News from the AI & ML world
DeeperML - #researchers
|
@colab.research.google.com
//
Google's Magenta project has unveiled Magenta RealTime (Magenta RT), an open-weights live music model designed for interactive music creation, control, and performance. This innovative model builds upon Google DeepMind's research in real-time generative music, providing opportunities for unprecedented live music exploration. Magenta RT is a significant advancement in AI-driven music technology, offering capabilities for both skill-gap accessibility and enhancement of existing musical practices. As an open-weights model, Magenta RT is targeted towards eventually running locally on consumer hardware, showcasing Google's commitment to democratizing AI music creation tools.
Magenta RT, an 800 million parameter autoregressive transformer model, was trained on approximately 190,000 hours of instrumental stock music. It leverages SpectroStream for high-fidelity audio (48kHz stereo) and a newly developed MusicCoCa embedding model, inspired by MuLan and CoCa. This combination allows users to dynamically shape and morph music styles in real-time by manipulating style embeddings, effectively blending various musical styles, instruments, and attributes. The model code is available on Github and the weights are available on Google Cloud Storage and Hugging Face under permissive licenses with some additional bespoke terms. Magenta RT operates by generating music in sequential chunks, conditioned on both previous audio output and style embeddings. This approach enables the creation of interactive soundscapes for performances and virtual spaces. Impressively, the model achieves a real-time factor of 1.6 on a Colab free-tier TPU (v2-8 TPU), generating two seconds of audio in just 1.25 seconds. This technology unlocks the potential to explore entirely new musical landscapes, experiment with never-before-heard instrument combinations, and craft unique sonic textures, ultimately fostering innovative forms of musical expression and performance. Recommended read:
References :
nftjedi@chatgptiseatingtheworld.com
//
Apple researchers recently published a study titled "The Illusion of Thinking," suggesting that advanced language models (LLMs) struggle with true reasoning, relying instead on pattern matching. The study presented findings based on tasks like the Tower of Hanoi puzzle, where models purportedly failed when complexity increased, leading to the conclusion that these models possess limited problem-solving abilities. However, these conclusions are now under scrutiny, with critics arguing the experiments were not fairly designed.
Alex Lawsen of Open Philanthropy has published a counter-study challenging the foundations of Apple's claims. Lawsen argues that models like Claude, Gemini, and OpenAI's latest systems weren't failing due to cognitive limits, but rather because the evaluation methods didn't account for key technical constraints. One issue raised was that models were often cut off from providing full answers because they neared their maximum token limit, a built-in cap on output text, which Apple's evaluation counted as a reasoning failure rather than a practical limitation. Another point of contention involved the River Crossing test, where models faced unsolvable problem setups. When the models correctly identified the tasks as impossible and refused to attempt them, they were still marked wrong. Furthermore, the evaluation system strictly judged outputs against exhaustive solutions, failing to credit models for partial but correct answers, pattern recognition, or strategic shortcuts. To illustrate, Lawsen demonstrated that when models were instructed to write a program to solve the Hanoi puzzle, they delivered accurate, scalable solutions even with 15 disks, contradicting Apple's assertion of limitations. Recommended read:
References :
@machinelearning.apple.com
//
Apple researchers have released a new study questioning the capabilities of Large Reasoning Models (LRMs), casting doubt on the industry's pursuit of Artificial General Intelligence (AGI). The research paper, titled "The Illusion of Thinking," reveals that these models, including those from OpenAI, Google DeepMind, Anthropic, and DeepSeek, experience a 'complete accuracy collapse' when faced with complex problems. Unlike existing evaluations primarily focused on mathematical and coding benchmarks, this study evaluates the reasoning traces of these models, offering insights into how LRMs "think".
Researchers tested various models, including OpenAI's o3-mini, DeepSeek-R1, and Claude 3.7 Sonnet, using puzzles like the Tower of Hanoi, Checker Jumping, River Crossing, and Blocks World. These environments allowed for the manipulation of complexity while maintaining consistent logical structures. The team discovered that standard language models surprisingly outperformed LRMs in low-complexity scenarios, while LRMs only demonstrated advantages in medium-complexity tasks. However, all models experienced a performance collapse when faced with highly complex tasks. The study suggests that the so-called reasoning of LRMs may be more akin to sophisticated pattern matching, which is fragile and prone to failure when challenged with significant complexity. Apple's research team identified three distinct performance regimes: low-complexity tasks where standard models outperform LRMs, medium-complexity tasks where LRMs show advantages, and high-complexity tasks where all models collapse. Apple has begun integrating powerful generative AI into its own apps and experiences. The new Foundation Models framework gives app developers access to the on-device foundation language model. Recommended read:
References :
Chris McKay@Maginative
//
Google's AI research notebook, NotebookLM, has introduced a significant upgrade that enhances collaboration by allowing users to publicly share their AI-powered notebooks with a simple link. This new feature, called Public Notebooks, enables users to share their research summaries and audio overviews generated by AI with anyone, without requiring sign-in or permissions. This move aims to transform NotebookLM from a personal research tool into an interactive, AI-powered knowledge hub, facilitating easier distribution of study guides, project briefs, and more.
The public sharing feature provides viewers with the ability to interact with AI-generated content like FAQs and overviews, as well as ask questions in chat. However, they cannot edit the original sources, ensuring the preservation of ownership while enabling discovery. To share a notebook, users can click the "Share" button, switch the setting to "Anyone with the link," and copy the link. This streamlined process is similar to sharing Google Docs, making it intuitive and accessible for users. This upgrade is particularly beneficial for educators, startups, and nonprofits. Teachers can share curated curriculum summaries, startups can distribute product manuals, and nonprofits can publish donor briefing documents without the need to build a dedicated website. By enabling easier sharing of AI-generated notes and audio overviews, Google is demonstrating how generative AI can be integrated into everyday productivity workflows, making NotebookLM a more grounded tool for sense-making of complex material. Recommended read:
References :
@techhq.com
//
References:
TechHQ
, futurumgroup.com
Dell Technologies and NVIDIA are collaborating to construct the "Doudna" supercomputer for the U.S. Department of Energy (DOE). Named after Nobel laureate Jennifer Doudna, the system will be housed at the Lawrence Berkeley National Laboratory's National Energy Research Scientific Computing Center (NERSC) and is slated for deployment in 2026. This supercomputer aims to revolutionize scientific research by merging artificial intelligence (AI) with simulation capabilities, empowering over 11,000 researchers across various disciplines, including fusion energy, astronomy, and life sciences. The project represents a significant federal investment in high-performance computing (HPC) infrastructure, designed to maintain U.S. leadership in AI and scientific discovery.
The Doudna supercomputer, also known as NERSC-10, promises a tenfold increase in scientific output compared to its predecessor, Perlmutter, while only consuming two to three times the power. This translates to a three-to-five-fold improvement in performance per watt, achieved through advanced chip design and system-level efficiencies. The system integrates high-performance CPUs with coherent GPUs, enabling direct data access and sharing across processors, addressing traditional bottlenecks in scientific computing workflows. Doudna will also be connected to DOE experimental and observational facilities through the Energy Sciences Network (ESnet), facilitating seamless data streaming and near real-time analysis. According to NVIDIA CEO Jensen Huang, Doudna is designed to accelerate scientific workflows and act as a "time machine for science," compressing years of discovery into days. Energy Secretary Chris Wright sees it as essential infrastructure for maintaining American technological leadership in AI and quantum computing. The supercomputer emphasizes coherent memory access between CPUs and GPUs, enabling data sharing in heterogeneous processors, which is a requirement for modern AI-accelerated scientific workflows. The Nvidia Vera Rubin supercomputer architecture introduces hardware-level optimizations designed specifically for the convergence of simulation, machine learning, and quantum algorithm development. Recommended read:
References :
|
BenchmarksBlogsResearch Tools |