IBM THINK: New open source Granite models, watsonx capabilities, and IBM Concert


IBM hosted its annual conference, THINK, this week, and at the event it unveiled its current strategy and vision around AI, along with several innovations to back it up. 

One of the points CEO Arvind Krishna touched on in his opening keynote is the desire to invest in and contribute to the open-source AI community. 

“We firmly believe in bringing open innovation to AI. We want to use the power of open source to do with AI what was successfully done with Linux and OpenShift,” said Krishna. “Open means choice. Open means more eyes on the code, more minds on the problems, and more hands on the solutions. For any technology to gain velocity and become ubiquitous, you’ve got to balance three things: competition, innovation, and safety. Open source is a great way to achieve all three.”

To that end, IBM has open sourced a new set of its Granite AI models, which can be used for tasks like application modernization, code generation, troubleshooting, documentation, and maintaining repositories, among other things. 

The models are now available on Hugging Face and GitHub, ranging from 3B to 34B parameter varieties. 

Additionally, the company, in partnership with Red Hat, launched InstructLab, which is a model alignment technique that enables developers to create models that are specific to their own companies or domains. An enterprise-ready version of InstructLab is available as part of RHEL AI. 

“It can enhance an LLM using far less human-generated information and far fewer computing resources than are typically used to retrain a model. And it makes it possible for upstream contributions to continuously make the model better,” IBM wrote in a blog post

The company also shared several upcoming improvements to watsonx. It will be releasing the watsonx Code Assistant for Enterprise Java Applications in October, watsonx Assistant for Z in June, and will expand watsonx Code Assistant for Z Service that same month to offer code explanation capabilities.

IBM will also add an AI assistant builder in watsonx Orchestrate, which will provide a low-code interface for creating new AI assistants. “The AI assistant builder enables you to extend your investments in automation and scale adoption through a highly conversational interface that is infused with AI that you can trust,” IBM wrote in a blog post

Next month, the company will also release IBM Concert, which uses generative AI to provide insights from a company’s applications, cloud infrastructure, source repositories, and CI/CD pipelines. It will help customers identify, predict, and respond to issues before they happen, according to IBM. 

“Powered by AI from watsonx, Concert can generate analyses, visualizations and recommendations that you can quickly turn into action. By helping you discover gaps, prioritize insights and instrument changes, Concert helps to reduce complexity and streamline operations so you can make your business more resilient, more innovative, and more cost-effective,” IBM wrote in a blog post

And finally, IBM announced it is continuing to improve collaborations with a number of companies to expand model choice. These include AWS, Adobe, Meta, Microsoft, Mistral, Palo Alto Networks, Salesforce, SAP, and SDAIA. 

“Our ecosystem of partners, large and small, are helping clients adopt and scale tailored AI across their businesses,” IBM said. 



Source link