IBM has made significant strides in advancing its pioneering platforms for artificial intelligence and quantum computing. This week, Red Hat launched its RHEL AI platform, making it easier for businesses to develop, test, and deploy generative models. The platform includes IBM Research’s Granite models and the model improvement method InstructLab.
Additionally, private previews of Qiskit Serverless are now available, enabling users to work with both classical and quantum resources in the cloud. This powerful programming model is well-suited for complex, long-running tasks and regularly executed workflows. Furthermore, IBM has open-sourced its fast HAP filter on Hugging Face, a toxic language filter built for speed that can detect hateful, abusive, and profane text at each phase of the large language model lifecycle. These innovations demonstrate IBM’s commitment to pushing the boundaries of AI and quantum computing, with potential applications in various industries.
Advancing AI and Quantum Computing Platforms
IBM Research has made significant strides in advancing its pioneering platforms for artificial intelligence (AI) and quantum computing. In this article, we will delve into the latest innovations and updates from IBM Research, including the launch of RHEL AI, Qiskit Serverless, and a fast HAP filter on Hugging Face.
Launch of RHEL AI
Red Hat has opened up its RHEL AI platform to all developers, making it easier for businesses to develop, test, and deploy generative models. The new platform includes many of the Granite models incubated at IBM Research, along with the model improvement method InstructLab. This launch is a significant step forward in democratizing access to AI tools and enabling more widespread adoption.
RHEL AI provides a comprehensive platform for developers to work with AI models, including development, testing, and deployment. The inclusion of Granite models and InstructLab will enable businesses to improve their AI capabilities and develop more accurate models. With the launch of RHEL AI, IBM Research is further solidifying its position as a leader in the field of AI research and development.
Qiskit Serverless and Qiskit Functions
Private previews of Qiskit Serverless are now available, enabling users to work with both classical and quantum resources in the cloud. This powerful programming model is well-suited for complex, long-running tasks and regularly executed workflows. IBM Research scientists have been improving Qiskit Serverless lately, making it an even more robust tool for developers.
Qiskit Serverless sets the stage for Qiskit Functions, which will enable users to work with quantum resources in a serverless environment. This will provide greater flexibility and scalability for developers working on complex projects. The development of Qiskit Serverless and Qiskit Functions is a significant step forward in advancing IBM’s quantum computing capabilities.
Fast HAP Filter on Hugging Face
IBM has open-sourced its fast HAP filter, granite-guardian-hap-38m, on Hugging Face. This AI filter is designed to detect hateful, abusive, and profane (HAP) text at incredible speed, making it an essential tool for developers working with large language models.
The granite-guardian-hap-38m filter is small enough to run on a CPU and quick enough to catch HAP-related text at each phase of the LLM lifecycle. This will enable developers to integrate HAP detection into their workflows seamlessly, ensuring that their AI models are trained on clean and respectful data. The open-sourcing of this filter is a significant contribution to the development of more responsible AI systems.
IBM has introduced the latest feature update to watsonx.ai, an open framework where users can access a catalogue of built-in models and patterns. With this update, users can also import custom foundation models, enabling them to bring in the best tool for the task, even if it resides outside of watsonx.ai.
This update is significant because it provides greater flexibility and customization options for developers working with AI models. By allowing users to import custom foundation models, IBM is further democratizing access to AI tools and enabling more widespread adoption.
IBM researchers have published several new papers that highlight the company’s continued innovation in AI and quantum computing. These papers include “Automated cut finding and circuit knitting on large quantum circuits,” “Scale-Adaptive Balancing of Exploration and Exploitation in Classical Planning,” and “Building Quantum Workflows using Composable Software within the Qiskit Patterns Framework.”
These publications demonstrate IBM Research’s continued commitment to advancing the state-of-the-art in AI and quantum computing. By pushing the boundaries of what is possible with these technologies, IBM researchers are enabling new applications and use cases that will transform industries and societies.
IBM Research has made significant strides in advancing its pioneering platforms for artificial intelligence (AI) and quantum computing. The launch of RHEL AI, Qiskit Serverless, and a fast HAP filter on Hugging Face demonstrate the company’s continued commitment to innovation and democratizing access to AI tools. By providing greater flexibility and customization options for developers, IBM is further solidifying its position as a leader in the field of AI research and development.
External Link: Click Here For More
