Microsoft partners with industry leaders to offer vertical SLMs

At its Microsoft Ignite 2024 show in Chicago this week, Microsoft and industry partner experts showed off the power of small language models (SLMs) with a new set of fine-tuned, pre-trained AI models using industry-specific data.“While organizations around the world recognize the value and potential of AI, for AI to be truly effective, it must be tailored to specific industry needs,” said Satish Thomas, corporate VP of business and industry solutions at Microsoft, in a blog post.SLMs are AI models that are smaller in scale and scope than large language models (LLMs), and are intended for specialized use cases. They’re also typically trained on more specific and higher quality datasets than LLMs, and because they’re more compact and efficient than LLMs, when used in appropriate use cases, they require less computational power and memory. Microsoft is working closely with partners — including Bayer, Cerence, Rockwell Automation, Saifr, Siemens Digital Industries Software, and Sight Machine — to leverage its Phi family of SLMs so it can provide these adapted AI models through the Azure AI model catalog. The company notes that customers can also use the models to configure agents in Microsoft Copilot Studio. Among the first models released is E.L.Y. Crop Protection, developed in collaboration with Bayer, which has also been blending gen AI and data science to help spur new agricultural solutions. The E.L.Y. Crop Protection SLM aims to help farmers’ decision-making processes around the application of crop treatments and pesticides. It leverages Bayer’s agricultural intelligence, trained on thousands of real-world questions on Bayer crop protection labels, along with data on regulatory requirements and environmental conditions. Microsoft said it’s scalable to farm operations of all types and sizes, and is customizable so that organizations can adapt the model to regional and crop-specific requirements.Microsoft will also be offering CaLLM Edge, an automotive-specific, embedded SLM developed by Cerence. CaLLM Edge enables drivers to control functions such as navigation and climate control, and is designed for use in conditions where vehicles might have limited or no cloud connectivity, like remote areas.Rockwell Automation is adding FT Optix Food & Beverage to the Azure AI catalog as well. This SLM helps frontline workers in manufacturing troubleshoot food and beverage assets. It provides factory floor workers and engineers with recommendations, explanations, and knowledge about specific manufacturing processes, machines, and inputs.Two addition manufacturing-aligned SLMs have also been announced, including a co-pilot for Siemens Digital Industries Software’s NX X industrial design software. The SLM enables users to ask natural language questions, access detailed technical insights, and streamline complex design tasks in an effort to streamline complex tasks prone to human error.Meanwhile, manufacturing data analytics company Sight Machine has developed Factory Namespace Manager, a model intended to help manufacturers rename and integrate factory data in their corporate data systems to facilitate analyzing and optimizing production alongside supply chain, sales, finance, and other corporate functions. Manufacturers often struggle with integrating and analyzing data from multiple sites due to inconsistent naming conventions for machines and processes. This SLM automatically translates the various names into a standard format.For financial institutions looking to better manage regulatory compliance, Saifr, a regulatory tech provider within Fidelity Investment’s innovation incubator, Fidelity Labs, has developed a set of four models: Retail Marketing Compliance, to ensure marketing materials adhere to industry regulations and standards; Risk Interpretation, to identify potential risks in marketing content; Language Suggestion, to enhance the compliance of marketing messages; and Image Detection, to analyze and verifying the appropriateness of images used in marketing campaigns.

At its Microsoft Ignite 2024 show in Chicago this week, Microsoft and industry partner experts showed off the power of small language models (SLMs) with a new set of fine-tuned, pre-trained AI models using industry-specific data. “While organizations around the world recognize the value and potential of AI, for AI to be truly effective, it must be tailored to specific industry needs,” said Satish Thomas, corporate VP of business and industry solutions at Microsoft, in a blog post. SLMs are AI models that are smaller in scale and scope than , and are .

They’re also typically trained on more specific and higher quality datasets than LLMs, and because they’re more compact and efficient than LLMs, when used in appropriate use cases, they require . Microsoft is working closely with partners — including Bayer, Cerence, Rockwell Automation, Saifr, Siemens Digital Industries Software, and Sight Machine — to leverage its Phi family of SLMs so it can provide these adapted AI models through the . The company notes that customers can also use the models to configure agents in Microsoft Copilot Studio.

Among the first models released is E.L.Y.

Crop Protection, developed in collaboration with Bayer, which has also been . The E.L.

Y. Crop Protection SLM aims to help farmers’ decision-making processes around the application of crop treatments and pesticides. It leverages Bayer’s agricultural intelligence, trained on thousands of real-world questions on Bayer crop protection labels, along with data on regulatory requirements and environmental conditions.

Microsoft said it’s scalable to farm operations of all types and sizes, and is customizable so that organizations can adapt the model to regional and crop-specific requirements. Microsoft will also be offering CaLLM Edge, an automotive-specific, embedded SLM developed by Cerence. CaLLM Edge enables drivers to control functions such as navigation and climate control, and is designed for use in conditions where vehicles might have limited or no cloud connectivity, like remote areas.

Rockwell Automation is adding FT Optix Food & Beverage to the Azure AI catalog as well. This SLM helps frontline workers in manufacturing troubleshoot food and beverage assets. It provides factory floor workers and engineers with recommendations, explanations, and knowledge about specific manufacturing processes, machines, and inputs.

Two addition manufacturing-aligned SLMs have also been announced, including a co-pilot for Siemens Digital Industries Software’s NX X industrial design software. The SLM enables users to ask natural language questions, access detailed technical insights, and streamline complex design tasks in an effort to streamline complex tasks prone to human error. Meanwhile, manufacturing data analytics company Sight Machine has developed Factory Namespace Manager, a model intended to help manufacturers rename and integrate factory data in their corporate data systems to facilitate analyzing and optimizing production alongside supply chain, sales, finance, and other corporate functions.

Manufacturers often struggle with integrating and analyzing data from multiple sites due to inconsistent naming conventions for machines and processes. This SLM automatically translates the various names into a standard format. For financial institutions looking to better manage regulatory compliance, Saifr, a regulatory tech provider within Fidelity Investment’s innovation incubator, Fidelity Labs, has developed a set of four models: Retail Marketing Compliance, to ensure marketing materials adhere to industry regulations and standards; Risk Interpretation, to identify potential risks in marketing content; Language Suggestion, to enhance the compliance of marketing messages; and Image Detection, to analyze and verifying the appropriateness of images used in marketing campaigns.

.