The challenge of tailoring general-purpose LLMs to specific tasks without extensive retraining or additional data persists even after significant advancements in the field. Adapting LMs for specialized tasks often requires substantial computational resources and...

The quest for clean, usable data for pretraining Large Language Models (LLMs) resembles searching for treasure amidst chaos. While rich with information, the digital realm is cluttered with extraneous content that complicates the extraction of valuable data. This...

The CHiME-8 MMCSG task focuses on the challenge of transcribing conversations recorded using smart glasses equipped with multiple sensors, including microphones, cameras, and inertial measurement units (IMUs). The dataset aims to help researchers to solve problems...

Point clouds serve as a prevalent representation of 3D data, with the extraction of point-wise features being crucial for various tasks related to 3D understanding. While deep learning methods have made significant strides in this domain, they often rely on large and...

Creating a model that excels at understanding, holding conversations, and solving complex problems has always been challenging in artificial intelligence. The goal is to develop a system that can chat like a human and think and reason through difficult questions. This...

The exploration of refining large language models (LLMs) to enhance their instruction-following prowess has surged, with Reinforcement Learning with AI Feedback (RLAIF) being a promising technique. This method traditionally involves an initial phase of Supervised...