Invited Speakers

Program

Schedule:  June 3, 8:30 AM – 12:00 PM

• 08:30 - 08:40 📢Workshop Kickoff and Opening Remarks 

• 08:40 - 09:40 🎉Six Oral Presentations (10 minutes each, 7 minutes for the talk and 3 minutes for Q&A) 

  • 08:40 - 08:50 Gen-SIS: Generative Self-augmentation Improves Self-supervised Learning

  • 08:50 - 09:00 BlendFusion - Scalable Synthetic Data Generation for Diffusion Model Training

  • 09:00 - 09:10  Analyzing Training-Free Corruption Detection for Object Detection Datasets

  • 09:10 - 09:20  OASIS: Generating Synthetic Skin Artifacts

  • 09:20 - 09:30 VGGHeads: 3D Multi Head Alignment with a Large-Scale Synthetic Dataset

  • 09:30 - 09:40  Less Detail, Better Answers: Degradation-Driven Prompting for VQA

• 09:40 - 10:20 🌟First Keynote Speech: Lily Weng  "Trustworthy Evaluation of Model Interpretability: A Unified Framework with Sanity Checks"

• 10:20 - 10:50 🌟Second Keynote Speech: Jason Alan Fries 

• 10:50 - 11:20 🌟Third Keynote Speech: Ludwig Schmidt "OpenThoughts"

• 11:20 - 11:40 🎉Two Oral Presentations (10 minutes each, 7 minutes for the talk and 3 minutes for Q&A) 

  • 11:20 - 11:30  Seeing the Evidence, Missing the Answer: Tool-Guided Vision-Language Models on Visual Illusions 

  • 11:30 - 11:40  Edit, Crop, and Prompt: Training-Free Visual Question Answering via Preprocessing Pipelines for Real-World Visual Illusions

• 11:40 - 11:45 📢 Challenge Introduction and Winner Announcement

• 11:45 – 12:00  Concluding Remark 

• 12:00 – 12:30  Poster Session

Contact Us

For additional information please contact us.