Artificial Intelligence and Machine Learning

AndreasForster

If you have a collection of FAQs that you want to be easily accessible for your business users, then a chatbot might be the answer. This blog explains how to create such a (non-hallucinating) chatbot on SAP's Business Technology Platform by leveraging the Generative AI Hub and SAP HANA Cloud's vector engine.

L_Skorwider

Hi Community, At the end of last week, Meta released the long-awaited Large Language Model - LLAMA3. What is unusual about it? The fact that it's open source and really delivers. For now we got the 8B and 70B versions, but 400B is also in the preparation. In my opinion, this is a big deal. And how is it doing in SAP's backyard? I won't reveal - see for yourself.

ChristophMorgen

With the SAP HANA Cloud 2024 Q1 release, several new embedded Machine Learning / AI functions have been released with the SAP HANA Cloud Predictive Analysis Library (PAL), for classification, regression, time series forecasting or financial data analysis on structured data stored in the SAP HANA Cloud database.

Sergiu · ‎10-23-2023

The blog State of GenAI in the SAP Community 09/2023 is a great analysis of the applicability of GenAI. I had the following conversation, which you can find in the comments: Sergiu: How did you collect statistics for Overview SAP Community Insights from 82 blogs? Peter: I would love to say I used a sophisticated method like accessing an API or at least did web scraping. But I simply invested one evening to search for the right keywords and another one to manually collect the numbers (likes/views). Sergiu: I thought you succeeded in giving a set of prompt instructions to ChatGPT Plus to access the web and return the results in CSV or JSON. One day we will get that without coding. 🙂 Open the online application SAP AI Blog Statistics I developed for collecting statistics on the SAP AI Blog. Challenge yourself to achieve the same results with Generative AI without coding or web scraping apps. Table of content Web scraping with Generative AI Online app SAP AI Blog Statistics Conclusions Web scraping with Generative AI Large Language Models (LLM) are an impressive piece of technology. Generative AI applications like ChatOpenAI and MONICA are great copilots for coding. ChatGPT Plus and MONICA Pro have a web access option which is great at summarizing; however, they can’t perform specific tasks like web scraping of statistics. With the prompt instruction “Collect the number of views,” I received the answer, “As an AI language model, I don’t have access to the exact number of views mentioned in the blog post.” Have you engineered a prompt to extract statistics of SAP blogs? Is it possible? Share the solution in the comments. Online app SAP AI Blog Statistics For now, you can collect statistics with Python code. 🐍 😊' Open the SAP AI Blog Statistics online application and select one of the options: SAP AI Gen State extracts the statistics for the list of blogs provided in the comments of the blog State of GenAI in the SAP Community 09/2023. SAP HANA ML Cloud Challenge 2022 extracts statistics for the blog series I wrote after I participated in the SAP HANA ML Challenge – Employee Churn and I came in second place. Upload the SAP blog list from the file. Select the list of blogs or upload a list of blogs and press the button extract. You can download the results in CSV or JSON formats. Screenshots: Select an option. Press Extract and wait for results. URLs are clickable, and columns are sortable. Press Download CSV or Download JSON to save Blog Statistics to a local file. In the GitHub repository LLM-Challenge, you can find the Jupyter Notebook for a playground and the class SapBlogStatistics. This allows you to integrate them into other applications. Conclusions Humans designed computer formats and languages to store, process, and transmit the information we perceive naturally. Somehow, we think that from now on, we can rely on LLMs to achieve any past and future tasks with computers using prompt instructions with texts or images so we never have to type keyboards thousands of line codes, test and debug hundreds of times. How will computer formats and languages evolve without further human contribution? Have you engineered a prompt to extract statistics of SAP blogs? Is it possible? Share the solution in the comments. What would you prefer for web scraping in terms of efficiency: LLM or apps and code? Why is the web scraping with generative AI important? I think achieving the extraction of data from specific elements of a web page leads to the next step of inserting data into web forms, which brings us into the territory of web automation with natural language instructions. With one more step of deploying generative AI into an application that interacts with other desktop applications, we can automate the entire workspace with natural language instructions. Discovering the capability of LLMs to work with specific elements could enhance user interaction from multi-screen apps operated by keyboard and mouse to the interaction of apps orchestrated by natural language. Update on November 09, 2023 Guess what, Bill Gates envisions the following "In the next five years, this will change completely. You won’t have to use different apps for different tasks. You’ll simply tell your device, in everyday language, what you want to do". 🙂 Read the article by Bill Gates "AI is about to completely change how you use computers". Enjoy! 🎈🎈🎈 Blog series SAP HANA Cloud Machine Learning Challenge “I quit!” – Understanding metrics Could machine learning build a model for prime numbers? “Hello, world!” your crafted chat GPT bot! SAP Machine Learning Embedding in OpenAI Building Trust in AI: YouTube Transcript OpenAI Assistant AI Challenge | Web scraping with Generative AI

amycross

SAP Business AI is already embedded in the SAP applications that customers use to run their most critical business processes. With SAP Business AI, you are empowered to interact with SAP software through Joule – SAP’s copilot available across SAP applications.

arabella

Discover AI-driven business solutions with SAP Business AI at our free-to-attend virtual session: unlock untapped potentials, gain competitive edge, and equip yourself to thrive in the AI-empowered future.

WojciechLewando

Hello experts, I am new to AI Core product, and I am trying to make some scenarios, but I faced problem that is stopping me for a while now. Background: I have been doing tutorial about SAP BTP AI Core: https://developers.sap.com/group.ai-core-get-started-basics.html Everything was working fine till the tutorial “Make Predictions for House Prices with SAP AI Core”: https://developers.sap.com/tutorials/ai-core-deploy.html There is step about to loading trained model to make predictions (I will refer to this later): model = pickle.load(open ('/mnt/models/model.pkl','rb')) Then there is the rest of the code for creating endpoints and using this model for prediction. Further configuration I did correct (Docker, yaml etc.), according to tutorial. The service started but was not working (error 500 when calling endpoint). My tests and analysis: I couldn’t find any potential problem for that, I rechecked everything and the web service from tutorial was still not working. I decided to create my own model and configuration for a very simple example of model which is instance of class making addition (very simple one, just for tests). But this time I created some more endpoints for this web service to check if problem is only with the one using (and loading) model: Again, everything was working except the part of using the model, so: The AWS S3 Bucket has this model (I check it manually using console) and path is correct. Python code is fine because I successfully tested it locally. Git with YAML file is correct because the service is running and other endpoints working. All actions in AI Core (deployments, executions etc.) were also successful. I implemented more functionalities into this webservice, so I was able to confirm that all other endpoints were working correctly and only the one loading model was returning error 500. One of the functionalities was defined as this: And the output was: AI Core Server: __pycache__|main.py|requirements.txt| So basically, there is no other folder (should be /mnt) or file in this location… Of course, I don’t know how exactly AI Core works internally, so maybe this approach for testing is wrong, but I also don’t know how else I can check this, when I see this log: “INFO:root:Successfully copied s3://hcp-… to /mnt/models” The location is 100% correct, I can see (and copy) this model using direct access to my AWS S3 Bucket. I did this and deployed locally, and it is working fine. My conclusion is that the model is not copied successfully (the path doesn’t even exist), or I am making some error in the implementation of this functionality. Question:s Is my approach for testing this correct (should this python function return all files and folders for my instance of AI Core in root location /)? How else I can debug/analyze this problem? Have you ever faced a similar issue? Is it possible that the free version of AI Core has some limitations like this or a hidden non-deterministic bug? I fount very similar problem here: https://community.sap.com/t5/technology-q-a/sap-ai-core-error-while-prediction/qaq-p/12769557 This problem has not been resolved on that forum as I see. I followed hints from there but with no results. Best regards

YatseaLi

In this blog post series, we will explore various options for running popular open-source Large Language Models to like Mistral, Mixtral, LlaVA, Gemma, Llama 2, etc., in SAP AI Core, which complements SAP Generative AI Hub with self-hostsed open-source LLMs. We'll utilize widely adopted open-source LLM tools or backends such as Ollama, LocalAI, llama.cpp and vLLM, among others, through the BYOM (Bring Your Own Model) approach.

Chr_Vogler · ‎09-05-2023

Customer behavior, post-pandemic economic recovery, extreme heatwaves, and drought due to climate change, inflation, supply chain disruption, and an energy crisis have caused many challenges for companies. This greatly influences the innovations that we bring to each SAP S/4HANA Cloud, public edition release, and our shift to include more of SAP´s Intelligent Technologies to help companies transform their processes and to better support them on the path to sustainable growth. In the following blog, I will give you an engineering expert view of some selected highlights of our SAP S/4HANA Cloud, public edition for Sales, to demonstrate that automation, system integration, and end-to-end transparency can help companies to transform their order-to-cash process with the shift to low-touch sales order processes. Background Low-touch sales order processing is designed so that the user needs little or no interaction with the system during the process. It is a process that takes place mainly automatically, each step must be thought out and designed to be as easy as possible for the user to pass without assistance. This concept brings several advantages: it leaves more room for spending time on “high-value” tasks by employees, instead of repetitive tasks with little added value; it reduces or eliminates the need for input moments in the system, reducing or eliminating errors and increasing efficiency; and it reduces the time needed to process orders. Automatic creation of sales orders from unstructured data A significant number of sales orders are captured from an unstructured format like PDF documents. To generate the sales order, the internal sales representative needs to check purchasing details in these files, identify sales data based on the purchasing data (for example, identify the sold-to party based on the purchaser’s name and address), and enter the sales data manually in the structured format as required to be processed into SAP S/4HANA Cloud, public edition. https://youtu.be/bAqARTEoSho Video 1: Get an overview of the automatic creation of sales orders from unstructured data. Creating sales order from unstructured data allows the extraction of the sales order data from unstructured PDF sources and the creation of the sales order automatically. The Business Benefits are very clear Cost Savings by reducing the manual effort to create sales orders Improvement of data quality by sales order process automation Automation shifts efforts toward better service quality The first version of this innovation was delivered with the SAP S/4HANA Cloud, public edition, 2011 release, and with SAP S/4HANA Cloud, public edition, 2102 release, the app has been enhanced to simulate order creation and to show these order simulation results, as well as to enable you to create custom logic (BAdI) to overwrite changeable sales order request fields based on your custom proposal rules. https://youtu.be/qDxfHhUkUQE Demo 1: Enhanced 'Create Sales Orders – Automatic Extraction' SAP Fiori app (note: no audio voice available). What is new with the SAP S/4HANA Cloud, public edition, 2102 release? Processing status is now visualized as micro process flow in the SAP Fiori list page Customers can now influence the result of master data determination via a Cloud BADI. For example, propose a sales area or sales order type according to the customer's own logic. Ship-to party is now determined from ship-to address automatically, customer can double check or change the determined the ship-to party Before creating a sales order from a sales order request, you can now simulate the order creation and get an overview of the to-be sales order: Simulated net value vs. extracted net value The actual sales order items to be created from the sales order request items The last simulation result will be persisted in the sales order request as a reference to your next actions. With the SAP S/4HANA Cloud, public edition, 2208 release, we continue our roadmap to bring Intelligence for Sales to automate routine tasks, and you can now open the purchase order file of a sales order request in an embedded pane and check file contents alongside the sales order request data. The embedded file view highlights extracted data by default, and you can hover over a highlighted area to show an extracted field or click on the eye icon to show all extracted fields. Previously, it was only possible to open the file in a new window. https://sapvideoa35699dc5.hana.ondemand.com/?entry_id=1_kasq8bql Demo 2: Increase salesforce efficiency and accuracy using the SAP Fiori app "Create Sales Orders - Automatic Extraction (including audio). Enhancements since SAP S/4HANA Cloud, public edition, 2011 release Since the first version of this innovation, we have brought new features for almost every release. Below you will find a summary: What’s New in SAP S/4HANA Cloud, public edition, 2105 release: You can upload image files in JPG, JPEG, TIF, TIFF, or PNG format. You can search for sales order requests using the order simulation filter. You can view the change log of a sales order request. The log shows changes that have been made to the sales order request fields. You can view the application log of a sales order request. The log shows system processing details regarding data extraction, data proposal, and BAdI calls. More fields are available in simulated item data, such as the item category, confirmed quantity, and confirmed delivery date. The product can be determined with reference to order creation history What’s New in SAP S/4HANA Cloud, public edition, 2108 release: View and change the requested delivery date in sales order request items. Control whether the ship-to party is required for creating sales orders in the configuration activity Adapt Settings by Company Code. Use a bot from the SAP Intelligent Robotic Process Automation (RPA) Factory to automate the creation of sales order requests from email attachments (PDF, JPG, JPEG, TIF, TIFF, and PNG file formats supported). For more information, see scope item 5LT, SAP Intelligent Robotic Process Automation (RPA) Factory, and SAP Intelligent RPA Store. Picture 1: An iRPA bot will be released on the SAP Cloud Platform the bot will automatically monitor the email inbox, once receiving a purchase order email, it will download the attached PDF file and upload it to the Fiori app. NOTE: The data extraction function in this app is provided by the Document Information Extraction service. What’s New in SAP S/4HANA Cloud, public edition, 2111 release: Simplified setup for scope item 4X9, which no longer requires an enterprise account on SAP Business Technology Platform (a separate license) to use this app. Activation of duplicate check on uploaded files in the configuration activity Adapt Settings by Company Code Two new situation templates, allow you to activate automatic notifications about sales order requests that require manual processing References To learn more about Document information Extraction, see What Is Document Information Extraction? To learn more about the iRPA (Intelligent robotic process automation), read the blog from eugene_517 in the SAP Community here For additional information about the automatic sales-order creation from unstructured data, see SAP Help here. Mass Changes of Sales Documents Low-touch sales order processing is a process that requires minimal human intervention. The Mass Changes of Sales Documents innovation continues to raise our journey to this low-touch sales order by automating as many steps as possible in the sales order processing without impacting the flexibility or speed of the process One more innovation concerning the low-touch sales order process is mass changes in sales documents. Typically, updating sales orders in SAP S/4HANA Cloud, public edition, requires the user to manually open and make changes to the correspondent Sales Orders Header Data or Sales Orders Items Data, one after the other. Maybe this is a possible process for a very small list of orders, but it becomes time-consuming for larger ones. Since the SAP S/4HANA Cloud, public edition, 2105 release, we have a new SAP Fiori app to manage mass changes of sales documents. This allows sales representatives to manage different sales documents with the possibility to change them massively without any development effort. We support the Mass Changes for: sales orders, sales orders without charge, sales scheduling agreements, sales contracts, debit- and credit memo requests, and customer returns. https://www.youtube.com/watch?v=ZWYEj2RMG-Q Video 2: The new SAP Fiori app enables you to trigger mass changes of sales documents and to manage and make changes to multiple documents at a time. Triggering of mass changes to sales documents across document categories With the SAP S/4HANA Cloud, public edition, 2202 release you can use the Mass Change of Sales Documents SAP Fiori app to access and make changes across all types of sales documents in one step. For example, you can make changes to multiple sales orders and sales quotations at the same time. Remark: The Mass Change of Sales Documents app now supports the consumption of multiple addresses for a customer for all the sales document categories. https://sapvideoa35699dc5.hana.ondemand.com/?entry_id=1_wjiavolw Demo 3: Ability to trigger mass changes at a header and item level across multiple sales document categories (including audio). Create new sales orders through spreadsheet upload With the SAP S/4HANA Cloud, public edition, 1911 release, we deliver the first version of the new SAP Fiori app “Create Sales Orders”. The SAP Fiori Create Sales Orders application is the new face of order management in SAP S/4HANA, public edition. With this application, the classic VA01 SAP GUI transaction is renovated into an SAP Fiori application for an intuitive user experience. Complexity has been reduced without compromising on transactional business requirements. In addition, the application offers in-app analytics leveraging SAP S/4HANA, public edition, and customization features for end-users. See here Journey of Managing Sales Orders in SAP Fiori: #1 Overview | SAP Blogs With SAP S/4HANA Cloud, public edition, 2005 release, we improve and increase sales force efficiency with a new option to create sales orders by uploading spreadsheets with the required data. You can use this process when creating sales orders as a result of mass upload capabilities. https://www.youtube.com/watch?v=7YAGvLjTOfE&t=1s Video 3: Click here to watch the video and see the improved efficiency in creating sales orders. Business background You regularly receive orders maintained in MS Excel files. As sales orders are fundamental to any inventory and order management system, sales orders should be entered into your system as soon as possible. You have now to create sales orders based on the data maintained in these MS Excel files, and you need to manually enter the order data and create them one by one in the Create Sales Orders app. With the new SAP Fiori app “Import Sales Orders”, you can batch-create sales orders directly from MS Excel spreadsheet files. After an order data file is uploaded, the system automatically creates sales orders, thereby reducing manual workload and enhancing work efficiency for internal sales reps. Note 1: The sales order fields provided in this app do not suffice for creating sales orders with a billing plan or for one-time customers. If you have created these sales orders using the app, you must make order data complete later in the “Change Sales Orders - VA02” app. Note 2: Technically, the import of sales orders requires field names that uniquely identify sales order fields. Currently, only a limited number of order fields are available for import. You can find their field names in the template For more info about this innovation see my blog for SAP S/4HANA Cloud 2005 here. What’s New in SAP S/4HANA Cloud, public edition, 2108 release: With the SAP S/4HANA Cloud, public edition 2108 release, we extend this feature, and you can now import pricing element data, texts, and custom fields using the Import Sales Orders app. Prior to an import, we recommend downloading the standard template directly from SAP S/4HANA Cloud, public edition. This template provides the list of sales order fields that are either required or optional for importing sales orders. https://sapvideoa35699dc5.hana.ondemand.com/?entry_id=1_fl6w8bpt Demo 4: Improved efficiency in creating sales orders, as a result of mass upload capabilities (including audio). Thanks for reading this blog post. Stay tuned! SAP S/4HANA Cloud, public edition release info here The Link Collection for Sales here Latest SAP S/4HANA Cloud, public edition release blog posts here and previous release highlights here Product videos on our SAP S/4HANA Cloud, Public Edition, and SAP S/4HANA YouTube playlist SAP S/4HANA PSCC Digital Enablement Wheel here Early Release Webinar Series here Inside SAP S/4HANA Podcast here openSAP Microlearnings for SAP S/4HANA here Best practices for SAP S/4HANA Cloud, public edition here SAP S/4HANA Cloud, Public Edition Community: here Feature Scope Description here What’s New here Help Portal Product Page here Follow us via @Sap and #S4HANA, or myself via @VoglerChrist and LinkedIn

YangYue01

Retrieval-Augmented Generation (RAG) techniques significantly improve the contextual relevance of large language model (LLM) outputs by incorporating real-time database lookups. SAP HANA Cloud vector engine plays a key role here, enabling efficient retrieval of pertinent data through vector similarity measures, thereby augmenting the performance of RAG tasks. This tutorial provides a comprehensive walkthrough, using Python, to showcase the utilization of SAP HANA cloud vector engine in a RAG-based LLM application.

Artificial Intelligence and Machine Learning

Artificial Intelligence and Machine Learning

Featured Topics

Welcome to the Artificial Intelligence Community Group!

Recent Activity

Hands-on Tutorial: Creating an FAQ Chatbot on BTP

LLAMA3 70B programming in ABAP

New Machine Learning features in SAP HANA Cloud 2024 Q1

AI Challenge | Web Scraping with Generative AI

Transforming Business Operations with an AI Copilot

Unveiling SAP Business AI's Potential at the Upcoming Solution Experience Live Session

AI Core - problem with copying object from AWS to /mnt/models

Bring Open-Source LLMs into SAP AI Core

Sales in SAP S/4HANA Cloud, Public Edition – Automated Order Process with SAP´s Intelligent Technologies

Mastering SAP HANA Cloud Vector Engine for RAG-driven LLM Applications: An Insightful Guide

Artificial Intelligence and Machine Learning at SAP

Mastering SAP HANA Cloud Vector Engine for RAG-dri...

Hands-on Tutorial: Creating an FAQ Chatbot on BTP

LLAMA3 70B programming in ABAP

Re: New Machine Learning features in SAP HANA Clou...

Re: New Machine Learning features in SAP HANA Clou...

Re: AI Challenge | Web Scraping with Generative AI

Re: Transforming Business Operations with an AI Co...