In order to gain insights about the decision-making of different visual recognition backbones, we propose two methodologies, sub-explanation counting and cross-testing, that systematically applies deep explanation algorithms on a dataset-wide basis, and compares the statistics generated from the amount and nature of the explanations. These methodologies reveal the difference among networks in terms of two properties called compositionality and disjunetivism. Transformers and ConvNeXt are found to be more compositional, in the sense that they jointly consider multiple parts of the image in building their decisions, whereas traditional CNNs and distilled transformers are less compositional and more disjunctive, which means that they use multiple diverse but smaller set of parts to achieve a confident prediction. Through further experiments, we pin-pointed the choice of normalization to be especially important in the compositionality of a model, in that batch normalization leads to less compositionality while group and layer normalization lead to more. Finally, we also analyze the features shared by different backbones and plot a land-scape of different models based on their feature-use similarity.

Despite extensive research on training generative adversarial networks (GANs) with limited training data, learning to generate images from long-tailed training distributions remains fairly unexplored. In the presence of imbalanced multiclass training data, GANs tend to favor classes with more samples, leading to the generation of low quality and less diverse samples in tail classes. In this study, we aim to improve the training of class-conditional GANs with long-tailed data. We propose a straightforward yet effective method for knowledge sharing, allowing tail classes to borrow from the rich information from classes with more abundant training data. More concretely, we propose modifications to existing class-conditional GAN architectures to ensure that the lower-resolution layers of the generator are trained entirely unconditionally while reserving class-conditional generation for the higher-resolution layers. Experiments on several long-tail benchmarks and GAN architectures demonstrate a significant improvement over existing methods in both the diversity and fidelity of the generated images. The code is available at https:github.com/khorrams/utlo.

Cross-speaker emotion transfer speech synthesis aims to synthesize emotional speech for a target speaker by transferring the emotion from reference speech recorded by another (source) speaker.In this task, extracting speaker-independent emotion embedding from reference speech plays an important role.However, the emotional information conveyed by such emotion embedding tends to be weakened in the process to squeeze out the source speaker's timbre information.In response to this problem, a prosody compensation module (PCM) is proposed in this paper to compensate for the emotional information loss.Specifically, the PCM tries to obtain speaker-independent emotional information from the intermediate feature of a pretrained ASR model.To this end, a prosody compensation encoder with global context (GC) blocks is introduced to obtain global emotional information from the ASR model's intermediate feature.Experiments demonstrate that the proposed PCM can effectively compensate the emotion embedding for the emotional information loss, and meanwhile maintain the timbre of the target speaker.Comparisons with state-of-the-art models show that our proposed method presents obvious superiority on the cross-speaker emotion transfer task.

In realistic scenarios, a user profiling model (e.g., gender classification or age regression) learned from one social media might perform rather poorly when tested on another social media due to the different data distributions in the two media. In this paper, we address cross-media user profiling by bridging the knowledge between the source and target media with a uniform user embedding learning approach. In our approach, we first construct a cross-media user-word network to capture the relationship among users through the textual information and a modified cross-media user-user network to capture the relationship among users through the social information. Then, we learn user embedding by jointly learning the heterogeneous network composed of above two networks. Finally, we train a classification (or regression) model with the obtained user embeddings as input to perform user profiling. Empirical studies demonstrate the effectiveness of the proposed approach to two cross-media user profiling tasks, i.e., cross-media gender classification and cross-media age regression.