IBM spendiert der KI- und Datenplattform watsonx neue generative KI-Fundamentalmodelle und Erweiterungen. |
IBM is giving its watsonx AI and data platform new generative AI foundation models and extensions. |
---|---|
IBM setzt große Hoffnungen in watsonx und baut das Angebot weiter aus. Die KI- und Datenplattform ist mit einer Reihe von KI-Funktionen ausgestattet, die Unternehmen dabei helfen, KI-Ergebnisse zu skalieren und zu beschleunigen. Diese Erweiterungen umfassen eine Technologievorschau für watsonx.governance, neue generative KI-Datendienste für watsonx.data und die geplante Integration von watsonx.ai Foundation Models in ausgewählte Software- und Infrastrukturprodukte.
Entwickler können viele dieser neuen Funktionen und KI-Modelle vom 11. bis 14. September auf der TechXchange, IBMs wichtigstem technischen Event in Las Vegas, erleben. Zu den neuen generativen KI-Modellen von IBM und Drittanbietern, die auf watsonx.ai verfügbar sind, gehören:
IBM hat für die Foundation Models einen Trainingsprozess eingerichtet, der auf den Grundsätzen von Vertrauen und Transparenz basiert und der mit einer konsequenten Datenerfassung beginnt und mit Kontrollpunkten endet. Das ermöglicht einen verantwortungsvollen Einsatz von Models und Anwendungen zur Governance, Risikobewertung, Datenschutzbedenken, Bias-Minderung und Compliance. IBM kündigt auch Pläne zur Einführung neuer Funktionen auf der watsonx -Plattform an. Watsonx.ai:
Watsonx.data:
Watsonx.governance:
„Wie die kontinuierliche Weiterentwicklung der watsonx-Plattform innerhalb weniger Monate seit der Markteinführung zeigt, unterstützen wir unsere Kunden während des gesamten KI-Lebenszyklus“, sagt Dinesh Nirmal, Senior Vice President, Products, IBM Software. „Als Transformationspartner arbeitet IBM mit Kunden zusammen, um ihnen dabei zu helfen, KI auf vertrauenswürdige Weise zu skalieren. Das reicht von der Unterstützung bei der Einführung grundlegender Elemente ihrer Datenstrategien über die Abstimmung von Models für ihre spezifischen Geschäftsanwendungen bis hin zur Unterstützung bei der Model-Steuerung darüber hinaus.“ Die IBM watsonx KI- und Datenplattform wird durch eine Reihe von KI-Assistenten ergänzt. Diese unterstützen Kunden dabei, die Vorteile der KI mit vertrauenswürdigen Daten in wichtigen Anwendungsfällen zu skalieren und zu beschleunigen, wie zum Beispiel:
IBM plant außerdem, die Innovationen von watsonx.ai in ihre hybriden Cloud-Software- und Infrastrukturlösungen einzubinden, darunter:
|
IBM has high hopes for watsonx and continues to expand its offerings. The AI and data platform is equipped with a number of AI capabilities that help companies scale and accelerate AI results. These enhancements include a technology preview for watsonx.governance, new generative AI data services for watsonx.data, and planned integration of watsonx.ai Foundation Models into select software and infrastructure products.
Developers can experience many of these new features and AI models Sept. 11-14 at TechXchange, IBM’s premier tech event in Las Vegas. New generative AI models from IBM and third-party vendors available on watsonx.ai include: – Granite Series Models: IBM plans to unveil the Granite Series models later this month. These use the „decoder“ architecture that gives today’s Large Language Models (LLMs) the ability to predict the next word in a sequence. They can also support enterprise NLP tasks, such as summarization, content generation and insight extraction. IBM plans to provide a list of data sources and a description of the data processing and testing steps. These were performed to create training data for the Granite Models series. (Planned availability in Q3 2023). – Third-party Models: IBM now offers the Metas Llama 2 chat model with 70 billion parameters and the StarCoder LLM for code generation in watsonx.ai in the IBM Cloud. (Available now) IBM has established a training process for Foundation Models based on the principles of trust and transparency that begins with consistent data collection and ends with checkpoints. This enables responsible use of Models and applications for governance, risk assessment, privacy concerns, bias mitigation and compliance. IBM also announced plans to introduce new capabilities to the watsonx platform. Watsonx.ai: – Tuning Studio: IBM plans to deliver the first version of Tuning Studio, which will include prompt tuning – an efficient, cost-effective way for clients to customize Foundation Models with their own enterprise data to their specific downstream tasks. (Planned availability in Q3 2023). – Synthetic Data Generator: IBM today launched a synthetic data generator that helps users create artificial tabular datasets consisting of custom data schemas or internal datasets. In this way, users can gain insights for training AI models with less risk. This facilitates decision making and accelerates time to market. (Available now) Watsonx.data: – Generative AI: IBM plans to integrate the generative AI capabilities of watsonx.ai into watsonx.data. The goal is to help users discover, extend, visualize and refine data for AI. This is done through a self-service offering based on a conversational natural language user interface. (Planned tech preview in Q4 2023). – Vector database feature: IBM plans to integrate a vector database feature into watsonx.data to support use cases for advanced querying of watsonx.ai. (Planned tech preview in Q4 2023). Watsonx.governance: – Model Risk Management for Generative AI: IBM is launching a tech preview for watsonx.governance. Customers can explore features for automated capture and documentation of foundation model specifics, as well as model risk management features, at the tech preview. These enable stakeholders to view relevant metrics in dashboards of their enterprise AI workflows with approvals so that people are engaged at the right time. „As evidenced by the continued evolution of the watsonx platform within months of launch, we support our clients throughout the AI lifecycle,“ said Dinesh Nirmal, senior vice president, Products, IBM Software. „As a transformation partner, IBM is working with clients to help them scale AI in a trusted way. This ranges from helping them implement foundational elements of their data strategies, to tuning models for their specific business applications, to helping them model control beyond that.“
The IBM watsonx AI and data platform is complemented by a suite of AI assistants. These will help clients scale and accelerate the benefits of AI with trusted data in key use cases, such as: – Application Modernization: IBM watsonx Code Assistant products launching later this year will use customized Foundation Models to convert code and generate recommendations for developers. IBM recently announced two AI-powered code products: watsonx Code Assistant for Z, to increase developer productivity and accelerate COBOL application modernization, and watsonx Code Assistant for Red Hat Ansible Lightspeed, to help developers of all levels write Ansible playbooks. – Customer support: IBM watsonx Assistant will help deliver consistent and intelligent customer service solutions with conversational AI. For example, IBM Support Insights Pro, expected to be available later this month, will leverage watsonx Assistant to help clients gain insights into their multivendor IT infrastructures, proactively assess support patterns and remediate risks. This leads to higher availability and security. – HR and personnel: IBM watsonx Orchestrate helps HR professionals automate repetitive tasks and back-office processes, such as scheduling interviews or posting job openings through an interaction platform. IBM also plans to incorporate watsonx.ai innovations into its hybrid cloud software and infrastructure solutions, including: – Intelligent IT Automation: IT automation products Instana and AIOps Insights, which will enter tech preview next week, include Intelligent Remediation. This incorporates generative AI Foundation Models from watsonx.ai to help IT Ops professionals summarize incident details and make workflow suggestions so technicians can quickly implement solutions. – Developer Services for watsonx: The SAP ABAP SDK for watsonx is expected to make it easier and faster for developers to more accurately apply functions of watsonx to their organizations‘ data on IBM Power for SAP workloads. It is expected to provide additional ways for clients to use AI to draw conclusions from their data on Power systems and to apply AI algorithms to their most sensitive data and transactions. (Planned availability in Q1 2024). |
Dr. Jakob Jung ist Chefredakteur Security Storage und Channel Germany. Er ist seit mehr als 20 Jahren im IT-Journalismus tätig. Zu seinen beruflichen Stationen gehören Computer Reseller News, Heise Resale, Informationweek, Techtarget (Storage und Datacenter) sowie ChannelBiz. Darüber hinaus ist er für zahlreiche IT-Publikationen freiberuflich tätig, darunter Computerwoche, Channelpartner, IT-Business, Storage-Insider und ZDnet. Seine Themenschwerpunkte sind Channel, Storage, Security, Datacenter, ERP und CRM.
Dr. Jakob Jung is Editor-in-Chief of Security Storage and Channel Germany. He has been working in IT journalism for more than 20 years. His career includes Computer Reseller News, Heise Resale, Informationweek, Techtarget (storage and data center) and ChannelBiz. He also freelances for numerous IT publications, including Computerwoche, Channelpartner, IT-Business, Storage-Insider and ZDnet. His main topics are channel, storage, security, data center, ERP and CRM.
Kontakt – Contact via Mail: jakob.jung@security-storage-und-channel-germany.de