South America
Evaluating Vision-Language Models in the Wild with Human Preferences
Our comprehensive analysis of 20K real-world interactions reveals important insights into the failure cases of top-performing VLMs.
Cracking the Code of Juxtaposition: Can AI Models Understand the Humorous Contradictions
Recent advancements in large multimodal language models have demonstrated remarkable proficiency across a wide range of tasks. Y et, these models still struggle with understanding the nuances of human humor through juxtaposition, particularly when it involves nonlinear narratives that underpin many jokes and humor cues. This paper investigates this challenge by focusing on comics with contradictory narratives, where each comic consists of two panels that create a humorous contradiction.
Metric Transforms and Low Rank Representations of Kernels for Fast Attention
This suggests that the low-rank fast attention only works for functions approx-imable by polynomials.
RGFN: Synthesizable Molecular Generation Using GFlowNets
However, most existing machine learning methods for small molecule generation suffer from poor synthesizability of candidate compounds, making experimental validation difficult.
DHA: Learning Decoupled-Head Attention from Transformer Checkpoints via Adaptive Heads Fusion Yilong Chen
Large language models (LLMs) with billions of parameters demonstrate impressive performance.
MMSite: A Multi-modal Framework for the Identification of Active Sites in Proteins Song Ouyang
MMSite is a two-stage ("First Align, Then Fuse") Understanding these sites is essential for elucidating enzyme mechanisms, designing inhibitors, and developing novel drugs. MNER, high-quality datasets for multi-modal active sites identification are less abundant.
