Stability AI has announced the next generation of its music generation model. Trained on properly licensed music, the model can generate up to 3 minutes of high-quality music. It also has audio-to-audio generation.
Researchers have developed an AI network where one AI can teach another to perform tasks using natural language processing, a capability not previously demonstrated. The system uses a model called S-Bert that allows AI to perform tasks given via instructions and then communicate that knowledge to another AI. This breakthrough has potential applications in robotics and could further understanding of human cognitive functions.
Opera has launched a new feature that allows users to download and run large language models locally on their computers, with over 150 models from more than 50 families available.
Researchers have developed DiJiang, a new approach that transforms existing Transformers into leaner, faster models without the heavy burden of retraining.
This study introduces a new method for creating driving paths for autonomous vehicles that combines diffusion models and transformers in a system called "World-Centric Diffusion Transformer" (WcDT).
Extracting information from datasets is critical for enterprise AI applications. These five new benchmark datasets can be used to measure general algorithmic performance for RAG applications.
This project introduces the concept of Unsolvable Problem Detection (UPD) in Vision Language Models, a new test to see if AI can identify when a problem can't be solved.
ASTRA is a Transformer-based model that is capable of identifying key moments during soccer matches and overcoming challenges like action localization and data imbalance.
Tools for Humanity has developed a secure and powerful computing environment for the Worldcoin Orb that utilizes NVIDIA Jetson for processing and Arm Cortex M4 microcontrollers for real-time functions. The Orb runs Rust applications and uses NVIDIA's TensorRT for neural network inference. It is powered by a custom security-focused GNU/Linux distribution called Orb OS. The system integrates a secure element for cryptography and supports trusted execution environments for backend authentication.
That Generative AI may turn out to be a disappointment. There are concerns about the technology's lack of profitability, security issues, and the inherent problem of hallucinations in language models. Unless a groundbreaking model like GPT-5 is released by the end of 2024, addressing key issues and offering a killer application, the hype surrounding Generative AI may start to dissipate.
Google is reportedly considering making the Search Generative Experience (SGE), which has been available through Search Labs for nearly a year, a paid feature as part of its Google One AI Premium subscription.