Identifying novel drug-protein interactions is crucial for drug discovery. For this purpose, many machine learning-based methods have been developed based on drug descriptors and one-dimensional (1D) protein sequences. However, protein sequence can't accurately reflect the interactions in 3D space. On the other hand, a direct input of 3D structure is of low efficiency due to the sparse 3D matrix, and is also prevented by limited number of co-crystal structures available for training. In this work, we propose an end-to-end deep learning framework to predict the interactions by representing proteins with 2D distance map from monomer structures (Image), and drugs with molecular linear notation (String), following the Visual Question Answering mode. For an efficient training of the system, we introduced a dynamic attentive convolutional neural network to learn fixed-size representations from the variable-length distance maps and a self-attentional sequential model to automatically extract semantic features from the linear notations. Extensive experiments demonstrate that our model obtains competitive performance against state-ofthe-art baselines on the DUD-E, Human and Bind-ingDB benchmark datasets. Further attention visualization provides biological interpretation to depict highlighted regions of both protein and drug molecules.

Protein sequence profile prediction aims to generate multiple sequences from structural information to advance the protein design. Protein sequence profile can be computationally predicted by energy-based method or fragment-based methods. By integrating these methods with neural networks, our previous method, SPIN2 has achieved a sequence recovery rate of 34%. However, SPIN2 employed only one dimensional (1D) structural properties that are not sufficient to represent 3D structures. In this study, we represented 3D structures by 2D maps of pairwise residue distances. and developed a new method (SPROF) to predict protein sequence profile based on an image captioning learning frame. To our best knowledge, this is the first method to employ 2D distance map for predicting protein properties. SPROF achieved 39.8% in sequence recovery of residues on the independent test set, representing a 5.2% improvement over SPIN2. We also found the sequence recovery increased with the number of their neighbored residues in 3D structural space, indicating that our method can effectively learn long range information from the 2D distance map. Thus, such network architecture using 2D distance map is expected to be useful for other 3D structure-based applications, such as binding site prediction, protein function prediction, and protein interaction prediction.

Paraphrase generation is of great importance to many downstream tasks in natural language processing. Recent efforts have focused on generating paraphrases in specific syntactic forms, which, generally, heavily relies on manually annotated paraphrase data that is not easily available for many languages and domains. In this paper, we propose a novel end-to-end framework to leverage existing large-scale bilingual parallel corpora to generate paraphrases under the control of syntactic exemplars. In order to train one model over the two languages of parallel corpora, we embed sentences of them into the same content and style spaces with shared content and style encoders using cross-lingual word embeddings. We propose an adversarial discriminator to disentangle the content and style space, and employ a latent variable to model the syntactic style of a given exemplar in order to guide the two decoders for generation. Additionally, we introduce cycle and masking learning schemes to efficiently train the model. Experiments and analyses demonstrate that the proposed model trained only on bilingual parallel data is capable of generating diverse paraphrases with desirable syntactic styles. Fine-tuning the trained model on a small paraphrase corpus makes it substantially outperform state-of-the-art paraphrase generation models trained on a larger paraphrase dataset.

In this paper, we propose a learning-based low-overhead beam alignment method for vehicle-to-infrastructure communication in vehicular networks. The main idea is to remotely infer the optimal beam directions at a target base station in future time slots, based on the CSI of a source base station in previous time slots. The proposed scheme can reduce channel acquisition and beam training overhead by replacing pilot-aided beam training with online inference from a sequence-to-sequence neural network. Simulation results based on ray-tracing channel data show that our proposed scheme achieves a 8.86% improvement over location-based beamforming schemes with a positioning error of 1m, and is within a 4.93% performance loss compared with the genieaided optimal beamformer.