Conventional unsupervised multi-source domain adaptation(UMDA) methods assume all source domains can be accessed directly. This neglects the privacy-preserving policy, that is,all the data and computations must be kept decentralized.There exists three problems in this scenario: (1)Minimizing the domain distance requires the pairwise calculation of the data from source and target domains, which is not accessible.(2)The communication cost and privacy security limit the application of UMDA methods (e.g.,the domain adversarial training).(3)Since users have no authority to checkthe data quality, the irrelevant or malicious source domainsare more likely to appear, which causes negative transfer. In this study, we propose a privacy-preserving UMDA paradigm named Knowledge Distillation based Decentralized Domain Adaptation(KD3A), which performs domain adaptation through the knowledge distillation on models from different source domains. KD3A solves the above problems with three components:(1)A multi-source knowledge distillation method named Knowledge Voteto learn high-quality domain consensus knowledge. (2)A dynamic weighting strategy named Consensus Focusto identify both the malicious and irrelevant domains.(3)A decentralized optimization strategy for computing domain distance named BatchNorm MMD.The extensive experiments on DomainNet demonstrate that KD3A is robust to the negative transfer and brings a 100x reduction of communication cost compared with other decentralized UMDA methods. Moreover, our KD3A significantly outperforms state-of-the-art UMDA approaches.
Any prediction from a model is made by a combination of learning history and test stimuli. This provides significant insights for improving model interpretability: {\it because of which part(s) of which training example(s), the model attends to which part(s) of a test example}. Unfortunately, existing methods to interpret a model's predictions are only able to capture a single aspect of either test stimuli or learning history, and evidences from both are never combined or integrated. In this paper, we propose an efficient and differentiable approach to make it feasible to interpret a model's prediction by jointly examining training history and test stimuli. Test stimuli is first identified by gradient-based methods, signifying {\it the part of a test example that the model attends to}. The gradient-based saliency scores are then propagated to training examples using influence functions to identify {\it which part(s) of which training example(s)} make the model attends to the test stimuli. The system is differentiable and time efficient: the adoption of saliency scores from gradient-based methods allows us to efficiently trace a model's prediction through test stimuli, and then back to training examples through influence functions. We demonstrate that the proposed methodology offers clear explanations about neural model decisions, along with being useful for performing error analysis, crafting adversarial examples and fixing erroneously classified examples.
Long-text generation remains a challenge. The difficulty of generating coherent long texts lies in the fact that existing models overwhelmingly focus on the tasks of local word prediction, and cannot make high level plans on what to generate or capture the high-level discourse dependencies between chunks of texts. Inspired by how humans write, where a list of bullet points or a catalog is first outlined, and then each bullet point is expanded to form the whole article, we propose {\it SOE}, a pipelined system that involves of summarizing, outlining and elaborating for long text generation: the model first outlines the summaries for different segments of long texts, and then elaborates on each bullet point to generate the corresponding segment. To avoid the labor-intensive process of summary soliciting, we propose the {\it reconstruction} strategy, which extracts segment summaries in an unsupervised manner by selecting its most informative part to reconstruct the segment.The proposed generation system comes with the following merits: (1) the summary provides high-level guidances for text generation and avoids the local minimum of individual word predictions; (2) the high-level discourse dependencies are captured in the conditional dependencies between summaries and are preserved during the summary expansion process and (3) additionally, we are able to consider significantly more contexts by representing contexts as concise summaries. Extensive experiments demonstrate that SOE produces long texts with significantly better quality, along with faster convergence speed.
In this paper, we investigate the problem of text-to-pedestrian synthesis, which has many potential applications in art, design, and video surveillance. Existing methods for text-to-bird/flower synthesis are still far from solving this fine-grained image generation problem, due to the complex structure and heterogeneous appearance that the pedestrians naturally take on. To this end, we propose the Multi-Grained Discrimination enhanced Generative Adversarial Network, that capitalizes a human-part-based Discriminator (HPD) and a self-cross-attended (SCA) global Discriminator in order to capture the coherence of the complex body structure. A fined-grained word-level attention mechanism is employed in the HPD module to enforce diversified appearance and vivid details. In addition, two pedestrian generation metrics, named Pose Score and Pose Variance, are devised to evaluate the generation quality and diversity, respectively. We conduct extensive experiments and ablation studies on the caption-annotated pedestrian dataset, CUHK Person Description Dataset. The substantial improvement over the various metrics demonstrates the efficacy of MGD-GAN on the text-to-pedestrian synthesis scenario.
Supervised neural networks, which first map an input $x$ to a single representation $z$, and then map $z$ to the output label $y$, have achieved remarkable success in a wide range of natural language processing (NLP) tasks. Despite their success, neural models lack for both robustness and generality: small perturbations to inputs can result in absolutely different outputs; the performance of a model trained on one domain drops drastically when tested on another domain. In this paper, we present methods to improve robustness and generality of NLP models from the standpoint of disentangled representation learning. Instead of mapping $x$ to a single representation $z$, the proposed strategy maps $x$ to a set of representations $\{z_1,z_2,...,z_K\}$ while forcing them to be disentangled. These representations are then mapped to different logits $l$s, the ensemble of which is used to make the final prediction $y$. We propose different methods to incorporate this idea into currently widely-used models, including adding an $L$2 regularizer on $z$s or adding Total Correlation (TC) under the framework of variational information bottleneck (VIB). We show that models trained with the proposed criteria provide better robustness and domain adaptation ability in a wide range of supervised learning tasks.
When patients need to take medicine, particularly taking more than one kind of drug simultaneously, they should be alarmed that there possibly exists drug-drug interaction. Interaction between drugs may have a negative impact on patients or even cause death. Generally, drugs that conflict with a specific drug (or label drug) are usually described in its drug label or package insert. Since more and more new drug products come into the market, it is difficult to collect such information by manual. We take part in the Drug-Drug Interaction (DDI) Extraction from Drug Labels challenge of Text Analysis Conference (TAC) 2018, choosing task1 and task2 to automatically extract DDI related mentions and DDI relations respectively. Instead of regarding task1 as named entity recognition (NER) task and regarding task2 as relation extraction (RE) task then solving it in a pipeline, we propose a two step joint model to detect DDI and it's related mentions jointly. A sequence tagging system (CNN-GRU encoder-decoder) finds precipitants first and search its fine-grained Trigger and determine the DDI for each precipitant in the second step. Moreover, a rule based model is built to determine the sub-type for pharmacokinetic interation. Our system achieved best result in both task1 and task2. F-measure reaches 0.46 in task1 and 0.40 in task2.
In this paper, we propose to investigate the problem of out-of-domain visio-linguistic pretraining, where the pretraining data distribution differs from that of downstream data on which the pretrained model will be fine-tuned. Existing methods for this problem are purely likelihood-based, leading to the spurious correlations and hurt the generalization ability when transferred to out-of-domain downstream tasks. By spurious correlation, we mean that the conditional probability of one token (object or word) given another one can be high (due to the dataset biases) without robust (causal) relationships between them. To mitigate such dataset biases, we propose a Deconfounded Visio-Linguistic Bert framework, abbreviated as DeVLBert, to perform intervention-based learning. We borrow the idea of the backdoor adjustment from the research field of causality and propose several neural-network based architectures for Bert-style out-of-domain pretraining. The quantitative results on three downstream tasks, Image Retrieval (IR), Zero-shot IR, and Visual Question Answering, show the effectiveness of DeVLBert by boosting generalization ability.
In e-commerce, a growing number of user-generated videos are used for product promotion. How to generate video descriptions that narrate the user-preferred product characteristics depicted in the video is vital for successful promoting. Traditional video captioning methods, which focus on routinely describing what exists and happens in a video, are not amenable for product-oriented video captioning. To address this problem, we propose a product-oriented video captioner framework, abbreviated as Poet. Poet firstly represents the videos as product-oriented spatial-temporal graphs. Then, based on the aspects of the video-associated product, we perform knowledge-enhanced spatial-temporal inference on those graphs for capturing the dynamic change of fine-grained product-part characteristics. The knowledge leveraging module in Poet differs from the traditional design by performing knowledge filtering and dynamic memory modeling. We show that Poet achieves consistent performance improvement over previous methods concerning generation quality, product aspects capturing, and lexical diversity. Experiments are performed on two product-oriented video captioning datasets, buyer-generated fashion video dataset (BFVD) and fan-generated fashion video dataset (FFVD), collected from Mobile Taobao. We will release the desensitized datasets to promote further investigations on both video captioning and general video analysis problems.
Synthesizing high-resolution realistic images from text descriptions is a challenging task. Almost all existing text-to-image methods employ stacked generative adversarial networks as the backbone, utilize cross-modal attention mechanisms to fuse text and image features, and use extra networks to ensure text-image semantic consistency. The existing text-to-image models have three problems: 1) For the backbone, there are multiple generators and discriminators stacked for generating different scales of images making the training process slow and inefficient. 2) For semantic consistency, the existing models employ extra networks to ensure the semantic consistency increasing the training complexity and bringing an additional computational cost. 3) For the text-image feature fusion method, cross-modal attention is only applied a few times during the generation process due to its computational cost impeding fusing the text and image features deeply. To solve these limitations, we propose 1) a novel simplified text-to-image backbone which is able to synthesize high-quality images directly by one pair of generator and discriminator, 2) a novel regularization method called Matching-Aware zero-centered Gradient Penalty which promotes the generator to synthesize more realistic and text-image semantic consistent images without introducing extra networks, 3) a novel fusion module called Deep Text-Image Fusion Block which can exploit the semantics of text descriptions effectively and fuse text and image features deeply during the generation process. Compared with the previous text-to-image models, our DF-GAN is simpler and more efficient and achieves better performance. Extensive experiments and ablation studies on both Caltech-UCSD Birds 200 and COCO datasets demonstrate the superiority of the proposed model in comparison to state-of-the-art models.