Search

Large scale vision and language models can achieve impressive zero-shot recognition performance by mapping class specific text queries to image content. Two distinct challenges that remain however, are high sensitivity to the choice of handcrafted class names that define queries, and the difficulty of adaptation to new, smaller datasets. Towards addressing these problems, we propose to leverage available data to learn, for each class, an optimal word embedding as a function of the visual content. By learning new word embeddings on an otherwise frozen model, we are able to retain zero-shot capabilities for new classes, easily adapt models to new datasets, and adjust potentially erroneous, non-descriptive or ambiguous class names. We show that our solution can easily be integrated in image classification and object detection pipelines, yields significant performance gains in multiple scenarios and provides insights into model biases and labelling errors., Comment: Accepted to CVPR 2023

Organic coatings are the most widely employed approach for the promotion of corrosion resistance of magnesium (Mg) alloys. Unfortunately, traditional organic coatings are weakly bonded to Mg substrates due to physical adsorption. Herein, a polyethylacrylate (PEA) coating was fabricated on Mg-Zn-Y-Nd alloy via electro-grafting. The surface structure and chemical composition were characterized by means of scanning electron microscope (SEM), energy dispersive X-ray spectroscopy (EDS), atomic force microscope (AFM) and Fourier transform infrared (FTIR) as well as time of flight-secondary ion mass spectrometer (ToF-SIMS). The results showed that the surface roughness of PEA coating was dominated by scan rate; while the coverage and integrity of PEA coating were mainly affected by the monomer concentration and sweep circles. ToF-SIMS results indicated that PEA coating was wholly covered on Mg alloy, and the presence of C2H3Mg− fragment confirmed the covalent bond between PEA coating and Mg alloy. In addition, DFT calculation results of the adsorption of EA molecules with Mg substrate agree well with the experimental phenomena and observation, suggesting that the electrons in 3 s orbit of Mg atoms and 2pz orbit of C1 atom participated in the formation of covalent bond between PEA coating and Mg substrate. Potentiodynamic polarization curves and immersion test demonstrated that the PEA coatings could effectively enhance the corrosion resistance of Mg alloy. The platelet adhesion results designated that platelets were barely visible on PEA coating, which implied that PEA coating could effectively prevent the thrombosis and coagulation of platelets. PEA coating might be a promising candidate coating of Mg alloy for cardiovascular stent.

Additional file 15: Figure S10. The Contribution of Family Cancer History and Pathological Features to the DrABC Model.

Additional file 21: Figure S13. The Suggested Pipeline of Genetic Testing for Women with Breast Cancer.

Additional file 19: Figure S12. Online Website for the DrABC Model.

Recent advances on large-scale pre-training have shown great potentials of leveraging a large set of Pre-Trained Models (PTMs) for improving Out-of-Distribution (OoD) generalization, for which the goal is to perform well on possible unseen domains after fine-tuning on multiple training domains. However, maximally exploiting a zoo of PTMs is challenging since fine-tuning all possible combinations of PTMs is computationally prohibitive while accurate selection of PTMs requires tackling the possible data distribution shift for OoD tasks. In this work, we propose ZooD, a paradigm for PTMs ranking and ensemble with feature selection. Our proposed metric ranks PTMs by quantifying inter-class discriminability and inter-domain stability of the features extracted by the PTMs in a leave-one-domain-out cross-validation manner. The top-K ranked models are then aggregated for the target OoD task. To avoid accumulating noise induced by model ensemble, we propose an efficient variational EM algorithm to select informative features. We evaluate our paradigm on a diverse model zoo consisting of 35 models for various OoD tasks and demonstrate: (i) model ranking is better correlated with fine-tuning ranking than previous methods and up to 9859x faster than brute-force fine-tuning; (ii) OoD generalization after model ensemble with feature selection outperforms the state-of-the-art methods and the accuracy on most challenging task DomainNet is improved from 46.5\% to 50.6\%. Furthermore, we provide the fine-tuning results of 35 PTMs on 7 OoD datasets, hoping to help the research of model zoo and OoD generalization. Code will be available at https://gitee.com/mindspore/models/tree/master/research/cv/zood.

A multitude of work has shown that machine learning-based medical diagnosis systems can be biased against certain subgroups of people. This has motivated a growing number of bias mitigation algorithms that aim to address fairness issues in machine learning. However, it is difficult to compare their effectiveness in medical imaging for two reasons. First, there is little consensus on the criteria to assess fairness. Second, existing bias mitigation algorithms are developed under different settings, e.g., datasets, model selection strategies, backbones, and fairness metrics, making a direct comparison and evaluation based on existing results impossible. In this work, we introduce MEDFAIR, a framework to benchmark the fairness of machine learning models for medical imaging. MEDFAIR covers eleven algorithms from various categories, nine datasets from different imaging modalities, and three model selection criteria. Through extensive experiments, we find that the under-studied issue of model selection criterion can have a significant impact on fairness outcomes; while in contrast, state-of-the-art bias mitigation algorithms do not significantly improve fairness outcomes over empirical risk minimization (ERM) in both in-distribution and out-of-distribution settings. We evaluate fairness from various perspectives and make recommendations for different medical application scenarios that require different ethical principles. Our framework provides a reproducible and easy-to-use entry point for the development and evaluation of future bias mitigation algorithms in deep learning. Code is available at https://github.com/ys-zong/MEDFAIR., Comment: Accepted to ICLR 2023

Additional file 13: Figure S8. The Distribution of the Predicted Probabilities in Non-carriers and CPGs-carriers by the DrABC Model.

Additional file 10: Figure S7. The Performance of the DrABC Model and Other Machine Learning Models Using an Inner Five-fold Cross-validation Strategy.

Additional file 5: Figure S3. Phenotype-genotype correlation and data interpretation.

Additional file 2: Figure S1. Summary of Variants in Cancer Predisposition Genes.

Additional file 9: Figure S6. Developing the DrABC Model through the Hierarchical Neural Network.

Additional file 14: Figure S9. Performance of Risk Prediction Models for Breast Cancer Patients with germline Pathogenic Variants in Cancer Predisposition Genes other than BRCA1/2.

Additional file 20. A user guide for the DrABC model.

Additional file 7: Figure S5. The Multi-center Validation of the DrABC Model.

Additional file 4: Figure S2. Association of Germline Variants with Age at Diagnosis (A), Family History (B), Histological Grade (C), and Molecular Subtypes (D).

The high surface roughness of dip polydopamine (dPDA) coating on magnesium alloy could decrease the quality of secondary modification. In the present study, we report an electropolymerized dopamine (ePDA) coating on Mg-Zn-Y-Nd alloy from sodium salicylate aqueous solutions by galvanostat method to address this problem. The coating formation process and coating structure were investigated. Then, a Poly ( l -lactic acid) (PLLA) coating was prepared on ePDA and dPDA treated samples. The SEM results reveal that the ePDA film appeared to be more uniform than dPDA coating and was doped with salicylate and magnesium ions. ASTM D3359 -09 “Tape test” shows excellent adhesion of the ePDA/PLLA layer. The electrochemical impedance spectroscopy and polarization curves demonstrate ePDA/PLLA film has a better corrosion resistance than dPDA/PLLA film, owing to the structural difference of polydopamine anchoring layer.

Gradient-based meta-learning and hyperparameter optimization have seen significant progress recently, enabling practical end-to-end training of neural networks together with many hyperparameters. Nevertheless, existing approaches are relatively expensive as they need to compute second-order derivatives and store a longer computational graph. This cost prevents scaling them to larger network architectures. We present EvoGrad, a new approach to meta-learning that draws upon evolutionary techniques to more efficiently compute hypergradients. EvoGrad estimates hypergradient with respect to hyperparameters without calculating second-order gradients, or storing a longer computational graph, leading to significant improvements in efficiency. We evaluate EvoGrad on three substantial recent meta-learning applications, namely cross-domain few-shot learning with feature-wise transformations, noisy label learning with Meta-Weight-Net and low-resource cross-lingual learning with meta representation transformation. The results show that EvoGrad significantly improves efficiency and enables scaling meta-learning to bigger architectures such as from ResNet10 to ResNet34., Comment: Accepted at NeurIPS 2021

Convolutional neural networks (CNNs) often have poor generalization performance under domain shift. One way to improve domain generalization is to collect diverse source data from multiple relevant domains so that a CNN model is allowed to learn more domain-invariant, and hence generalizable representations. In this work, we address domain generalization with MixStyle, a plug-and-play, parameter-free module that is simply inserted to shallow CNN layers and requires no modification to training objectives. Specifically, MixStyle probabilistically mixes feature statistics between instances. This idea is inspired by the observation that visual domains can often be characterized by image styles which are in turn encapsulated within instance-level feature statistics in shallow CNN layers. Therefore, inserting MixStyle modules in effect synthesizes novel domains albeit in an implicit way. MixStyle is not only simple and flexible, but also versatile -- it can be used for problems whereby unlabeled images are available, such as semi-supervised domain generalization and unsupervised domain adaptation, with a simple extension to mix feature statistics between labeled and pseudo-labeled instances. We demonstrate through extensive experiments that MixStyle can significantly boost the out-of-distribution generalization performance across a wide range of tasks including object recognition, instance retrieval, and reinforcement learning., Comment: Extension of https://openreview.net/forum?id=6xHJ37MVxxp. Code available at https://github.com/KaiyangZhou/mixstyle-release

This paper is concerned with contrastive learning (CL) for low-level image restoration and enhancement tasks. We propose a new label-efficient learning paradigm based on residuals, residual contrastive learning (RCL), and derive an unsupervised visual representation learning framework, suitable for low-level vision tasks with noisy inputs. While supervised image reconstruction aims to minimize residual terms directly, RCL alternatively builds a connection between residuals and CL by defining a novel instance discrimination pretext task, using residuals as the discriminative feature. Our formulation mitigates the severe task misalignment between instance discrimination pretext tasks and downstream image reconstruction tasks, present in existing CL frameworks. Experimentally, we find that RCL can learn robust and transferable representations that improve the performance of various downstream tasks, such as denoising and super resolution, in comparison with recent self-supervised methods designed specifically for noisy inputs. Additionally, our unsupervised pre-training can significantly reduce annotation costs whilst maintaining performance competitive with fully-supervised image reconstruction., Comment: Accepted by IJCAI 2022

Most existing multi-source domain adaptation (MSDA) methods minimize the distance between multiple source-target domain pairs via feature distribution alignment, an approach borrowed from the single source setting. However, with diverse source domains, aligning pairwise feature distributions is challenging and could even be counter-productive for MSDA. In this paper, we introduce a novel approach: transferable attribute learning. The motivation is simple: although different domains can have drastically different visual appearances, they contain the same set of classes characterized by the same set of attributes; an MSDA model thus should focus on learning the most transferable attributes for the target domain. Adopting this approach, we propose a domain attention consistency network, dubbed DAC-Net. The key design is a feature channel attention module, which aims to identify transferable features (attributes). Importantly, the attention module is supervised by a consistency loss, which is imposed on the distributions of channel attention weights between source and target domains. Moreover, to facilitate discriminative feature learning on the target data, we combine pseudo-labeling with a class compactness loss to minimize the distance between the target features and the classifier's weight vectors. Extensive experiments on three MSDA benchmarks show that our DAC-Net achieves new state of the art performance on all of them., Comment: Accepted to BMVC 2021 as oral presentation

Data augmentation (DA) techniques aim to increase data variability, and thus train deep networks with better generalisation. The pioneering AutoAugment automated the search for optimal DA policies with reinforcement learning. However, AutoAugment is extremely computationally expensive, limiting its wide applicability. Followup works such as Population Based Augmentation (PBA) and Fast AutoAugment improved efficiency, but their optimization speed remains a bottleneck. In this paper, we propose Differentiable Automatic Data Augmentation (DADA) which dramatically reduces the cost. DADA relaxes the discrete DA policy selection to a differentiable optimization problem via Gumbel-Softmax. In addition, we introduce an unbiased gradient estimator, RELAX, leading to an efficient and effective one-pass optimization strategy to learn an efficient and accurate DA policy. We conduct extensive experiments on CIFAR-10, CIFAR-100, SVHN, and ImageNet datasets. Furthermore, we demonstrate the value of Auto DA in pre-training for downstream detection problems. Results show our DADA is at least one order of magnitude faster than the state-of-the-art while achieving very comparable accuracy. The code is available at https://github.com/VDIGPKU/DADA.

The study of neural generative models of human sketches is a fascinating contemporary modeling problem due to the links between sketch image generation and the human drawing process. The landmark SketchRNN provided breakthrough by sequentially generating sketches as a sequence of waypoints. However this leads to low-resolution image generation, and failure to model long sketches. In this paper we present B��zierSketch, a novel generative model for fully vector sketches that are automatically scalable and high-resolution. To this end, we first introduce a novel inverse graphics approach to stroke embedding that trains an encoder to embed each stroke to its best fit B��zier curve. This enables us to treat sketches as short sequences of paramaterized strokes and thus train a recurrent sketch generator with greater capacity for longer sketches, while producing scalable high-resolution results. We report qualitative and quantitative results on the Quick, Draw! benchmark., Accepted as poster at ECCV 2020

Off-Policy Actor-Critic (Off-PAC) methods have proven successful in a variety of continuous control tasks. Normally, the critic's action-value function is updated using temporal-difference, and the critic in turn provides a loss for the actor that trains it to take actions with higher expected return. In this paper, we introduce a novel and flexible meta-critic that observes the learning process and meta-learns an additional loss for the actor that accelerates and improves actor-critic learning. Compared to the vanilla critic, the meta-critic network is explicitly trained to accelerate the learning process; and compared to existing meta-learning algorithms, meta-critic is rapidly learned online for a single task, rather than slowly over a family of tasks. Crucially, our meta-critic framework is designed for off-policy based learners, which currently provide state-of-the-art reinforcement learning sample efficiency. We demonstrate that online meta-critic learning leads to improvements in avariety of continuous control environments when combined with contemporary Off-PAC methods DDPG, TD3 and the state-of-the-art SAC., Comment: NeurIPS 2020

Sketch as an image search query is an ideal alternative to text in capturing the fine-grained visual details. Prior successes on fine-grained sketch-based image retrieval (FG-SBIR) have demonstrated the importance of tackling the unique traits of sketches as opposed to photos, e.g., temporal vs. static, strokes vs. pixels, and abstract vs. pixel-perfect. In this paper, we study a further trait of sketches that has been overlooked to date, that is, they are hierarchical in terms of the levels of detail -- a person typically sketches up to various extents of detail to depict an object. This hierarchical structure is often visually distinct. In this paper, we design a novel network that is capable of cultivating sketch-specific hierarchies and exploiting them to match sketch with photo at corresponding hierarchical levels. In particular, features from a sketch and a photo are enriched using cross-modal co-attention, coupled with hierarchical node fusion at every level to form a better embedding space to conduct retrieval. Experiments on common benchmarks show our method to outperform state-of-the-arts by a significant margin., Comment: Accepted for ORAL presentation in BMVC 2020

In vitro degradation is an important approach to screening appropriate biomedical magnesium (Mg) alloys at low cost. However, corrosion products deposited on Mg alloys exert a critical impact on corrosion resistance. There are no acceptable criteria on the evaluation on degradation rate of Mg alloys. Understanding the degradation behavior of Mg alloys in presence of Tris buffer is necessary. An investigation was made to compare the influence of Tris-HCl and Tris on the corrosion behavior of Mg alloy AZ31 in the presence of various anions of simulated body fluids via hydrogen evolution, pH value and electrochemical tests. The results demonstrated that the Tris-HCl buffer resulted in general corrosion due to the inhibition of the formation of corrosion products and thus increased the corrosion rate of the AZ31 alloy. Whereas Tris gave rise to pitting corrosion or general corrosion due to the fact that the hydrolysis of the amino-group of Tris led to an increase in solution pH value, and promoted the formation of corrosion products and thus a significant reduction in corrosion rate. In addition, the corrosion mechanisms in the presence of Tris-HCl and Tris were proposed. Tris-HCl as a buffer prevented the formation of precipitates of HCO3−, SO42−, HPO42− and H2PO4− ions during the corrosion of the AZ31 alloy due to its lower buffering pH value (x.x). Thus, both the hydrogen evolution rate and corrosion current density of the alloy were approximately one order of magnitude higher in presence of Tris-HCl than Tris and Tris-free saline solutions.

Based on the physical phenomenon of the fiber reinforced polymer material after the erosion of the liquid medium, the erosion law of the liquid medium on the FRP is analyzed. And considering the anisotropy of the FRP material, a FRP point source erosion depth model based on Fick’s law is proposed. Taking the glass fiber reinforced polymer (GFRP) in alkali solution as an example, we carried out GFRP corrosion and aging test in alkaline solution, and the erosion depth of GFRP specimens was measured at 30, 90, 180, 360 and 720 days, respectively. At the same time, the water absorption rate of GFRP was used to calculate the erosion diffusion coefficient of solution medium. Through the comparison between the experimental value of erosion depth and the calculated value of erosion depth model, the error between the experimental data and the calculated data is within 10%, and the overall regularity is consistent. It shows that the erosion depth model of liquid medium proposed in this paper is reasonable and has high accuracy.

Deep neural networks have been proven powerful at processing perceptual data, such as images and audio. However for tabular data, tree-based models are more popular. A nice property of tree-based models is their natural interpretability. In this work, we present Deep Neural Decision Trees (DNDT) -- tree models realised by neural networks. A DNDT is intrinsically interpretable, as it is a tree. Yet as it is also a neural network (NN), it can be easily implemented in NN toolkits, and trained with gradient descent rather than greedy splitting. We evaluate DNDT on several tabular datasets, verify its efficacy, and investigate similarities and differences between DNDT and vanilla decision trees. Interestingly, DNDT self-prunes at both split and feature-level., Comment: presented at 2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018), Stockholm, Sweden

We propose to model complex visual scenes using a non-parametric Bayesian model learned from weakly labelled images abundant on media sharing sites such as Flickr. Given weak image-level annotations of objects and attributes without locations or associations between them, our model aims to learn the appearance of object and attribute classes as well as their association on each object instance. Once learned, given an image, our model can be deployed to tackle a number of vision problems in a joint and coherent manner, including recognising objects in the scene (automatic object annotation), describing objects using their attributes (attribute prediction and association), and localising and delineating the objects (object detection and semantic segmentation). This is achieved by developing a novel Weakly Supervised Markov Random Field Stacked Indian Buffet Process (WS-MRF-SIBP) that models objects and attributes as latent factors and explicitly captures their correlations within and across superpixels. Extensive experiments on benchmark datasets demonstrate that our weakly supervised model significantly outperforms weakly supervised alternatives and is often comparable with existing strongly supervised models on a variety of tasks including semantic segmentation, automatic image annotation and retrieval based on object-attribute associations., Comment: Accepted in IEEE Transaction on Pattern Analysis and Machine Intelligence

Generating natural language descriptions of images is an important capability for a robot or other visual-intelligence driven AI agent that may need to communicate with human users about what it is seeing. Such image captioning methods are typically trained by maximising the likelihood of ground-truth annotated caption given the image. While simple and easy to implement, this approach does not directly maximise the language quality metrics we care about such as CIDEr. In this paper we investigate training image captioning methods based on actor-critic reinforcement learning in order to directly optimise non-differentiable quality metrics of interest. By formulating a per-token advantage and value computation strategy in this novel reinforcement learning based captioning model, we show that it is possible to achieve the state of the art performance on the widely used MSCOCO benchmark.

We propose a novel and flexible approach to meta-learning for learning-to-learn from only a few examples. Our framework is motivated by actor-critic reinforcement learning, but can be applied to both reinforcement and supervised learning. The key idea is to learn a meta-critic: an action-value function neural network that learns to criticise any actor trying to solve any specified task. For supervised learning, this corresponds to the novel idea of a trainable task-parametrised loss generator. This meta-critic approach provides a route to knowledge transfer that can flexibly deal with few-shot and semi-supervised conditions for both reinforcement and supervised learning. Promising results are shown on both reinforcement and supervised learning problems., Comment: Technical report, 12 pages, 3 figures, 2 tables

We propose a framework for training multiple neural networks simultaneously. The parameters from all models are regularised by the tensor trace norm, so that each neural network is encouraged to reuse others' parameters if possible -- this is the main motivation behind multi-task learning. In contrast to many deep multi-task learning models, we do not predefine a parameter sharing strategy by specifying which layers have tied parameters. Instead, our framework considers sharing for all shareable layers, and the sharing strategy is learned in a data-driven way., Comment: Submission to Workshop track - ICLR 2017

Most contemporary multi-task learning methods assume linear models. This setting is considered shallow in the era of deep learning. In this paper, we present a new deep multi-task representation learning framework that learns cross-task sharing structure at every layer in a deep network. Our approach is based on generalising the matrix factorisation techniques explicitly or implicitly used by many conventional MTL algorithms to tensor factorisation, to realise automatic learning of end-to-end knowledge sharing in deep networks. This is in contrast to existing deep learning approaches that need a user-defined multi-task sharing strategy. Our approach applies to both homogeneous and heterogeneous MTL. Experiments demonstrate the efficacy of our deep multi-task representation learning in terms of both higher accuracy and fewer design choices., Comment: 9 pages, Accepted to ICLR 2017 Conference Track. This is a conference version of the paper. For the multi-domain learning part (not in this version), please refer to https://arxiv.org/pdf/1605.06391v1.pdf