A.I. Less is More: A Unified Architecture for Device-Directed Speech Detection with Multiple Invocation Types 06/09/2023
A.I. Meet Pix2Act: an AI agent that can interact with GUIs using the same conceptual interface humans commonly use through pixel-based screenshots and generic keyboard and mouse actions 06/08/2023
A.I. Scaling Generative Retrieval: Google Research and University of Waterloo empirical study of generative retrieval at various corpus scales, including a deep dive into the MS MARCO task of 8.8 million passages 06/08/2023
A.I. Michelangelo’s AI cousin: Neuralangelo is an AI model that can achieve high-fidelity 3D surface reconstruction 06/08/2023
A.I. AI agents can learn to think while acting: New AI research introduces a new imitation learning framework called thought cloning 06/08/2023
A.I. Exploring AVFormer: Google AI’s innovative approach to augmenting audio-only models with visual feedback and optimized domain adaptation 06/08/2023
A.I. Meet STEVE-1: An instructive generative AI model for Minecraft that follows visual and text instructions and only costs $60 to train 06/07/2023
A.I. This AI paper proposes a self-monitored music comprehension model called MERT that achieves overall SOTA performance on 14 MIR tasks. 06/07/2023
Speculative Sampling: Explained Intuitively and Exhaustively | by Daniel Warfield | December 2023 12/15/2023
