About
Research & Development
Beyond Mimicry: Unpacking How Large Language Models Develop Understanding
TLDR: A research paper explores how Large Language Models (LLMs) move beyond simple pattern matching to develop genuine understanding, using mechanistic interpretability. It proposes...
Research & Development
Navigating In-Context Learning: A Deep Dive into How Examples Shape Multimodal AI for Image Captioning
TLDR: This research paper explores how different configurations of in-context examples affect Large Multimodal Models (LMMs) in image captioning. Through external experiments, it reveals...
Research & Development
Boosting LoRA Performance: A Constraint-Based Approach to Adapter Initialization
TLDR: ConsNoTrainLoRA (CNTLoRA) is a novel data-driven method for initializing Low-rank Adapters (LoRA) in foundation models. By using constraints on pre-training and fine-tuning activations,...
Research & Development
Developing a Secure and Efficient AI Framework for Respiratory Disease Diagnosis with Limited Data
TLDR: This research introduces a new AI framework called PFFL that helps diagnose respiratory diseases even when there's limited patient data and strict privacy...
Research & Development
Estimating Building Heights Globally with Single Satellite Radar Images and Deep Learning
TLDR: A new deep learning method accurately estimates building heights using single, very high-resolution SAR images from COSMO-SkyMed satellites. Tested across eight diverse cities...
Research & Development
New AI Framework Reconstructs Dynamic Human-Object Interactions from Single-Camera Video
TLDR: A new research paper introduces a framework for reconstructing dynamic human-object interactions from monocular video, addressing challenges like occlusions and temporal inconsistencies. The...