Apr 27, 2025

Unauthorized AI Experiment

Researchers from the University of Zurich conducted an unauthorized study on CMV users, deploying AI-generated comments to assess their persuasiveness without prior disclosure or consent, violating community rules.

LLMs can see and hear without any training

MILS enables LLMs to process visual and auditory data without prior training, showcasing a significant advancement in cross-modal learning capabilities.

How NASA Is Using Graph Technology and LLMs to Build a People Knowledge Graph

NASA's People Knowledge Graph leverages graph databases and large language models (LLMs) to enhance people analytics, enabling real-time insights into expertise, project similarities, and skill gaps across the agency.

CosAE: Learnable Fourier Series for Image Restoration

CosAE (Cosine Autoencoder) innovatively combines Fourier series with a feed-forward neural network, enabling extreme spatial compression while preserving image detail during restoration.

Intuition behind Load-Balancing Loss in the paper OUTRAGEOUSLY LARGE NEURAL NETWORKS: THE SPARSELY-GATED MIXTURE-OF-EXPERTS LAYER

The Load-Balancing Loss in the paper "OUTRAGEOUSLY LARGE NEURAL NETWORKS" aims to ensure that all experts in a sparsely-gated mixture-of-experts model are utilized effectively, preventing any single expert from becoming overloaded while others remain underused.

Research Paper and Presentation about Multi-Agent Reinforcement Learning

Multi-Agent Reinforcement Learning (MARL) focuses on competitive Game AI, requiring a nuanced understanding of game models like Stochastic Games and Partial Observable Stochastic Games (POSG) to effectively capture its complexities.