1. IRIS RAG Demo

IRIS RAG Demo

This demo showcases the powerful synergy between IRIS Vector Search and RAG (Retrieval Augmented Generation), providing a cutting-edge approach to interacting with documents through a conversational interface. Utilizing InterSystems IRIS's newly introduced Vector Search capabilities, this application sets a new standard for retrieving and generating information based on a knowledge base.
The backend, crafted in Python and leveraging the prowess of IRIS and IoP, the LLM model is orca-mini and served by the ollama server.
The frontend is an chatbot written with Streamlit.

17 3
2 1K

Keywords: Jupyter Notebook, Tensorflow GPU, Keras, Deep Learning, MLP, and HealthShare

1. Purpose and Objectives

In previous"Part I" we have set up a deep learning demo environment. In this "Part II" we will test what we could do with it.

Many people at my age had started with the classic MLP (Multi-Layer Perceptron) model. It is intuitive hence conceptually easier to start with.

1 2
3 996

Keywords: IRIS, IntegratedML, Machine Learning, Covid-19, Kaggle

Purpose

Recently I noticed a Kaggle dataset for the prediction of whether a Covid-19 patient will be admitted to ICU. It is a spreadsheet of 1925 encounter records of 231 columns of vital signs and observations, with the last column of "ICU" being 1 for Yes or 0 for No. The task is to predict whether a patient will be admitted to ICU based on known data.

2 1
1 919

Artificial intelligence has solved countless human challenges – and medical coding might be next.
As organizations prepare for ICD-11, medical coding is about to become more complicated. Healthcare organizations in the United States already manage 140,000+ codes in ICD-10. With ICD-11, that number will rise.
Some propose artificial intelligence as a solution. AI could aid computer-based medical coding systems, identifying errors, enhancing patient care, and optimizing revenue cycles, among other benefits.

1 0
0 845

On this GitHub you can find all the information on how to use a HuggingFace machine learning / AI model on the IRIS Framework using python.

1. iris-huggingface

Usage of Machine Learning models in IRIS using Python; For text-to-text, text-to-image or image-to-image models.

6 5
1 758

The invention and popularization of Large Language Models (such as OpenAI's GPT-4) has launched a wave of innovative solutions that can leverage large volumes of unstructured data that was impractical or even impossible to process manually until recently.

27 4
5 757

Challenges of real-time AI/ML computations

We will start from the examples that we faced as Data Science practice at InterSystems:

  • A “high-load” customer portal is integrated with an online recommendation system. The plan is to reconfigure promo campaigns at the level of the entire retail network (we will assume that instead of a “flat” promo campaign master there will be used a “segment-tactic” matrix). What will happen to the recommender mechanisms? What will happen to data feeds and updates into the recommender mechanisms (the volume of input data having increased 25000 times)? What will happen to recommendation rule generation setup (the need to reduce 1000 times the recommendation rule filtering threshold due to a thousandfold increase of the volume and “assortment” of the rules generated)?
  • An equipment health monitoring system uses “manual” data sample feeds. Now it is connected to a SCADA system that transmits thousands of process parameter readings each second. What will happen to the monitoring system (will it be able to handle equipment health monitoring on a second-by-second basis)? What will happen once the input data receives a new bloc of several hundreds of columns with data sensor readings recently implemented in the SCADA system (will it be necessary, and for how long, to shut down the monitoring system to integrate the new sensor data in the analysis)?
  • A complex of AI/ML mechanisms (recommendation, monitoring, forecasting) depend on each other’s results. How many man-hours will it take every month to adapt those AI/ML mechanisms’ functioning to changes in the input data? What is the overall “delay” in supporting business decision making by the AI/ML mechanisms (the refresh frequency of supporting information against the feed frequency of new input data)?

4 0
1 726

Hi all. We are going to find duplicates in a dataset using Apache Spark Machine Learning algorithms.

Note: I have done the following on Ubuntu 18.04, Python 3.6.5, Zeppelin 0.8.0, Spark 2.1.1

Introduction

In previous articles we have done the following:

0 0
1 725

We all know that having a set of proper test data before deploying an application to production is crucial for ensuring its reliability and performance. It allows to simulate real-world scenarios and identify potential issues or bugs before they impact end-users. Moreover, testing with representative data sets allows to optimize performance, identify bottlenecks, and fine-tune algorithms or processes as needed. Ultimately, having a comprehensive set of test data helps to deliver a higher quality product, reducing the likelihood of post-production issues and enhancing the overall user experience.

In this article, let's look at how one can use generative AI, namely Gemini by Google, to generate (hopefully) meaningful data for the properties of multiple objects. To do this, I will use the RESTful service to generate data in a JSON format and then use the received data to create objects.

29 4
0 695

What is Distributed Artificial Intelligence (DAI)?

Attempts to find a “bullet-proof” definition have not produced result: it seems like the term is slightly “ahead of time”. Still, we can analyze semantically the term itself – deriving that distributed artificial intelligence is the same AI (see our effort to suggest an “applied” definition) though partitioned across several computers that are not clustered together (neither data-wise, nor via applications, not by providing access to particular computers in principle). I.e., ideally, distributed artificial intelligence should be arranged in such a way that none of the computers participating in that “distribution” have direct access to data nor applications of another computer: the only alternative becomes transmission of data samples and executable scripts via “transparent” messaging. Any deviations from that ideal should lead to an advent of “partially distributed artificial intelligence” – an example being distributed data with a central application server. Or its inverse. One way or the other, we obtain as a result a set of “federated” models (i.e., either models trained each on their own data sources, or each trained by their own algorithms, or “both at once”).

2 0
1 667

Problem

In a fast-paced clinical environment, where quick decision-making is crucial, the lack of streamlined document storage and access systems poses several obstacles. While storage solutions for documents exist (e.g, FHIR), accessing and effectively searching for specific patient data within those documents meaningfully can be a significant challenge.

7 0
2 625

​Keywords: ChatGPT, COS, Lookup Table, IRIS, AI

Purpose


Here is another quick note before we move on to GPT-4 assisted automation journey. Below are some "little" helps ChatGPT had already been offering, here and there, during daily works.

And what could be the perceived gaps, risks and traps to LLMs assisted automation, if you happen to explore this path too. I'd also love to hear anyone's use cases and experiences on this front too.

6 3
2 622

With the advent of Embedded Python, a myriad of use cases are now possible from within IRIS directly using Python libraries for more complex operations. One such operation is the use of natural language processing tools such as textual similarity comparison.

14 4
4 582

Hi Community!

Just want to share with you an exercise I made to create "my own" chat with GPT in Telegram.

It became possible because of two components on Open Exchange: Telegram Adapter by @Nikolay Solovyev and IRIS Open-AI by @Kurro Lopez

So with this example you can setup your own chat with ChatGPT in Telegram.

Let's see how to make it work!

6 8
1 574

Hello everyone, this is with great pleasure that I announce the V2 of my application 'Contest-FHIR'.

In this new version, I used new tools and techniques I discovered at the EUROPEAN HEALTHCARE HACKATHON in which I was invited by InterSystems as a guest and as a mentor to display the multiple projects I did in my intership back in April 2022.

Today I present to you the V2 of my application, it can now transform CSV to FHIR to SQL to JUPYTER notebook.

4 0
0 546

This is my introduction to a series of posts explaining how to create an end-to-end Machine Learning system.

Starting with one problem

Our IRIS Development Community has several posts without tags or wrong tagged. As the posts keep growing the organization
of each tag and the experience of any community member browsing the subjects tends to decrease.

First solutions in mind

We can think some usual solutions for this scenario, like:

5 11
2 531

Artificial Intelligence (AI) is getting a lot of attention lately because it can change many areas of our lives. Better computer power and more data have helped AI do amazing things, like improving medical tests and making self-driving cars. AI can also help businesses make better decisions and work more efficiently, which is why it's becoming more popular and widely used. How can one integrate the OpenAI API calls into an existing IRIS Interoperability application?

13 6
4 518

A few months ago, I read this interesting article from MIT Technology Review, explaing how COVID-19 pandemic are issuing challenges to IT teams worldwide regarding their machine learning (ML) systems.

Such article inspire me to think about how to deal with performance issues after a ML model was deployed.

2 2
0 480

Problem

Do you resonate with this - A capability and impact of a technology being truly discovered when it's packaged in a right way to it's audience. Finest example would be, how the Generative AI took off when ChatGPT was put in the public for easy access and not when Transformers/RAG's capabilities were identified. At least a much higher usage came in, when the audience were empowered to explore the possibilities.

8 6
6 480

Artificial intelligence is not limited only to generating images through text with instructions or creating narratives with simple directions.

You can also make variations of a picture or include a special background to an already existing one.

Additionally, you can obtain the transcription of audio regardless of its language and the speed of the speaker.

So, let's analyze how the file management works.

9 2
1 476

As you have seen in the latest community publications, InterSystems IRIS has included since version 2024.1 the possibility of including vector data types in its database and based on this type of data vector searches have been implemented. Well, these new features reminded me of the article I published a while ago that was based on facial recognition using Embedded Python.

7 2
2 470

In today's data landscape, businesses encounter a number of different challenges. One of them is to do analytics on top of unified and harmonized data layer available to all the consumers. A layer that can deliver the same answers to the same questions irrelative to the dialect or tool being used.

11 2
1 456
Article
· May 14, 2024 11m read
Q&A Chatbot with IRIS and langchain

TL;DR

This article introduces using the langchain framework supported by IRIS for implementing a Q&A chatbot, focusing on Retrieval Augmented Generation (RAG). It explores how IRIS Vector Search within langchain-iris facilitates storage, retrieval, and semantic search of data, enabling precise and up-to-date responses to user queries. Through seamless integration and processes like indexing and retrieval/generation, RAG applications powered by IRIS enable the capabilities of GenAI systems for InterSystems developers.

5 3
3 422

What is Unstructured Data?
Unstructured data refers to information lacking a predefined data model or organization. In contrast to structured data found in databases with clear structures (e.g., tables and fields), unstructured data lacks a fixed schema. This type of data includes text, images, videos, audio files, social media posts, emails, and more.

8 3
0 417

In this article, I am trying to identify the multiple areas to develop the features we can able to do using python and machine learning.

Each hospital is every moment trying to improve its quality of service and efficiency using technology and services.

The healthcare sector is one of the very big and vast areas of service options available and python is one of the best technology for doing machine learning.

In every hospital, humans will come with some feelings, if this feeling will understand using technology is make a chance to provide better service.

2 2
2 416