Since the milestone study by Alex Krizhevsky and colleagues in 2012 [1], Deep Learning (DL), with particular emphasis on Convolutional Neural Networks (CNNs), is the state-of-the-art method for image classification in many different applications. Besides classification performance, the reason for the success of CNNs is twofold: i) the recent boost of graphical processing units (GPUs) and parallel processing, that allows to train very large models; ii) the ever-growing availability of massive annotated task-specific datasets. Nonetheless, in many realistic applications many concerns may be raised about the reliability of such datasets both in terms of image and labelling quality, and consequently on the robustness of the CNN models trained and tested on them. As regards to image quality, standard CNNs are supposed to be fed with high quality samples. Nevertheless, in practical scenarios different kinds of image degradation can heavily affect the performance of a CNN both in the training and in the inference phase. Problems concerning image acquisition devices, poor image sensor, lighting conditions, focus, stabilization, exposure time or partial occlusion of the cameras may lead to produce low quality samples, which have been demonstrated to be one of the chief reasons for troublesome learning procedures of CNN models in many applications [2, 3, 4]. On the other hand, even though the CNN had been proficiently trained and validated on high quality data, noisy inputs can heavily affect the inference phase, and cause classification errors at run-time. A typical example are self-driving cars, where a partial occlusion of the image acquisition device may lead to misinterpret a road sign, with catastrophic consequences. In such settings, the well-known limitations of standard CNNs to broadcast information about how much the given input resembles the ones the model was trained on - and hence, whether the associated prediction should (or should not) be trusted - is also playing a major role.

Mixup-based data augmentation has achieved great success as regularizer for deep neural networks. However, existing mixup methods require explicitly designed mixup policies. In this paper, we present a flexible, general Automatic Mixup (AutoMix) framework which utilizes discriminative features to learn a sample mixing policy adaptively. We regard mixup as a pretext task and split it into two sub-problems: mixed samples generation and mixup classification. To this end, we design a lightweight mix block to generate synthetic samples based on feature maps and mix labels. Since the two sub-problems are in the nature of Expectation-Maximization (EM), we also propose a momentum training pipeline to optimize the mixup process and mixup classification process alternatively in an end-to-end fashion. Extensive experiments on six popular classification benchmarks show that AutoMix consistently outperforms other leading mixup methods and improves generalization abilities to downstream tasks. We hope AutoMix will motivate the community to rethink the role of mixup in representation learning. The code will be released soon.

The paper considers a distributed version of deep reinforcement learning (DRL) for multi-agent decision-making process in the paradigm of federated learning. Since the deep neural network models in federated learning are trained locally and aggregated iteratively through a central server, frequent information exchange incurs a large amount of communication overheads. Besides, due to the heterogeneity of agents, Markov state transition trajectories from different agents are usually unsynchronized within the same time interval, which will further influence the convergence bound of the aggregated deep neural network models. Therefore, it is of vital importance to reasonably evaluate the effectiveness of different optimization methods. Accordingly, this paper proposes a utility function to consider the balance between reducing communication overheads and improving convergence performance. Meanwhile, this paper develops two new optimization methods on top of variation-aware periodic averaging methods: 1) the decay-based method which gradually decreases the weight of the model's local gradients within the progress of local updating, and 2) the consensus-based method which introduces the consensus algorithm into federated learning for the exchange of the model's local gradients. This paper also provides novel convergence guarantees for both developed methods and demonstrates their effectiveness and efficiency through theoretical analysis and numerical simulation results.

With the rapid increase in the amount of public code repositories, developers maintain a great desire to retrieve precise code snippets by using natural language. Despite existing deep learning based approaches(e.g., DeepCS and MMAN) have provided the end-to-end solutions (i.e., accepts natural language as queries and shows related code fragments retrieved directly from code corpus), the accuracy of code search in the large-scale repositories is still limited by the code representation (e.g., AST) and modeling (e.g., directly fusing the features in the attention stage). In this paper, we propose a novel learnable deep Graph for Code Search (calleddeGraphCS), to transfer source code into variable-based flow graphs based on the intermediate representation technique, which can model code semantics more precisely compared to process the code as text directly or use the syntactic tree representation. Furthermore, we propose a well-designed graph optimization mechanism to refine the code representation, and apply an improved gated graph neural network to model variable-based flow graphs. To evaluate the effectiveness of deGraphCS, we collect a large-scale dataset from GitHub containing 41,152 code snippets written in C language, and reproduce several typical deep code search methods for comparison. Besides, we design a qualitative user study to verify the practical value of our approach. The experimental results have shown that deGraphCS can achieve state-of-the-art performances, and accurately retrieve code snippets satisfying the needs of the users.

Digitalization leads to data transparency for production systems that we can benefit from with data-driven analysis methods like neural networks. For example, automated anomaly detection enables saving resources and optimizing the production. We study using rarely occurring information about labeled anomalies into Variational Autoencoder neural network structures to overcome information deficits of supervised and unsupervised approaches. This method outperforms all other models in terms of accuracy, precision, and recall. We evaluate the following methods: Principal Component Analysis, Isolation Forest, Classifying Neural Networks, and Variational Autoencoders on seven time series datasets to find the best performing detection methods. We extend this idea to include more infrequently occurring meta information about production processes. This use of sparse labels, both of anomalies or production data, allows to harness any additional information available for increasing anomaly detection performance.

Motivation: Several accurate deep learning models have been proposed to predict drug-target affinity (DTA). However, all of these models are black box hence are difficult to interpret and verify its result, and thus risking acceptance. Explanation is necessary to allow the DTA model more trustworthy. Explanation with counterfactual provides human-understandable examples. Most counterfactual explanation methods only operate on single input data, which are in tabular or continuous forms. In contrast, the DTA model has two discrete inputs. It is challenging for the counterfactual generation framework to optimize both discrete inputs at the same time. Results: We propose a multi-agent reinforcement learning framework, Multi-Agent Counterfactual Drug-target binding Affinity (MACDA), to generate counterfactual explanations for the drug-protein complex. Our proposed framework provides human-interpretable counterfactual instances while optimizing both the input drug and target for counterfactual generation at the same time. The result on the Davis dataset shows the advantages of the proposed MACDA framework compared with previous works.

We introduce deep learning models to the two most important stages in product search at JD.com, one of the largest e-commerce platforms in the world. Specifically, we outline the design of a deep learning system that retrieves semantically relevant items to a query within milliseconds, and a pairwise deep re-ranking system, which learns subtle user preferences. Compared to traditional search systems, the proposed approaches are better at semantic retrieval and personalized ranking, achieving significant improvements.

Cognitive grammar suggests that the acquisition of language grammar is grounded within visual structures. While grammar is an essential representation of natural language, it also exists ubiquitously in vision to represent the hierarchical part-whole structure. In this work, we study grounded grammar induction of vision and language in a joint learning framework. Specifically, we present VLGrammar, a method that uses compound probabilistic context-free grammars (compound PCFGs) to induce the language grammar and the image grammar simultaneously. We propose a novel contrastive learning framework to guide the joint learning of both modules. To provide a benchmark for the grounded grammar induction task, we collect a large-scale dataset, \textsc{PartIt}, which contains human-written sentences that describe part-level semantics for 3D objects. Experiments on the \textsc{PartIt} dataset show that VLGrammar outperforms all baselines in image grammar induction and language grammar induction. The learned VLGrammar naturally benefits related downstream tasks. Specifically, it improves the image unsupervised clustering accuracy by 30\%, and performs well in image retrieval and text retrieval. Notably, the induced grammar shows superior generalizability by easily generalizing to unseen categories.

Predicting multiple plausible future trajectories of the nearby vehicles is crucial for the safety of autonomous driving. Recent motion prediction approaches attempt to achieve such multimodal motion prediction by implicitly regularizing the feature or explicitly generating multiple candidate proposals. However, it remains challenging since the latent features may concentrate on the most frequent mode of the data while the proposal-based methods depend largely on the prior knowledge to generate and select the proposals. In this work, we propose a novel transformer framework for multimodal motion prediction, termed as mmTransformer. A novel network architecture based on stacked transformers is designed to model the multimodality at feature level with a set of fixed independent proposals. A region-based training strategy is then developed to induce the multimodality of the generated proposals. Experiments on Argoverse dataset show that the proposed model achieves the state-of-the-art performance on motion prediction, substantially improving the diversity and the accuracy of the predicted trajectories. Demo video and code are available at https://decisionforce.github.io/mmTransformer.

We investigate the problem of zero-shot semantic image painting. Instead of painting modifications into an image using only concrete colors or a finite set of semantic concepts, we ask how to create semantic paint based on open full-text descriptions: our goal is to be able to point to a location in a synthesized image and apply an arbitrary new concept such as "rustic" or "opulent" or "happy dog." To do this, our method combines a state-of-the art generative model of realistic images with a state-of-the-art text-image semantic similarity network. We find that, to make large changes, it is important to use non-gradient methods to explore latent space, and it is important to relax the computations of the GAN to target changes to a specific region. We conduct user studies to compare our methods to several baselines.