them; 2) the interference is smaller when the inputs and outputs of the shared operator are more similar. Inspired by these two observations, we propose two approaches to mitigate the interference: 1) rather than randomly sampling child models for optimization, we propose a gradual modification scheme by modifying one operator between adjacent optimization steps to minimize the interference on the shared operators; 2) forcing the inputs and outputs of the operator across all child models to be similar to reduce the interference. Experiments on a BERT search space verify that mitigating interference via each of our proposed methods improves the rank correlation of super-pet and combining both methods can achieve better results. Our discovered architecture outperforms RoBERTa$_{\rm base}$ by 1.1 and 0.6 points and ELECTRA$_{\rm base}$ by 1.6 and 1.1 points on the dev and test set of GLUE benchmark. Extensive results on the BERT compression, reading comprehension and large-scale image classification tasks also demonstrate the effectiveness and generality of our proposed methods.

Tue 19 July 8:00 - 8:05 PDT

Spotlight
Reverse Engineering $\ell_p$ attacks: A block-sparse optimization approach with recovery guarantees

Darshan Thaker · Paris Giampouras · Rene Vidal

Deep neural network-based classifiers have been shown to be vulnerable to imperceptible perturbations to their input, such as $\ell_p$-bounded norm adversarial attacks. This has motivated the development of many defense methods, which are then broken by new attacks, and so on. This paper focuses on a different but related problem of reverse engineering adversarial attacks. Specifically, given an attacked signal, we study conditions under which one can determine the type of attack ($\ell_1$, $\ell_2$ or $\ell_\infty$) and recover the clean signal. We pose this problem as a block-sparse recovery problem, where both the signal and the attack are assumed to lie in a union of subspaces that includes one subspace per class and one subspace per attack type. We derive geometric conditions on the subspaces under which any attacked signal can be decomposed as the sum of a clean signal plus an attack. In addition, by determining the subspaces that contain the signal and the attack, we can also classify the signal and determine the attack type. Experiments on digit and face classification demonstrate the effectiveness of the proposed approach.

Tue 19 July 8:05 - 8:25 PDT

Oral
Unified Scaling Laws for Routed Language Models

Aidan Clark · Diego de Las Casas · Aurelia Guy · Arthur Mensch · Michela Paganini · Jordan Hoffmann · Bogdan Damoc · Blake Hechtman · Trevor Cai · Sebastian Borgeaud · George van den Driessche · Eliza Rutherford · Tom Hennigan · Matthew Johnson · Albin Cassirer · Chris Jones · Elena Buchatskaya · David Budden · Laurent Sifre · Simon Osindero · Oriol Vinyals · Marc'Aurelio Ranzato · Jack Rae · Erich Elsen · Koray Kavukcuoglu · Karen Simonyan

The performance of a language model has been shown to be effectively modeled as a power-law in its parameter count. Here we study the scaling behaviors of Routing Networks: architectures that conditionally use only a subset of their parameters while processing an input. For these models, parameter count and computational requirement form two independent axes along which an increase leads to better performance. In this work we derive and justify scaling laws defined on these two variables which generalize those known for standard language models and describe the performance of a wide range of routing architectures trained via three different techniques. Afterwards we provide two applications of these laws: first deriving an Effective Parameter Count along which all models scale at the same rate, and then using the scaling coefficients to give a quantitative comparison of the three routing techniques considered. Our analysis derives from an extensive evaluation of Routing Networks across five orders of magnitude of size, including models with hundreds of experts and hundreds of billions of parameters.

Tue 19 July 8:25 - 8:30 PDT

Spotlight
DRAGONN: Distributed Randomized Approximate Gradients of Neural Networks

Zhuang Wang · Zhaozhuo Xu · Xinyu Wu · Anshumali Shrivastava · T. S. Eugene Ng

Data-parallel distributed training (DDT) has become the de-facto standard for accelerating the training of most deep learning tasks on massively parallel hardware. In the DDT paradigm, the communication overhead of gradient synchronization is the major efficiency bottleneck. A widely adopted approach to tackle this issue is gradient sparsification (GS). However, the current GS methods introduce significant new overhead in compressing the gradients, outweighing the communication overhead and becoming the new efficiency bottleneck. In this paper, we propose DRAGONN, a randomized hashing algorithm for GS in DDT. DRAGONN can significantly reduce the compression time by up to 70% compared to state-of-the-art GS approaches, and achieve up to 3.52x speedup in total training throughput.

Tue 19 July 8:30 - 8:35 PDT

Spotlight
A deep convolutional neural network that is invariant to time rescaling

Brandon G Jacques · Zoran Tiganj · Aakash Sarkar · Marc Howard · Per Sederberg

Human learners can readily understand speech, or a melody, when it is presented slower or faster than usual. This paper presents a deep CNN (SITHCon) that uses a logarithmically compressed temporal representation at each level. Because rescaling the time of the input results in a translation of $\log$ time, and because the output of the convolution is invariant to translations, this network can generalize to out-of-sample data that are temporal rescalings of a learned pattern. We compare the performance of SITHCon to a Temporal Convolution Network (TCN) on classification and regression problems with both univariate and multivariate time series. We find that SITHCon, unlike TCN, generalizes robustly over rescalings of about an order of magnitude. Moreover, we show that the network can generalize over exponentially large scales without retraining the weights simply by extending the range of the logarithmically-compressed temporal memory.

Tue 19 July 8:35 - 8:40 PDT

Spotlight
LyaNet: A Lyapunov Framework for Training Neural ODEs

Ivan Dario Jimenez Rodriguez · Aaron Ames · Yisong Yue

We propose a method for training ordinary differential equations by using a control-theoretic Lyapunov condition for stability. Our approach, called LyaNet, is based on a novel Lyapunov loss formulation that encourages the inference dynamics to converge quickly to the correct prediction. Theoretically, we show that minimizing Lyapunov loss guarantees exponential convergence to the correct solution and enables a novel robustness guarantee. We also provide practical algorithms, including one that avoids the cost of backpropagating through a solver or using the adjoint method. Relative to standard Neural ODE training, we empirically find that LyaNet can offer improved prediction performance, faster convergence of inference dynamics, and improved adversarial robustness. Our code is available at https://github.com/ivandariojr/LyapunovLearning.

Tue 19 July 8:40 - 8:45 PDT

Spotlight
Transfer and Marginalize: Explaining Away Label Noise with Privileged Information

Mark Collier · Rodolphe Jenatton · Efi Kokiopoulou · Jesse Berent

Supervised learning datasets often have privileged information, in the form of features which are available at training time but are not available at test time e.g. the ID of the annotator that provided the label. We argue that privileged information is useful for explaining away label noise, thereby reducing the harmful impact of noisy labels. We develop a simple and efficient method for supervised learning with neural networks: it transfers via weight sharing the knowledge learned with privileged information and approximately marginalizes over privileged information at test time. Our method, TRAM (TRansfer and Marginalize), has minimal training time overhead and has the same test-time cost as not using privileged information. TRAM performs strongly on CIFAR-10H, ImageNet and Civil Comments benchmarks.

Tue 19 July 8:45 - 8:50 PDT

Spotlight
On Collective Robustness of Bagging Against Data Poisoning

Ruoxin Chen · Zenan Li · Jie Li · Junchi Yan · Chentao Wu

Bootstrap aggregating (bagging) is an effective ensemble protocol, which is believed can enhance robustness by its majority voting mechanism. Recent works further prove the sample-wise robustness certificates for certain forms of bagging (e.g. partition aggregation). Beyond these particular forms, in this paper, we propose the first collective certification for general bagging to compute the tight robustness against the global poisoning attack. Specifically, we compute the maximum number of simultaneously changed predictions via solving a binary integer linear programming (BILP) problem. Then we analyze the robustness of vanilla bagging and give the upper bound of the tolerable poison budget. Based on this analysis, we propose hash bagging to improve the robustness of vanilla bagging almost for free. This is achieved by modifying the random subsampling in vanilla bagging to a hash-based deterministic subsampling, as a way of controlling the influence scope for each poisoning sample universally. Our extensive experiments show the notable advantage in terms of applicability and robustness. Our code is available at https://github.com/Emiyalzn/ICML22-CRB.

Tue 19 July 8:50 - 8:55 PDT

Spotlight
Hindering Adversarial Attacks with Implicit Neural Representations

Andrei A Rusu · Dan Andrei Calian · Sven Gowal · Raia Hadsell

We introduce the Lossy Implicit Network Activation Coding (LINAC) defence, an input transformation which successfully hinders several common adversarial attacks on CIFAR-10 classifiers for perturbations up to 8/255 in Linf norm and 0.5 in L2 norm. Implicit neural representations are used to approximately encode pixel colour intensities in 2D images such that classifiers trained on transformed data appear to have robustness to small perturbations without adversarial training or large drops in performance. The seed of the random number generator used to initialise and train the implicit neural representation turns out to be necessary information for stronger generic attacks, suggesting its role as a private key. We devise a Parametric Bypass Approximation (PBA) attack strategy for key-based defences, which successfully invalidates an existing method in this category. Interestingly, our LINAC defence also hinders some transfer and adaptive attacks, including our novel PBA strategy. Our results emphasise the importance of a broad range of customised attacks despite apparent robustness according to standard evaluations.

Tue 19 July 8:55 - 9:00 PDT

Spotlight
From Noisy Prediction to True Label: Noisy Prediction Calibration via Generative Model

HeeSun Bae · Seungjae Shin · Byeonghu Na · JoonHo Jang · Kyungwoo Song · IL CHUL MOON

Noisy labels are inevitable yet problematic in machine learning society. It ruins the generalization of a classifier by making the classifier over-fitted to noisy labels. Existing methods on noisy label have focused on modifying the classifier during the training procedure. It has two potential problems. First, these methods are not applicable to a pre-trained classifier without further access to training. Second, it is not easy to train a classifier and regularize all negative effects from noisy labels, simultaneously. We suggest a new branch of method, Noisy Prediction Calibration (NPC) in learning with noisy labels. Through the introduction and estimation of a new type of transition matrix via generative model, NPC corrects the noisy prediction from the pre-trained classifier to the true label as a post-processing scheme. We prove that NPC theoretically aligns with the transition matrix based methods. Yet, NPC empirically provides more accurate pathway to estimate true label, even without involvement in classifier learning. Also, NPC is applicable to any classifier trained with noisy label methods, if training instances and its predictions are available. Our method, NPC, boosts the classification performances of all baseline models on both synthetic and real-world datasets. The implemented code is available at https://github.com/BaeHeeSun/NPC.