All Stories

  1. SDG-MLLM: Injecting Structured Dialogue Graphs into MLLM for Multimodal Conversational Aspect-Based Sentiment Analysis
  2. DiffHarmony++: Enhancing Image Harmonization with Harmony-VAE and Inverse Harmonization Model
  3. Triple Alignment Strategies for Zero-shot Phrase Grounding under Weak Supervision
  4. EP-BERTGCN: A Simple but Effective Power Equipment Fault Recognition Method
  5. S2TD: A Tree-Structured Decoder for Image Paragraph Captioning
  6. Maintenance Decision Generator for Electrical Equipment Based on Reinforcement Learning
  7. Learning Visual Features from Product Title for Image Retrieval
  8. Correspondence Autoencoders for Cross-Modal Retrieval