The following features enable businesses to leverage AI technology optimized for the IBM Z platform to accelerate your AI journey. Data scientists can find benefits from familiar tools accelerated for IBM Z. AI engineers may find ways to improve their development cycles with the IBM Z integrations.
Overview:
IBM has delivered a range of AI capabilities, empowering application architects to effectively deploy AI solutions that enhance the AI on z/OS ecosystem. The following capabilities are designed to optimize the deployment and utilization of AI applications on z/OS:
- Python AI Toolkit. Python AI Toolkit for IBM z/OS 1.13.0 release is delivering IBM-owned open-source Python packages, relevant for AI and Machine Learning (ML) workloads running on IBM z/OS with the latest Python SDK 3.13.x.
- IBM Z Platform for Apache Spark. IBM Z Platform for Apache Spark v1.1 (zSpark), generally available today, has exciting capabilities that IBM delivered in 3Q2024. New with zSpark, there is a currency update from Apache Spark 3.2 to Apache Spark 3.5.1 allowing to stay in line with the open-source community.
- IBM Synthetic Data Sets. IBM Synthetic Data Sets are artificially generated pre-built data sets designed to help jumpstart AI projects in the financial services industry by enhancing AI models with rich, labelled data, and enabling the creation of solutions with synthetic transactional data. This offering is intended to avoid using real client seed data that could include Personal Identifiable Information (PII), allowing to keep private data secure.
----------------------------------------------------------------------------------------------------------------------------
Description:
Python AI Toolkit
Python AI Toolkit for IBM z/OS 1.1.6 is upgrading to Python AI Toolkit for IBM z/OS 1.13.0 to align version numbers with the IBM Open Enterprise SDK and open-source Python interpreters. Python AI Toolkit for IBM z/OS 1.13.0 supports IBM Open Enterprise SDK version 3.13.x and the Python 3.13.x interpreter. New function PAIT v1.13.0 will include py_zcrypto which enables developers to implement cryotographic operations within their Python applications and other currency upgrades. See the official announcement for more information.
IBM Z Platform for Apache Spark
IBM Z Platform for Apache Spark (zSpark) supports Apache Spark the latest version of Apache Spark v3.5.1. This update ensures compatibility with the latest Spark features. Moreover, zSpark has incorporated security enhancements. These include updates to the secure all ports documentation, including a new feature that allows web UI ports to be covered with ATTLS (Application Transparent Transport Layer Security). This feature bolsters security by providing an additional layer of encryption and authentication, thereby safeguarding sensitive data and maintaining compliance with stringent security standard. zSpark now supports Java 17, which extends the Java support to Java 8, Java 11 and Java 17. Please note that IBM strongly recommends to use Java 11 and Java 17. Find out more about IBM Z Platform for Apache Spark 1.1.0 - IBM Documentation. With the PTF for APAR PH59867, Spark 3.5.1.0 is now available.
IBM Synthetic Data Sets
IBM Synthetic Data Sets are a family of enterprise-grade artificially generated data sets that need no real client source or seed data and are specifically created for AI model training and solutions that benefit IBM Z and LinuxONE customers. These Synthetic Data Sets are easily downloadable pre-built enterprise-grade .CSV data sets that require low client effort, are platform agnostic, include IBM industry expertise and domain knowledge, and don’t use any real client seed data which avoids the usage of Personal Identifiable Information (PII) to keep private data secure. The primary objective of these customized data sets is to facilitate real-time AI applications on IBM Z and LinuxONE (such as fraud detection, anti-money laundering, and insurance claims processing) that enable the generation of valuable business insights while strictly adhering to privacy regulations and safeguarding sensitive information. Thus, organizations can keep real data secure from threats by training models with artificial data and meet industry requirements by leveraging data that uses no real PII and therefore requires no encryption or redaction.
Leverage IBM Synthetic Data Sets to expedite your AI initiatives and promptly initiate AI projects, as these data sets are suitable for constructing, refining, and validating AI models, thereby fostering a robust and efficient AI development process. For more information, read the full announcement letter, visit the IBM Synthetic Data Sets product page, and read the IBM Redbooks publication.
Questions? Reach out to start a conversation!