Search Results for author:
Found 7 papers, 4 papers with code
Enhancing Visual Continual Learning with Language-Guided Supervision
Existing methods commonly utilize the one-hot labels and randomly initialize the classifier head.
INTERS: Unlocking the Power of Large Language Models in Search with Instruction Tuning
Despite this, their application to information retrieval (IR) tasks is still challenging due to the infrequent occurrence of many IR-specific concepts in natural language.
Knowledge-injected Prompt Learning for Chinese Biomedical Entity Normalization
However, existing research falls short in tackling the more complex Chinese BEN task, especially in the few-shot scenario with limited medical data, and the vast potential of the external medical knowledge base has yet to be fully harnessed.
SADM: Sequence-Aware Diffusion Model for Longitudinal Medical Image Generation
To this end, we propose a sequence-aware diffusion model (SADM) for the generation of longitudinal medical images.
Continual Stereo Matching of Continuous Driving Scenes With Growing Architecture
The deep stereo models have achieved state-of-the-art performance on driving scenes, but they suffer from severe performance degradation when tested on unseen scenes.
Knowledge Mining and Transferring for Domain Adaptive Object Detection
Specifically, KTNet is constructed on a base detector with intrinsic knowledge mining and relational knowledge constraints.
Brief Announcement: On the Limits of Parallelizing Convolutional Neural Networks on GPUs
However, popular deep learning (DL) frameworks such as TensorFlow and PyTorch launch the majority of neural network operations, especially convolutions, serially on GPUs and do not exploit this inter-op parallelism.
