IBM collaborates with Watsonx to lead the development of enterprise-level artificial intelligence.

2023-09-12

This week, IBM held its annual TechXchange conference in Las Vegas. Although the agenda was geared towards all IT professionals, it clearly emphasized how IT organizations can harness the power of artificial intelligence to transform their businesses. IBM is in a favorable position to lead these conversations, having recently released a suite of AI solutions.

IBM's updated AI product portfolio, called "watsonx," is an enterprise-focused AI and data platform that includes three interconnected solution stacks:

  • watsonx.ai Studio focuses on foundational models, generative AI, and machine learning.
  • watsonx.data is a data repository based on an open lakehouse architecture.
  • watsonx.governance provides a toolkit for responsible AI workflows (coming soon).

Let's take a look at what the watsonx product portfolio offers.

watsonx.ai

IBM watsonx.ai allows AI developers to leverage models provided by IBM and the Hugging Face community to tackle various AI development tasks. These models are pre-trained and suitable for handling various natural language processing (NLP) tasks, including question answering, content generation, summarization, text classification, and data extraction. Examples include:

  • Retrieval-Augmented Generation (RAG): RAG can develop context-aware chatbots and question answering capabilities based on customer-specific content.
  • Summarization: Transforms text of domain-specific content into personalized answers.
  • Content Generation: Text generation for specific purposes, such as marketing and other business content.
  • Named Entity Recognition: Identifies and extracts important information from unstructured text.
  • Insight Extraction: Analyzes existing unstructured text content to generate domain-specific insights.
  • Classification: Reads and classifies written input without the need for examples.

An upcoming release will expand the range of proprietary foundational models trained by IBM, enabling efficient specialization in specific domains and tasks. IBM's watsonx also offers various AI models for natural language processing tasks from IBM and the Hugging Face community.

Last month, IBM expanded watsonx.ai by announcing support for Meta's Llama 2-chat 70 billion parameter model in the watsonx.ai studio. This collaboration builds on their joint work in open AI innovation, including projects like PyTorch and Presto.

Adding Llama 2 to watsonx.ai is an important step in IBM's generative AI roadmap, with plans to introduce more AI models and features. IBM prioritizes trust and security, allowing users to eliminate harmful language with AI safeguards.

watsonx.data

IBM's watsonx.data aims to help customers overcome challenges related to data scale, complexity, cost, and governance in order to scale their AI workloads. The platform enables users to seamlessly access their data stored in the cloud or within the enterprise through a single entry point. This approach greatly simplifies data access for non-technical users while ensuring security and compliance.

Furthermore, the significance of watsonx.data extends beyond data scientists and engineers. It empowers non-technical users to self-serve access enterprise-grade trusted data on a unified collaboration platform, granting them greater capabilities. Simultaneously, it strengthens security and compliance through centralized governance and local automated policy enforcement.

In the near future, watsonx.data will leverage the capabilities of watsonx.ai foundational models to simplify and expedite user interactions with data. This will enable users to perform tasks such as discovery, enhancement, refinement, and visualization of data and metadata using natural language, resulting in a more conversational and user-friendly experience.

watsonx.governance

As artificial intelligence becomes increasingly integrated into everyday workflows, implementing proactive governance within organizations to ensure responsible and ethical decision-making becomes more necessary. Watsonx.governance utilizes IBM's powerful AI governance capabilities to help organizations implement comprehensive end-to-end lifecycle governance, mitigate risks, and effectively manage compliance with evolving AI and industry regulations.

Watsonx.governance empowers organizations to lead and oversee their AI initiatives. The tool enhances your ability to reduce risks, address regulatory requirements, and consider ethical factors through software automation, without the need for expensive data science platform transformations for models created using third-party tools.

IBM's watsonx.governance is currently in the technical preview stage and is expected to be officially available later this year.

watsonx Infrastructure

Today, almost every enterprise's IT infrastructure is a hybrid cloud infrastructure. IBM recognizes this and offers watsonx capabilities to both on-premises and cloud users.

For example, IBM partnered with Amazon Web Services (AWS) to bring watsonx.data into the AWS infrastructure. Enterprises can accelerate their cloud data modernization efforts by leveraging the openness, performance, and governance of IBM watsonx.data alongside the scalability, flexibility, and cost-effectiveness of AWS cloud infrastructure.

IBM provides this flexibility by building watsonx on its enhanced version of Red Hat OpenShift technology, ensuring support for more efficient foundational model workloads. OpenShift also enables seamless integration of watsonx with various IBM infrastructure products.

While the cloud brings significant benefits to IT organizations, enterprises have rich expertise in using IBM's high-performance computing, including its Z series mainframes and IBM "Vela" AI supercomputer, when specialized servers, storage, and accelerators designed for AI workloads are required.

IBM also plans to enhance its IBM Cloud to better support the needs of AI workloads. The company has informed us that it expects to launch a comprehensive, high-performance, and adaptive infrastructure service for foundational models optimized for AI training and deployment through IBM Cloud in the coming months.

RECENT AI NEWS

RECENT AI TOOLS