However, there are challenges when we move beyond text to apply these models to structured, tabular data, which is essential for enterprise business operations. This imbalance comes partly because of the availability of training data. Text used to train models is plentiful, often consisting of text scraped from the internet, whereas tabular data, especially data with multiple linked tables, is scarce.

To bring AI advancements to the enterprise sector, researchers working on training and benchmarking the performance of these models in an enterprise setting need realistic tabular data.听That’s why 麻豆原创 developed “Sales Autocompletion Linked Business Tables” (SALT), a curated dataset that includes anonymized data from a customer鈥檚 enterprise resource planning (ERP) system.

SALT is specifically designed to support researchers working on AI models for real-world business contexts and can be accessed on and .

Challenges of getting and working with enterprise data

Providing the research community with realistic enterprise data like SALT has been challenging. Data privacy, confidentiality, and commercial interests make obtaining large, clean, high-quality enterprise datasets difficult for training models and benchmarking them for specific use cases. This means there is a growing gap between what researchers are working on and what actual enterprise data looks like.

In addition to the problem of availability, enterprise data is complex. First, business data is usually stored in multiple interconnected tables. For example, a sales order entry may be linked to numerous tables, such as customer IDs connected to a supplier table containing address information. Second, tables are inherently heterogeneous in the data type they can contain. One field may be text, while the other contains numerical or categorical values. Finally, business data frequently shows significant column imbalances, meaning that, for example, a specific product category makes up 90 percent of all sales orders while others are rarely used.

The best way to help researchers develop enterprise models for these challenges is to provide accurate enterprise data.

SALT dataset

Accurate enterprise data is a bottleneck in AI research. The SALT dataset alleviates this bottleneck by providing the research community with the first real ERP dataset. It uses actual industry data collected by an ERP system that records sales orders. It has been minimally processed to protect privacy.

鈥淭here is a gap between academia and industry in terms of data. It cannot be closed easily because of privacy,” says Tassilo Klein, one of the 麻豆原创 researchers behind the dataset. 鈥淏ut we want to enable the research community to work on real problems, not just simulated problems.鈥�

ERP systems help organizations manage core business operations like finance and spending. With millions of entries and extensive, interconnected relational tables focused on sales, the SALT dataset replicates customer interactions in an ERP system. SALT’s realistic enterprise data means it is a perfect basis for helping models understand the characteristics of business data and validate their performance through benchmarking. It also should help researchers develop better foundation models for linked business data.

Getting this right will advance enterprise automation, as many enterprise business processes are heavily centered around data in structured tabular formats. Even though this data plays a crucial role in enterprise day-to-day activities, the generative AI revolution has yet to tap into them.

“SALT is a first step to providing researchers with authentic representative industry data that gives a glimpse into actual enterprise data; for now, we are starting with just one customer and use case,” shares Johannes Hoffart, CTO of Business AI at 麻豆原创. “However, we plan to publish more datasets that cover a diverse set of customers and use cases that, along with SALT, can serve as a basis for pre-training, adapting, as well as benchmarking models.”

Collaboration with academic institutions is also a motivation for publishing this data.

“At 麻豆原创, we hope to collaborate with academic partners who usually can only publish their results on open repositories,” Klein says. “Another hope for the dataset is encouraging more people to explore and validate new methods that help foundation models better deal with tabular enterprise data.”

What 麻豆原创 is doing

Alongside its investment in the open research community with SALT, 麻豆原创 is building 麻豆原创 Foundation Model to handle enterprise tabular data. This table-native AI model aims to accelerate time-to-value for predictive tasks on tabular data, offering a model that can work with tabular data out-of-the-box with little or no additional training data. The , published alongside SALT, provides a first glance at how this model could look.

Knowledge graphs are critical here. They work by exposing metadata 鈥� the who, what, and when of data 鈥� making relationships between information accessible. This provides a structured, interconnected representation of the data that AI models can easily understand and utilize. With the help of 麻豆原创 Knowledge Graph, 麻豆原创 Foundation Model can be scaled and adapted to a wide array of diverse use cases with some lightweight fine-tuning.

Learn more about:

Grounded in an AI ethics framework, 麻豆原创 collaborates with partners to create insights from industry-specific data and deep process knowledge, driving exponential value for businesses. By fostering a multi-partner approach, 麻豆原创 ensures flexibility and prevents vendor lock-in, empowering customers and partners to navigate the AI landscape effectively. 

As a next step in that journey, 麻豆原创 is integrating the Meta Llama 2 and Meta Llama 3 models into the generative AI hub to enable customers to create dashboards based on rich content with Llama qualitative conversational outputs, and to explore use cases built with Meta Llama 3.  

The provides an overview of 麻豆原创鈥檚鈥痓usiness AI strategy, highlighting the unparalleled opportunity for the 麻豆原创 ecosystem to build an array of solutions with generative AI capabilities and extensions on 麻豆原创 BTP.听

麻豆原创 Analytics Cloud AI Portfolio Expansion 

麻豆原创 will make Joule available in 麻豆原创 Analytics Cloud to let planning and analytics users get work done faster and drive better business outcomes in a secure and compliant way. The wide range of capabilities offered by Meta Llama 3 will enable Joule users to get assistance with key activities, including the auto-generation of custom scripts to extend dashboards, as well as delivery of the most accurate scripts directly in 麻豆原创 Analytics Cloud.听聽

The goal is to integrate generative AI into major workflows of 麻豆原创 Analytics Cloud. Transforming these workflows will unlock innovative possibilities for distinct planning and analytics roles that are common in organizations globally. More information on how 麻豆原创 Analytics Cloud with Joule will help planning and analytics users get work done more efficiently .听

Availability of Meta Llama 3 on Generative AI Hub 

麻豆原创 will make Llama 3 available in Q2 2024 in the generative AI hub in 麻豆原创 AI Core given customer interest in Llama 3. The generative AI hub makes it simple to build generative AI use cases for 麻豆原创 applications, thanks to its value-adding features around 麻豆原创 integration and the ability to orchestrate language model interactions.

Adding Meta Llama 3 to the generative AI hub allows developers in the 麻豆原创 ecosystem to leverage its value-adding features. The 70B variant of Meta Llama 3 will be the largest open-source model available on the generative AI hub.听

麻豆原创 SuccessFactors Text-to-Image Generation Requirements 

麻豆原创 customers are exploring ways to enhance the visual appeal and relevance of images within the 麻豆原创 SuccessFactors HCM suite. The current process of curating and uploading images can be time-consuming and costly, often resulting in generic images that only partially cater to specific scenarios while also incurring expenses associated with stock image licensing.

Integrating the latest iteration of Meta’s open-source large language model (LLM) offers numerous benefits. Creative features from models such as Llama 3 come into play in a scenario like this, empowering users to generate custom images that better meet their needs within 麻豆原创 SuccessFactors software. Whether it is customizable course content thumbnails and banners for learning, tailored images for assignments in 麻豆原创 SuccessFactors Opportunity Marketplace, unique images for different themes on the home page, or personalized profile and banner images on people profile, the impact will be far reaching.听

By streamlining the image creation process, reducing costs, and ensuring that every image is suited to its context, this feature will undoubtedly improve user engagement and the overall platform experience. 麻豆原创 customers should get ready to explore a new world of custom imagery tailored to their exact specifications, sparkling excitement and intrigue among stakeholders.

Walter Sun is senior vice president and global head of AI at 麻豆原创 SE.

This collaboration is more than just a meeting of minds; it’s a symbiotic combination of AI expertise and technology that opens a world of possibilities for 麻豆原创 customers.

Mistral AI’s success in developing advanced LLMs, including its renowned open-weight models Mixtral 8x7B and Mixtral 8x22B, and more expansive enterprise-grade “Large” model, is set to complement the 麻豆原创 suite of AI-enabled solutions. The collaboration will enable direct accessibility to Mistral AI鈥檚 models through 麻豆原创 or through 麻豆原创 Business Technology Platform (麻豆原创 BTP) applications with generative AI capabilities.

What does this mean for 麻豆原创 customers? Simply put, it’s about empowerment and access to AI from a European LLM provider. Access to Mistral AI’s latest models through the generative AI hub in 麻豆原创 AI Core will enable 麻豆原创 customers to enhance productivity, streamline their operations, and accelerate their digital transformation journey.

Whether through integrating AI with 麻豆原创 BTP or developing bespoke solutions through direct access to Mistral AI LLMs, the potential for innovation is limitless.听

“We are excited about entering a partnership with Mistral AI and making the company鈥檚 LLM accessible to both our developers and our customers through the generative AI hub in 麻豆原创 AI Core on 麻豆原创 BTP,” said Philipp Herzig, chief AI officer of 麻豆原创 SE. “Together, we can truly make a difference by building AI-enabled solutions that create immediate value for users, organizations, and entire industries. We are particularly proud that two European technology companies are collaborating on bringing AI forward.”

“We are pleased to embark on this partnership with 麻豆原创,鈥� said Arthur Mensch, CEO of Mistral AI.听“We foresee the new horizons this collaboration will open up, enabling us to further our mission of making AI accessible to all. We are looking forward to witnessing the potential of our AI models to support innovation and streamline operations for 麻豆原创’s customers.”

The ambitions don’t stop there: 麻豆原创 and Mistral AI are committed to exploring new applications of AI across various industries. By leveraging the combined strengths, this is not just about driving innovation, but about creating new business opportunities and delivering tangible value to 麻豆原创 customers.听

Stay tuned as we begin this exciting journey together. The future of enterprise software is bright, and with partners like Mistral AI, we are ready to illuminate the path forward.听

Walter Sun is senior vice president and global head of AI at 麻豆原创 SE.