ICLR 2023 论文列表

Motivated by the fact that forward and backward passes of a deep network naturally form symmetric mappings between input and output representations, we introduce a simple yet effective self-supervised vision model pretraining framework inspired by energy-based models (EBMs). In the proposed framework, we model energy estimation and data restoration as the forward and backward passes of a single network without any auxiliary components, e.g., an extra decoder. For the forward pass, we fit a network to an energy function that assigns low energy scores to samples that belong to an unlabeled dataset, and high energy otherwise. For the backward pass, we restore data from corrupted versions iteratively using gradient-based optimization along the direction of energy minimization. In this way, we naturally fold the encoder-decoder architecture widely used in masked image modeling into the forward and backward passes of a single vision model. Our framework accepts a wide range of pretext tasks with different data corruption methods, and permits models to be pretrained from masked image modeling, patch sorting, and image restoration, including super-resolution, denoising, and colorization. We support our findings with extensive experiments, and show the proposed method delivers comparable and even better performance with remarkably fewer epochs of training compared to the state-of-the-art self-supervised vision model pretraining methods. Our findings shed light on further exploring self-supervised vision model pretraining and pretext tasks beyond masked image modeling.

State-of-the-art computer vision models are mostly trained with supervised learning using human-labeled images, which limits their scalability due to the expensive annotation cost. While self-supervised representation learning has achieved impressive progress, it still requires a second stage of finetuning on labeled data. On the other hand, models pre-trained with large-scale text supervision (e.g., CLIP) have enabled zero-shot transfer to downstream image classification tasks. However, the zero-shot performance of CLIP-like models are often insufficient for real-world adoption. In this paper, we aim to leverage the abundant unlabeled data from a target domain to improve the performance of a pre-trained zero-shot classifier, by unsupervised finetuning of the pre-trained model. We propose Masked Unsupervised Self-Training (MUST), a new approach which leverages two different and complimentary sources of training signals: pseudo-labels and raw images. MUST jointly optimizes three objectives to learn both class-level global feature and pixel-level local feature and enforces a regularization between the two. We demonstrate the efficacy of MUST on 8 downstream tasks across a variety of domains, where it improves upon CLIP by a large margin. MUST also outperforms supervised few-shot adaptation methods. It achieves a top-1 accuracy of 77.7% on ImageNet using ViT-B, +9.4% higher than CLIP, and +6.2% higher than 16-shot CLIP adaptation. Our code is available at https://github.com/salesforce/MUST.

Reliable application of machine learning-based decision systems in the wild is one of the major challenges currently investigated by the field. A large portion of established approaches aims to detect erroneous predictions by means of assigning confidence scores. This confidence may be obtained by either quantifying the model's predictive uncertainty, learning explicit scoring functions, or assessing whether the input is in line with the training distribution. Curiously, while these approaches all state to address the same eventual goal of detecting failures of a classifier upon real-world application, they currently constitute largely separated research fields with individual evaluation protocols, which either exclude a substantial part of relevant methods or ignore large parts of relevant failure sources. In this work, we systematically reveal current pitfalls caused by these inconsistencies and derive requirements for a holistic and realistic evaluation of failure detection. To demonstrate the relevance of this unified perspective, we present a large-scale empirical study for the first time enabling benchmarking confidence scoring functions w.r.t all relevant methods and failure sources. The revelation of a simple softmax response baseline as the overall best performing method underlines the drastic shortcomings of current evaluation in the plethora of publicized research on confidence scoring. Code and trained models are at https://github.com/https://github.com/IML-DKFZ/fd-shifts

In this paper, we study a novel inference paradigm, termed as schema inference, that learns to deductively infer the explainable predictions by rebuilding the prior deep neural network (DNN) forwarding scheme, guided by the prevalent philosophical cognitive concept of schema. We strive to reformulate the conventional model inference pipeline into a graph matching policy that associates the extracted visual concepts of an image with the pre-computed scene impression, by analogy with human reasoning mechanism via impression matching. To this end, we devise an elaborated architecture, termed as SchemaNet, as a dedicated instantiation of the proposed schema inference concept, that models both the visual semantics of input instances and the learned abstract imaginations of target categories as topological relational graphs. Meanwhile, to capture and leverage the compositional contributions of visual semantics in a global view, we also introduce a universal Feat2Graph scheme in SchemaNet to establish the relational graphs that contain abundant interaction information. Both the theoretical analysis and the experimental results on several benchmarks demonstrate that the proposed schema inference achieves encouraging performance and meanwhile yields a clear picture of the deductive process leading to the predictions. Our code is available at https://github.com/zhfeing/SchemaNet-PyTorch.

Data-free model stealing aims to replicate a target model without direct access to either the training data or the target model. To accomplish this, existing methods use a generator to produce samples in order to train a student model to match the target model outputs. To this end, the two main challenges are estimating gradients of the target model without access to its parameters, and generating a diverse set of training samples that thoroughly explores the input space. We propose a Dual Student method where two students are symmetrically trained in order to provide the generator a criterion to generate samples that the two students disagree on. On one hand, disagreement on a sample implies at least one student has classified the sample incorrectly when compared to the target model. This incentive towards disagreement implicitly encourages the generator to explore more diverse regions of the input space. On the other hand, our method utilizes gradients of student models to indirectly estimate gradients of the target model. We show that this novel training objective for the generator network is equivalent to optimizing a lower bound on the generator's loss if we had access to the target model gradients. In other words, our method alters the standard data-free model stealing paradigm by substituting the target model with a separate student model, thereby creating a lower bound which can be directly optimized without additional target model queries or separate synthetic datasets. We show that our new optimization framework provides more accurate gradient estimation of the target model and better accuracies on benchmark classification datasets. Additionally, our approach balances improved query efficiency with training computation cost. Finally, we demonstrate that our method serves as a better proxy model for transfer-based adversarial attacks than existing data-free model stealing methods.

Self-attention is key to the remarkable success of transformers in sequence modeling tasks including many applications in natural language processing and computer vision. Like neural network layers, these attention mechanisms are often developed by heuristics and experience. To provide a principled framework for constructing attention layers in transformers, we show that the self-attention corresponds to the support vector expansion derived from a support vector regression problem, whose primal formulation has the form of a neural network layer. Using our framework, we derive popular attention layers used in practice and propose two new attentions: 1) the Batch Normalized Attention (Attention-BN) derived from the batch normalization layer and 2) the Attention with Scaled Head (Attention-SH) derived from using less training data to fit the SVR model. We empirically demonstrate the advantages of the Attention-BN and Attention-SH in reducing head redundancy, increasing the model's accuracy, and improving the model's efficiency in a variety of practical applications including image and time-series classification.

Contrastive learning is a cornerstone underlying recent progress in multi-view and multimodal learning, e.g., in representation learning with image/caption pairs. While its effectiveness is not yet fully understood, a line of recent work reveals that contrastive learning can invert the data generating process and recover ground truth latent factors shared between views. In this work, we present new identifiability results for multimodal contrastive learning, showing that it is possible to recover shared factors in a more general setup than the multi-view setting studied previously. Specifically, we distinguish between the multi-view setting with one generative mechanism (e.g., multiple cameras of the same type) and the multimodal setting that is characterized by distinct mechanisms (e.g., cameras and microphones). Our work generalizes previous identifiability results by redefining the generative process in terms of distinct mechanisms with modality-specific latent variables. We prove that contrastive learning can block-identify latent factors shared between modalities, even when there are nontrivial dependencies between factors. We empirically verify our identifiability results with numerical simulations and corroborate our findings on a complex multimodal dataset of image/text pairs. Zooming out, our work provides a theoretical basis for multimodal representation learning and explains in which settings multimodal contrastive learning can be effective in practice.

Direct speech-to-speech translation (S2ST) with discrete units leverages recent progress in speech representation learning. Specifically, a sequence of discrete representations derived in a self-supervised manner are predicted from the model and passed to a vocoder for speech reconstruction, while still facing the following challenges: 1) Acoustic multimodality: the discrete units derived from speech with same content could be indeterministic due to the acoustic property (e.g., rhythm, pitch, and energy), which causes deterioration of translation accuracy; 2) high latency: current S2ST systems utilize autoregressive models which predict each unit conditioned on the sequence previously generated, failing to take full advantage of parallelism. In this work, we propose TranSpeech, a speech-to-speech translation model with bilateral perturbation. To alleviate the acoustic multimodal problem, we propose bilateral perturbation (BiP), which consists of the style normalization and information enhancement stages, to learn only the linguistic information from speech samples and generate more deterministic representations. With reduced multimodality, we step forward and become the first to establish a non-autoregressive S2ST technique, which repeatedly masks and predicts unit choices and produces high-accuracy results in just a few cycles. Experimental results on three language pairs demonstrate that BiP yields an improvement of 2.9 BLEU on average compared with a baseline textless S2ST model. Moreover, our parallel decoding shows a significant reduction of inference latency, enabling speedup up to 21.4x than autoregressive technique. Audio samples are available at https://TranSpeech.github.io

Transformers have seen growing interest in processing different modalities, including language and image data. As a result, we can process vision and language data using transformers that are architecturally similar. Leveraging this feature of transformers, we propose weight sharing across two transformer backbones and within the same transformer backbone and pruning across two backbones in a unified framework. More specifically, we investigate weight sharing and pruning for two components of the transformers: (1) Multi-Head Attention (MSA) and (2) Feed-Forward Network (FFN) layers. To jointly perform weight sharing and pruning, we propose to use a regularization term to align model weights and the desired structure during the multimodal pre-training step. The structure vectors of sharing and pruning are generated by using a hypernetwork, which can capture complex interactions between pruning and sharing across layers and modalities. We train the hypernetwork and model weights iteratively so that the learned structure evolves along with model weights. After minimizing the proposed objective in the pre-training step, we perform weight sharing and pruning and fine-tune the compressed model on downstream tasks. Finally, we perform experiments on vision and language tasks, including Referring Expression Comprehension (REC), Visual Question Answering (VQA), and Object Detection using the state-of-the-art grounding based models: MDETR and GLIP. Our experiments show that we can compress these models by $35-40\%$ by sharing and pruning MSA and FFN weights without almost any loss in accuracy.

In this work we introduce a novel optimization algorithm for image recovery under learned sparse and low-rank constraints, which are parameterized with weighted extensions of the $\ell_p^p$-vector and $\mathcal{S}_p^p$ Schatten-matrix quasi-norms for $0\!<p\!\le1$, respectively. Our proposed algorithm generalizes the Iteratively Reweighted Least Squares (IRLS) method, used for signal recovery under $\ell_1$ and nuclear-norm constrained minimization. Further, we interpret our overall minimization approach as a recurrent network that we then employ to deal with inverse low-level computer vision problems. Thanks to the convergence guarantees that our IRLS strategy offers, we are able to train the derived reconstruction networks using a memory-efficient implicit back-propagation scheme, which does not pose any restrictions on their effective depth. To assess our networks' performance, we compare them against other existing reconstruction methods on several inverse problems, namely image deblurring, super-resolution, demosaicking and sparse recovery. Our reconstruction results are shown to be very competitive and in many cases outperform those of existing unrolled networks, whose number of parameters is orders of magnitude higher than that of our learned models.

The recent success of Vision Transformers is shaking the long dominance of Convolutional Neural Networks (CNNs) in image recognition for a decade. Specifically, in terms of robustness on out-of-distribution samples, recent research finds that Transformers are inherently more robust than CNNs, regardless of different training setups. Moreover, it is believed that such superiority of Transformers should largely be credited to their \emph{self-attention-like architectures per se}. In this paper, we question that belief by closely examining the design of Transformers. Our findings lead to three highly effective architecture designs for boosting robustness, yet simple enough to be implemented in several lines of code, namely a) patchifying input images, b) enlarging kernel size, and c) reducing activation layers and normalization layers. Bringing these components together, we are able to build pure CNN architectures without any attention-like operations that are as robust as, or even more robust than, Transformers. We hope this work can help the community better understand the design of robust neural architectures. The code is publicly available at https://github.com/UCSC-VLAA/RobustCNN.

Learning from large amounts of unsupervised data and a small amount of supervision is an important open problem in computer vision. We propose a new semi-supervised learning method, Semantic Positives via Pseudo-Labels (SEMPPL), that combines labelled and unlabelled data to learn informative representations. Our method extends self-supervised contrastive learning—where representations are shaped by distinguishing whether two samples represent the same underlying datum (positives) or not (negatives)—with a novel approach to selecting positives. To enrich the set of positives, we leverage the few existing ground-truth labels to predict the missing ones through a k-nearest neighbors classifier by using the learned embeddings of the labelled data. We thus extend the set of positives with datapoints having the same pseudo-label and call these semantic positives. We jointly learn the representation and predict bootstrapped pseudo-labels. This creates a reinforcing cycle. Strong initial representations enable better pseudo-label predictions which then improve the selection of semantic positives and lead to even better representations. SEMPPL outperforms competing semi-supervised methods setting new state-of-the-art performance of 76% and 68.5% top-1accuracy when using a ResNet-50 and training on 10% and 1% of labels on ImageNet, respectively. Furthermore, when using selective kernels, SEMPPL significantly outperforms previous state-of-the-art achieving 72.3% and 78.3% top-1accuracy on ImageNet with 1% and 10% labels, respectively, which improves absolute +7.8% and +6.2% over previous work. SEMPPL also exhibits state-of-the-art performance over larger ResNet models as well as strong robustness, out-of-distribution and transfer performance. We release the checkpoints and the evaluation code at https://github.com/deepmind/semppl.

The transferability of adversarial perturbations between image models has been extensively studied. In this case, an attack is generated from a known surrogate \eg, the ImageNet trained model, and transferred to change the decision of an unknown (black-box) model trained on an image dataset. However, attacks generated from image models do not capture the dynamic nature of a moving object or a changing scene due to a lack of temporal cues within image models. This leads to reduced transferability of adversarial attacks from representation-enriched \emph{image} models such as Supervised Vision Transformers (ViTs), Self-supervised ViTs (\eg, DINO), and Vision-language models (\eg, CLIP) to black-box \emph{video} models. In this work, we induce dynamic cues within the image models without sacrificing their original performance on images. To this end, we optimize \emph{temporal prompts} through frozen image models to capture motion dynamics. Our temporal prompts are the result of a learnable transformation that allows optimizing for temporal gradients during an adversarial attack to fool the motion dynamics. Specifically, we introduce spatial (image) and temporal (video) cues within the same source model through task-specific prompts. Attacking such prompts maximizes the adversarial transferability from image-to-video and image-to-image models using the attacks designed for image models. As an example, an iterative attack launched from image model Deit-B with temporal prompts reduces generalization (top1 \% accuracy) of a video model by 35\% on Kinetics-400. Our approach also improves adversarial transferability to image models by 9\% on ImageNet w.r.t the current state-of-the-art approach. Our attack results indicate that the attacker does not need specialized architectures, \eg, divided space-time attention, 3D convolutions, or multi-view convolution networks for different data modalities. Image models are effective surrogates to optimize an adversarial attack to fool black-box models in a changing environment over time. Code is available at \url{https://bit.ly/3Xd9gRQ}

We introduce compositional soft prompting (CSP), a parameter-efficient learning technique to improve the zero-shot compositionality of large-scale pretrained vision-language models (VLMs) like CLIP. We develop CSP for compositional zero-shot learning, the task of predicting unseen attribute-object compositions (e.g., old cat and young tiger). VLMs have a flexible text encoder that can represent arbitrary classes as natural language prompts but they often underperform task-specific architectures on the compositional zero-shot benchmark datasets. CSP treats the attributes and objects that define classes as learnable tokens of vocabulary. During training, the vocabulary is tuned to recognize classes that compose tokens in multiple ways (e.g., old cat and white cat). At test time, we recompose the learned attribute-object vocabulary in new combinations to recognize novel classes. We show that CSP outperforms the CLIP on benchmark datasets by an average of 10.9 percentage points on AUC. CSP also outperforms CoOp, a soft prompting method that fine-tunes the prefix context tokens, by an average of 5.8 percentage points on AUC. We perform additional experiments to show that CSP improves generalization to higher-order attribute-attribute-object compositions (e.g., old white cat) and combinations of pretrained attributes and fine-tuned objects. The code is available at https://github.com/BatsResearch/csp.

Deep learning has had tremendous success at learning low-dimensional representations of high-dimensional data. This success would be impossible if there was no hidden low-dimensional structure in data of interest; this existence is posited by the manifold hypothesis, which states that the data lies on an unknown manifold of low intrinsic dimension. In this paper, we argue that this hypothesis does not properly capture the low-dimensional structure typically present in image data. Assuming that data lies on a single manifold implies intrinsic dimension is identical across the entire data space, and does not allow for subregions of this space to have a different number of factors of variation. To address this deficiency, we consider the union of manifolds hypothesis, which states that data lies on a disjoint union of manifolds of varying intrinsic dimensions. We empirically verify this hypothesis on commonly-used image datasets, finding that indeed, observed data lies on a disconnected set and that intrinsic dimension is not constant. We also provide insights into the implications of the union of manifolds hypothesis in deep learning, both supervised and unsupervised, showing that designing models with an inductive bias for this structure improves performance across classification and generative modelling tasks. Our code is available at https://github.com/layer6ai-labs/UoMH.

High-resolution images are prevalent in various applications, such as autonomous driving and computer-aided diagnosis. However, training neural networks on such images is computationally challenging and easily leads to out-of-memory errors even on modern GPUs. We propose a simple method, Iterative Patch Selection (IPS), which decouples the memory usage from the input size and thus enables the processing of arbitrarily large images under tight hardware constraints. IPS achieves this by selecting only the most salient patches, which are then aggregated into a global representation for image recognition. For both patch selection and aggregation, a cross-attention based transformer is introduced, which exhibits a close connection to Multiple Instance Learning. Our method demonstrates strong performance and has wide applicability across different domains, training regimes and image sizes while using minimal accelerator memory. For example, we are able to finetune our model on whole-slide images consisting of up to 250k patches (>16 gigapixels) with only 5 GB of GPU VRAM at a batch size of 16.

The ability to predict future visual observations conditioned on past observations and motor commands can enable embodied agents to plan solutions to a variety of tasks in complex environments. This work shows that we can create good video prediction models by pre-training transformers via masked visual modeling. Our approach, named MaskViT, is based on two simple design decisions. First, for memory and training efficiency, we use two types of window attention: spatial and spatiotemporal. Second, during training, we mask a variable percentage of tokens instead of a fixed mask ratio. For inference, MaskViT generates all tokens via iterative refinement where we incrementally decrease the masking ratio following a mask scheduling function. On several datasets we demonstrate that MaskViT outperforms prior works in video prediction, is parameter efficient, and can generate high resolution videos ($256 \times $256). Further, we demonstrate the benefits of inference speedup (up to $512 \times$) due to iterative decoding by using MaskViT for planning on a real robot. Our work suggests that we can endow embodied agents with powerful predictive models by leveraging the general framework of masked visual modeling with minimal domain knowledge.

This paper presents a new Convolutional Neural Network, named Contextual Convolutional Network, that capably serves as a general-purpose backbone for visual recognition. Most existing convolutional backbones follow the representation-to-classification paradigm, where representations of the input are firstly generated by category-agnostic convolutional operations, and then fed into classifiers for specific perceptual tasks (e.g., classification and segmentation). In this paper, we deviate from this classic paradigm and propose to augment potential category memberships as contextual priors in the convolution for contextualized representation learning. Specifically, top-k likely classes from the preceding stage are encoded as a contextual prior vector. Based on this vector and the preceding features, offsets for spatial sampling locations and kernel weights are generated to modulate the convolution operations. The new convolutions can readily replace their plain counterparts in existing CNNs and can be easily trained end-to-end by standard back-propagation without additional supervision. The qualities of Contextual Convolutional Networks make it compatible with a broad range of vision tasks and boost the state-of-the-art architecture ConvNeXt-Tiny by 1.8% on top-1 accuracy of ImageNet classification. The superiority of the proposed model reveals the potential of contextualized representation learning for vision tasks. Code is available at: \url{https://github.com/liang4sx/contextual_cnn}.

This paper presents Universal Vision-Language Dense Retrieval (UniVL-DR), which builds a unified model for multi-modal retrieval. UniVL-DR encodes queries and multi-modality resources in an embedding space for searching candidates from different modalities. To learn a unified embedding space for multi-modal retrieval, UniVL-DR proposes two techniques: 1) Universal embedding optimization strategy, which contrastively optimizes the embedding space using the modality-balanced hard negatives; 2) Image verbalization method, which bridges the modality gap between images and texts in the raw data space. UniVL-DR achieves the state-of-the-art on the multi-modal open-domain question answering benchmark, WebQA, and outperforms all retrieval models on the two subtasks, text-text retrieval and text-image retrieval. It demonstrates that universal multi-modal search is feasible to replace the divide-and-conquer pipeline with a united model and also benefits single/cross modality tasks. All source codes of this work are available at https://github.com/OpenMatch/UniVL-DR.