About
Research & Development
Balancing Data Privacy and Utility with Curvature-Guided Perturbation
TLDR: This research introduces a novel bilevel optimization framework to address the privacy-utility trade-off in data publication. It uses an upper-level task to maximize...
Research & Development
Adapting Robot Intelligence: A New Framework for Vision-Language-Action Models
TLDR: Align-Then-stEer (ATE) is a novel framework that efficiently adapts pre-trained Vision-Language-Action (VLA) models to new robot embodiments and tasks. It achieves this by...
Research & Development
Optimizing LLM Prompts Through Comparative Learning
TLDR: A new framework called Contrastive Reasoning Prompt Optimization (CRPO) improves Large Language Model (LLM) prompt quality. It works by retrieving reference prompts from...
News & Current Events
Google Cloud Launches Second Gen AI Exchange Hackathon to Drive Real-World Generative AI Solutions in India
TLDR: Google Cloud, in partnership with Hack2skill, has announced the second edition of its Gen AI Exchange Hackathon, aiming to transition generative AI from...
Research & Development
Enhancing Language Models for Philippine Language Named Entity Recognition
TLDR: This research explores meta-pretraining with Model-Agnostic Meta-Learning (MAML) to improve zero-shot Named Entity Recognition (NER) in low-resource Philippine languages like Tagalog and Cebuano,...
Research & Development
Unlocking Protein Motion: A New Look at Simulating Molecular Dynamics with AI
TLDR: This research explores new AI-driven methods to simulate the complex, long-term movements of proteins. By breaking down the simulation into an encoder-propagator-decoder system,...