This paper presents a layout- and style-based architecture for generative adversarial networks (LostGANs) that can be trained end-to-end to generate images from reconfigurable layout and style. The proposed LostGAN can learn fine-grained mask maps i…

We presented a cross-modality framework, LXMERT, for learning the connections between vision and language. We build the model based on Transfermer encoders and our novel cross-modality encoder. This model is then pre-trained with diverse pre-trainin…

In this paper, we proposed and evaluated a two-step framework which leverages features of two time scales for predicting DLMO. This framework is the extension of previously used models with input of either frequently sampled data (i.e., light exposu…

We address the motion synthesis of an interactive task—sitting onto a chair. We introduce a hierarchical reinforcement learning approach which relies on a collection of subtask controllers trained to imitate reusable mocap motions, and a meta contro…

This paper proposes a novel framework for blind image deblurring that uses deep generative networks as priors rather than in a conventional end-to-end manner. We report convincing deblurring results under the generative priors in comparison to the e…

In this paper, we have essentially replicated the study of McDonald and Nivre (2007, 2011) for neural parsers. In the baseline setting, where parsers use pre-trained word embeddings and character representations fed through a BiLSTM, we can still di…

This work was supported by a research grant (15334) from VILLUM FONDEN. This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement no 757360).

In this paper, we showed that synthesizing multi-modal LGE images from labelled bSSFP images to finetune a pre-trained segmentation network shows impressive segmentation performance on LGE images even though the network has not seen real labelled LG…

In this work, we propose a novel Semantic-Specific Graph Representation Learning framework, in which a semantic guided attentional mechanism is designed to learn semantic-related feature vectors and a graph propagation network is introduced to simul…

We present a novel framework to synthesize consecutive photorealistic video frames according to certain input texts, the synthesized virtual individual has appropriate mouth movements and facial expressions while speaking these input texts, which is…

Universal adversarial triggers expose new vulnerabilities for NLP—they are transferable across both examples and models. Previous work on adversarial attacks exposes input-specific model biases; triggers highlight input-agnostic biases, i.e., global…