InfoQ Homepage AI, ML & Data Engineering Content on InfoQ
-
Google Develops Voice Transfer AI for Restoring Voices
A team at Google Research developed a zero-shot voice transfer (VT) model that can be used to customize a text-to-speech (TTS) with a specific person's voice. This allows speakers who have lost their voice, for example from Parkinson's disease or ALS, to use a TTS device to replicate their original voice. The model also works across languages.
-
PyTorch Conference 2024: PyTorch 2.4/Upcoming 2.5, and Llama 3.1
The PyTorch Conference 2024, held by The Linux Foundation, showcased groundbreaking advancements in AI, featuring insights on PyTorch 2.4, Llama 3.1, and open-source projects like OLMo. Key discussions on LLM deployment, ethical AI, and innovative libraries like Torchtune and TorchChat emphasized collaboration and responsible practices in the evolving landscape of generative AI.
-
Anthropic Unveils Contextual Retrieval for Enhanced AI Data Handling
Anthropic has announced Contextual Retrieval, a significant advancement in AI systems' interaction with extensive knowledge bases. This technique addresses the challenge of context loss in Retrieval-Augmented Generation (RAG) systems by enriching text chunks with contextual information before embedding or indexing.
-
Microsoft Launches Azure AI Inference SDK for .NET
Microsoft launched Azure AI Inference SDK for .NET, streamlining access to generative AI models in the Azure AI Studio model catalog. This catalog includes models from providers like Azure OpenAI Service, Mistral, Meta, Cohere, NVIDIA, and Hugging Face, organized into three collections: Curated by Azure AI, Azure OpenAI Models, and Open Models from Hugging Face Hub.
-
Uber Creates GenAI Gateway Mirroring OpenAI API to Support over 60 LLM Use Cases
Uber created a unified platform for serving large language models (LLMs) from external vendors and self-hosted ones and opted to mirror OpenAI API to help with internal adoption. GenAI Gateway provides a consistent and efficient interface and serves over 60 distinct LLM use cases across many areas.
-
Study Shows AI Coding Assistant Improves Developer Productivity
Researchers from Microsoft, MIT, Princeton University, and the Wharton School of the University of Pennsylvania recently published a study that showed the use of GitHub Copilot increased developer productivity. The team conducted three separate randomized controlled trials (RCT) involving over 4,000 developers; the ones using Copilot achieved a 26% increase in productivity.
-
Google Proposes Adding Pipe Syntax to SQL
In a recent paper titled "SQL Has Problems. We Can Fix Them", a research team at Google proposed the introduction of a new Pipe Syntax in SQL. Google's solution to address perceived limitations in SQL is currently available in the GoogleSQL and ZetaSQL dialects, but it has received mixed feedback from the community.
-
Stability AI Announces Integration of Top Text-to-Image Models with Amazon Bedrock
Stability AI has introduced three new text-to-image models to Amazon Bedrock: Stable Image Ultra, Stable Diffusion 3 Large, and Stable Image Core. These models focus on improving performance in multi-subject prompts, image quality, and typography. They are designed to generate high-quality visuals for various use cases in marketing, advertising, media, entertainment, retail, and more.
-
AWS Announces General Availability of EC2 P5e Instances, Powered by NVIDIA H100 Tensor Core GPUs
Amazon Web Services (AWS) has launched EC2 P5e instances featuring NVIDIA H100 Tensor Core GPUs, substantially boosting AI and HPC performance. With enhanced memory bandwidth, these instances reduce latency for real-time applications. Ideal for tasks like LLM training and simulations, they offer improved scalability and cost-efficiency, making them pivotal for modern cloud computing.
-
Apple Open-Sources Multimodal AI Model 4M-21
Researchers at Apple and the Swiss Federal Institute of Technology Lausanne (EPFL) have open-sourced 4M-21, a single any-to-any AI model that can handle 21 input and output modalities. 4M-21 performs well "out of the box" on several vision benchmarks and is available under the Apache 2.0 license.
-
Security Experts Exploit Airport Security Loophole with SQL Injection
In the article "Bypassing airport security via SQL injection," two security researchers recently demonstrated how they executed a simple SQL injection attack on a service that enables pilots and flight attendants to bypass airport security screening.
-
Google Announces Game Simulation AI GameNGen
A research team from Google recently published a paper on GameNGen, a generative AI model that can simulate the video game Doom. GameNGen can simulate the game at 20 frames-per-second (FPS) and in human evaluations was preferred only slightly less often than the actual game.
-
HelixML Announces Helix 1.0 Release
HelixML has announced their Helix platform for Generative AI is production ready at version 1.0. Described as a "Private GenAI Stack," the platform provides an interface layer and applications that can be connected to a variety of LLMs. It can be used to prototype apps, starting with a laptop, with all components version controlled to ease subsequent deployment and scaling.
-
Microsoft .NET Conf: Focus on AI
Last month the latest edition of the .NET Conf: Focus series for 2024 took place, featuring AI development topics. The event targeted developers of all levels, with both informative and hands-on sessions showcasing how to use artificial intelligence within the .NET ecosystem. This was one of the most technical events ever in the Focus series, with many high-quality, deep-dive sessions.
-
Leveraging the Transformer Architecture for Music Recommendation on YouTube
Google has described an approach to use transformer models, which ignited the current generative AI boom, for music recommendation. This approach, which is currently being applied experimentally on YouTube, aims to build a recommender that can understand sequences of user actions when listening to music to better predict user preferences based on their context.