You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Human-centered multimodal sexism detection for EXIST 2026 (CLEF) Task 2: a five-stream FiLM-conditioned cross-attention model that fuses meme text, image, annotator demographics, and biometric signals. Team VANGUARD.
MemeSense leverages CLIP for multimodal meme analysis, integrating vision and language features through cosine similarity to classify humor, hate speech,stance.
Meme Sommelier: a multimodal ML + VLM workbench for Memotion7K meme humor analysis. Combines text, image, CLIP fusion, Gemini-generated H-layer mechanisms, unsupervised clustering, and a Streamlit agent to explain why memes land or fail.