Efficient and accurate identification of drug-target affinity (DTA) is crucial in the virtual screening stage of drug discovery and the reuse of existing drugs. Deep neural networks are becoming popular tools for providing fast and accurate binding affinity prediction. MGraphDTA is one of the best performing models for DTA prediction, utilizing a super-deep graph neural network to extract features, demonstrating good generalization and interpretation capabilities. However, the complex deep neural networks come with a higher computational cost of model inference. Meanwhile, due to the expansion of drug databases, rapid completion of large-scale virtual screening still presents challenges. In this work, we propose an efficient binding affinity prediction method called DCUI-MGraphDTA through a series of optimizations on MGraphDTA to achieve large-scale virtual screening for millions of molecules on the Deep Computing Unit (DCU) cluster. The optimized DCUI-MGraphDTA model achieves a speedup of approximately 5.83 times on a single DCU compared to the original MGraphDTA. As the number of DCUs expands from 1 to 32, the total time for the model to process 1.28 million data significantly decreases, achieving a speedup of approximately 11.9 times.

Fourier Neural Operator (FNO) has been proven to be a universal and effective deep learning framework capable of achieving remarkable accuracy on Partial Differential Equation (PDE) solution problem. However, certain key components of emerging FNO-based models cannot leverage hardware potential, which makes it difficult to apply in high resolution and high realtime demand scenario. This paper presents a high optimized model called Speed Galerkin Transformer, including multilevel parallel SliceK-SplitK-ReduceK strategy for batched skinny matrix multiplication, memory layout optimization for QKV matrices and positional encodings and multi-head layer normalization fusion, as well as batched transposition optimization with strided scattering and gathering in 2D FNO, and these strategies can achieve $10.29 \mathrm{x}, 4.41 \mathrm{x}$ and 2.38 x speedup respectively under specific configuration. When solving the Darcy Flow equation at 512x512 resolution, the Speed Galerkin Transformer model can achieve about 1.72 x speedup, and achieve more than $\mathbf{9 0 \%}$ parallel efficiency on 8 GPUs.

Few-shot learning task aims to explore a model that is able to quickly learn new concepts by learning a few examples. The current approaches learning new categories with few images or even a single image are only based on the the visual modality. However, it is difficult to learn the representative features of new categories by a few images. This is because some categories are similar in vision. Moreover, due to the viewpoint, luminosity and that sometimes individuals of the same species appear markedly different from one another, the models are not able to learn the exact representation of classes. Therefore, considering that semantic information can enhance understanding when visual information is limited, we propose Knowledge-Augmented Networks (KAN), which combines the visual features with the semantic information extracted from knowledge graph to represent the features of each class. We demonstrate the effectiveness of our method on standard few-shot learning tasks, and further observe that with the augmented semantic information from knowledge graph, KAN is able to learn more disentangled representations. Experiments show that our model outperforms the state-of-the-art methods.