Implementing Advanced RAG Techniques with Elasticsearch Vector DB

July 1, 2024

‍

‍

Imagine a world where applications understand your queries and provide the most relevant information instantly. This article provides context about creating RAG applications using Elasticsearch and advanced techniques like reranking and auto-merging. Discover how these cutting-edge technologies can change the way you interact with data, enabling seamless and highly accurate information retrieval.

Introduction

In this article, we will explore how to develop a Retrieval-Augmented Generation-based chatbot implementing advanced RAG techniques with Elasticsearch as the vector database and Llama 3 as the large language model.

But let’s first try to understand, “What are advanced RAG techniques?”

Advanced RAG techniques are a modification over native RAG techniques, which help in better retrieval of data from documents by using a number of sophisticated techniques.

There are various advanced RAG techniques available, but in this article we will be utilizing two of them:

Auto-Merging: In Auto-Merging technique, the split document is converted into nodes with different levels of chunk sizes. The larger chunk sizes are referred to as parent nodes while the smaller ones are referred to as child nodes. Each node has a unique id, which helps the child node to point to the parent node.

Reranking: Reranking is one of the most basic techniques in advanced RAG. As the name suggests, it reranks chunks of documents based on the similarity score achieved by any reranker model.

‍

Details of how our application will work can be understood through this flowchart:

GitHub

You can access the full code and implementation on GitHub.

Let’s Code

Setting Up the Environment

Let's start by installing the necessary libraries that we’ll be using in the project.


pip install llama-index
pip install elastic-search
pip install sentence-transformers
pip install langchain
pip install groq

Data Preparation

We’ll be taking up a dataset from planetfp.org/, which contains a pdf, “Introduction to the Indian Financial System and Markets”,

detailing the financial system of India.

You can access the dataset here.

Our first step will be to load the documents into the memory to perform operations over it.


from llama_index.readers.file import PyMuPDFReader
loader = PyMuPDFReader()
docs = loader.load(file_path="Chapter-2-Introduction-to-the-Indian-FS-and-Markets.pdf")


from llama_index.core import Document
doc_text = "\n\n".join([d.get_content() for d in docs])
docs = [Document(text=doc_text)]

Auto-Merging

The next step will be to split the dataset into smaller and larger chunks (child nodes and parent nodes) to perform auto-merging over it. For this we will be using the HierarchicalNodeParser.


from llama_index.core.node_parser import HierarchicalNodeParser,get_leaf_nodes


node_parser = HierarchicalNodeParser.from_defaults()


nodes = node_parser.get_nodes_from_documents(docs)
leaf_nodes = get_leaf_nodes(nodes)


nodes_by_id = {node.node_id: node for node in nodes}

Let’s initiate our embedding model, which will be used to convert the sentences into their respective vector embeddings.


from sentence_transformers import SentenceTransformer
model = SentenceTransformer('all-mpnet-base-v2')

Data Visualization

It's time to look over our data so it can be pushed into the Elasticsearch database. For this purpose, we need to initiate a list containing all the ids of child nodes pointing to the parent node, as well as a list containing small chunks of data from the child nodes that were created earlier and stored in child nodes that will be converted to vectors.


parent_ids_list = []
for i in range(0,len(leaf_nodes)):
    parent_ids_list.append(leaf_nodes[i].parent_node.node_id)





child_ids_list = []
for i in range(0,len(leaf_nodes)):
    child_ids_list.append(leaf_nodes[i].node_id)



child_contexts_list = []
for i in range(0,len(leaf_nodes)):
    child_contexts_list.append(leaf_nodes[i].text)

Now let’s create a dataframe using Pandas for better visualization.


import pandas as pd

df = pd.DataFrame({'parent_id':parent_ids_list,'child_id':child_ids_list,'child_context':child_contexts_list})

df["vectors"] = df["child_context"].apply(lambda x: model.encode(x))

Our dataframe looks like this:

‍

Storing the Data

For indexing the data in the database, it needs to be converted into the JSON format as Elasticsearch can only take up JSON format files; so we need to first define the structure of the data in JSON.


indexMapping = {
    "properties":{
        "parent_id":{
            "type":"text"
        },
        "child_id":{
            "type":"text"
        },
        "child_context":{
            "type":"text"
        },
        "vectors":{
            "type":"dense_vector",
            "dims":768,
            "index":True,
            "similarity":"l2_norm"
        }
    }
}

Next, let’s make a connection with the Elasticsearch server. If the connection is established successfully, the code will return true.

Please refer to this article for installation of Elasticsearch database.


from elasticsearch import Elasticsearch


es = Elasticsearch(
    "https://localhost:9200",
    basic_auth=("elastic","YOUR_PASSWORD_HERE"),
    ca_certs="../elasticsearch-8.14.1/config/certs/http_ca.crt"
)
es.ping()

Note: Replace YOUR_PASSWORD_HERE with the password you created at the time of installation.

It's time to initiate our database in the Elasticsearch server by running this code snippet.


es.indices.create(index="finance",mappings= indexMapping)

Now let's convert our Pandas dataframe into the JSON format structure as defined above and index that into the database.


record_list = df.to_dict("records")


for record in record_list:
    try:
        es.index(index="finance",document=record)
    except Exception as e:
        print(e)

Fetching the Results

Let’s first define a function that will take a query from the user and return the parent ids of the most similar vectors present in the Elasticsearch database.


test_query = "What are the Features of Capital Markets ?"


def find_matching_parent_ids(input_query):
    vector_of_query = model.encode(input_query)
    query = {
            "field":"vectors",
            "query_vector":vector_of_query,
            "k":25,
            "num_candidates":421,
        }
    results = es.knn_search(index="finance",
                            knn=query,
                            source=["child_context","parent_id"]
                            )
    fetched_ids = []




    for i in results["hits"]["hits"]:
        fetched_ids.append(i['_source']['parent_id'])
    return fetched_ids




fetched_parent_ids = find_matching_parent_ids(test_query)

Next, let's create a function to find the most frequently occurring parent ids in order to select the most relevant parent node for further analysis.


def most_frequent_parent_ids(list_of_id):
    frequency_dict = {}
    threshold = 5
    for element in list_of_id:
        if element in frequency_dict:
            frequency_dict[element] += 1
        else:
            frequency_dict[element] = 1




    sorted_elements = sorted(frequency_dict.items(), key=lambda item: item[1], reverse=True)
    most_common_ids = []
    for i in range(0,threshold):
        most_common_ids.append(sorted_elements[i][0])
    return most_common_ids


most_common_ids=most_frequent_parent_ids(fetched_parent_ids)

Based on the parent ids received, let’s fetch the data we created earlier as the larger chunks.


parent_context_list = []


for y in most_common_ids:
    for i in range(0,len(leaf_nodes)):
        if(y== leaf_nodes[i].parent_node.node_id):
            parent_context_list.append(nodes_by_id[leaf_nodes[i].parent_node.node_id].text)
            break

From this we have the parent chunks of the most relevant dataset based on the user query. This completes our first Advanced RAG technique auto-merging.

Let's move on to our second advanced RAG technique, that is, reranking.

Reranking

For this, we will again need to initiate an encoder that will rerank our data based on the similarity with the user query.


from sentence_transformers import CrossEncoder
rankmodel = CrossEncoder("jinaai/jina-reranker-v1-tiny-en")

query = test_query
results = rankmodel.rank(query, parent_context_list, return_documents=True, top_k=5)

Let's put back the reranked data into a list for further processing.


reranked_list = []
for i in range(0,5):
    reranked_list.append(results[i]['corpus_id'])

Initiating the LLM

We will need a large language model for formatting and to show the outputs in natural language. Here, we are employing Llama 3 as the LLM with the support of Groq.


from langchain_groq import ChatGroq

llm = ChatGroq(temperature=0, model_name="llama3-70b-8192",groq_api_key="YOUR_API_KEY")

Note: Please replace API_KEY with your own API key, which can be generated from here.

Now let's create a prompt for the language model in order to get a concise and to-the-point answer from the context.

We need to provide the context and the question to the LLM and, based on that context, the LLM will provide the answer to the question.

So let's first prepare the context from the reranked list. We will be considering the top three contexts from the reranked_list in order to get the highest relevance as they will be most related to the question.


context = ""

for i in reranked_list[0:3]:
    context+=parent_context_list[i]+"\n\n"

It's time to define our prompt template using LangChain to make the work easy.


from langchain import PromptTemplate
from langchain import LLMChain


prompt_template = PromptTemplate(
    template="These are few Context: {context} for this question Question: {question} base on this context generate a relevant concise Answer from thi context:",
    input_variables=["context", "question"]
)


llm_chain = LLMChain(llm=llm, prompt=prompt_template)

Let’s create a function that will take context and questions as input parameters and give the context-based answer generated by the language model.


def generate_answer(context, question):
    input_data = {
        "context": context,
        "question": question
    }
    answer = llm_chain(input_data)
    return answer

It's time to test the model before creating a chatbot over it.

So let's call out the function created above genrate_answer to get a response from the LLM.


print("Your Question:  \n"+test_query+"\n\n"+"Bot Reply:  \n"+generate_answer(context, test_query)['text'])

For the test query we provided above, we are getting this as the response:


test_query = "What are the Features of Capital Markets ?"

‍

Since our model is almost ready, let's move on to create a chatbot. For this purpose, we need to call all the functions declared above in a loop so the model doesn’t stop at just one question.


while True:
    test_query = input("\n\n\n\nEnter your query: ")
    if test_query = "":
      break
    fetched_parent_ids = find_matching_parent_ids(test_query)
    most_common_ids=most_frequent_parent_ids(fetched_parent_ids)
    parent_context_list = []
    for y in most_common_ids:
        for i in range(0,len(leaf_nodes)):
            if(y== leaf_nodes[i].parent_node.node_id):
                parent_context_list.append(nodes_by_id[leaf_nodes[i].parent_node.node_id].text)
                break
   
    query = test_query
    results = rankmodel.rank(query, parent_context_list, return_documents=True, top_k=5)


    reranked_list = []
    for i in range(0,5):
        reranked_list.append(results[i]['corpus_id'])


    context = ""


    for i in reranked_list[0:3]:
        context+=parent_context_list[i]+"\n\n"
    print(context)


    print("'Your Question':  \n"+test_query+"\n\n"+"'Bot Reply':  \n"+generate_answer(context, test_query)['text'])

Final Output

When a user puts in their query in the input box, the model initially converts the query into the respective vector embeddings. These are then matched with the child nodes data present in the database to find the best matches. Then the frequently occurring child nodes respective to their parent nodes are selected and their parent nodes’ data is put through reranking, which is further processed with an LLM and the final answer shows up.

Conclusion

In summary, this project illustrates the effectiveness of building RAG applications using Elasticsearch and advanced techniques like reranking and auto-merging. By following a step-by-step guide, we've demonstrated how to harness the capabilities of powerful tools like Elasticsearch to develop efficient retrieval-augmented generation systems.

References

https://docs.llamaindex.ai/en/stable/examples/retrievers/auto_merging_retriever/

https://www.elastic.co/guide/en/elasticsearch/client/python-api/current/getting-started-python.html

https://python.langchain.com/v0.1/docs/integrations/chat/groq/

Sign up for Free Trial

Latest Blogs

October 11, 2024

Step-by-Step Guide to Fine-Tune Flux.1 with AI Toolkit and Generate Images for Ecommerce

October 18, 2022

A Complete Guide To Customer Acquisition For Startups

Any business is enlivened by its customers. Therefore, a strategy to constantly bring in new clients is an ongoing requirement. In this regard, having a proper customer acquisition strategy can be of great importance.

So, if you are just starting your business, or planning to expand it, read on to learn more about this concept.

The problem with customer acquisition

As an organization, when working in a diverse and competitive market like India, you need to have a well-defined customer acquisition strategy to attain success. However, this is where most startups struggle. Now, you may have a great product or service, but if you are not in the right place targeting the right demographic, you are not likely to get the results you want.

To resolve this, typically, companies invest, but if that is not channelized properly, it will be futile.

So, the best way out of this dilemma is to have a clear customer acquisition strategy in place.

How can you create the ideal customer acquisition strategy for your business?

Define what your goals are

You need to define your goals so that you can meet the revenue expectations you have for the current fiscal year. You need to find a value for the metrics –

MRR – Monthly recurring revenue, which tells you all the income that can be generated from all your income channels.
CLV – Customer lifetime value tells you how much a customer is willing to spend on your business during your mutual relationship duration.
CAC – Customer acquisition costs, which tells how much your organization needs to spend to acquire customers constantly.
Churn rate – It tells you the rate at which customers stop doing business.

All these metrics tell you how well you will be able to grow your business and revenue.

Identify your ideal customers

You need to understand who your current customers are and who your target customers are. Once you are aware of your customer base, you can focus your energies in that direction and get the maximum sale of your products or services. You can also understand what your customers require through various analytics and markers and address them to leverage your products/services towards them.

Choose your channels for customer acquisition

How will you acquire customers who will eventually tell at what scale and at what rate you need to expand your business? You could market and sell your products on social media channels like Instagram, Facebook and YouTube, or invest in paid marketing like Google Ads. You need to develop a unique strategy for each of these channels.

Communicate with your customers

If you know exactly what your customers have in mind, then you will be able to develop your customer strategy with a clear perspective in mind. You can do it through surveys or customer opinion forms, email contact forms, blog posts and social media posts. After that, you just need to measure the analytics, clearly understand the insights, and improve your strategy accordingly.

Combining these strategies with your long-term business plan will bring results. However, there will be challenges on the way, where you need to adapt as per the requirements to make the most of it. At the same time, introducing new technologies like AI and ML can also solve such issues easily. To learn more about the use of AI and ML and how they are transforming businesses, keep referring to the blog section of E2E Networks.

‍

Reference Links

https://www.helpscout.com/customer-acquisition/

https://www.cloudways.com/blog/customer-acquisition-strategy-for-startups/

https://blog.hubspot.com/service/customer-acquisition

This is a decorative image for: Constructing 3D objects through Deep Learning

October 18, 2022

Image-based 3D Object Reconstruction State-of-the-Art and trends in the Deep Learning Era

3D reconstruction is one of the most complex issues of deep learning systems. There have been multiple types of research in this field, and almost everything has been tried on it — computer vision, computer graphics and machine learning, but to no avail. However, that has resulted in CNN or convolutional neural networks foraying into this field, which has yielded some success.

The Main Objective of the 3D Object Reconstruction

Developing this deep learning technology aims to infer the shape of 3D objects from 2D images. So, to conduct the experiment, you need the following:

Highly calibrated cameras that take a photograph of the image from various angles.
Large training datasets can predict the geometry of the object whose 3D image reconstruction needs to be done. These datasets can be collected from a database of images, or they can be collected and sampled from a video.

By using the apparatus and datasets, you will be able to proceed with the 3D reconstruction from 2D datasets.

State-of-the-art Technology Used by the Datasets for the Reconstruction of 3D Objects

The technology used for this purpose needs to stick to the following parameters:

Input

Training with the help of one or multiple RGB images, where the segmentation of the 3D ground truth needs to be done. It could be one image, multiple images or even a video stream.

The testing will also be done on the same parameters, which will also help to create a uniform, cluttered background, or both.

Output

The volumetric output will be done in both high and low resolution, and the surface output will be generated through parameterisation, template deformation and point cloud. Moreover, the direct and intermediate outputs will be calculated this way.

Network architecture used

The architecture used in training is 3D-VAE-GAN, which has an encoder and a decoder, with TL-Net and conditional GAN. At the same time, the testing architecture is 3D-VAE, which has an encoder and a decoder.

Training used

The degree of supervision used in 2D vs 3D supervision, weak supervision along with loss functions have to be included in this system. The training procedure is adversarial training with joint 2D and 3D embeddings. Also, the network architecture is extremely important for the speed and processing quality of the output images.

Practical applications and use cases

Volumetric representations and surface representations can do the reconstruction. Powerful computer systems need to be used for reconstruction.

Given below are some of the places where 3D Object Reconstruction Deep Learning Systems are used:

3D reconstruction technology can be used in the Police Department for drawing the faces of criminals whose images have been procured from a crime site where their faces are not completely revealed.
It can be used for re-modelling ruins at ancient architectural sites. The rubble or the debris stubs of structures can be used to recreate the entire building structure and get an idea of how it looked in the past.
They can be used in plastic surgery where the organs, face, limbs or any other portion of the body has been damaged and needs to be rebuilt.
It can be used in airport security, where concealed shapes can be used for guessing whether a person is armed or is carrying explosives or not.
It can also help in completing DNA sequences.

So, if you are planning to implement this technology, then you can rent the required infrastructure from E2E Networks and avoid investing in it. And if you plan to learn more about such topics, then keep a tab on the blog section of the website.

‍

Reference Links

https://tongtianta.site/paper/68922

https://github.com/natowi/3D-Reconstruction-with-Deep-Learning-Methods

This is a decorative image for: Comprehensive Guide to Deep Q-Learning for Data Science Enthusiasts

October 18, 2022

A Comprehensive Guide To Deep Q-Learning For Data Science Enthusiasts

For all data science enthusiasts who would love to dig deep, we have composed a write-up about Q-Learning specifically for you all. Deep Q-Learning and Reinforcement learning (RL) are extremely popular these days. These two data science methodologies use Python libraries like TensorFlow 2 and openAI’s Gym environment.

So, read on to know more.

What is Deep Q-Learning?

Deep Q-Learning utilizes the principles of Q-learning, but instead of using the Q-table, it uses the neural network. The algorithm of deep Q-Learning uses the states as input and the optimal Q-value of every action possible as the output. The agent gathers and stores all the previous experiences in the memory of the trained tuple in the following order:

State> Next state> Action> Reward

The neural network training stability increases using a random batch of previous data by using the experience replay. Experience replay also means the previous experiences stocking, and the target network uses it for training and calculation of the Q-network and the predicted Q-Value. This neural network uses openAI Gym, which is provided by taxi-v3 environments.

Now, any understanding of Deep Q-Learning is incomplete without talking about Reinforcement Learning.

What is Reinforcement Learning?

Reinforcement is a subsection of ML. This part of ML is related to the action in which an environmental agent participates in a reward-based system and uses Reinforcement Learning to maximize the rewards. Reinforcement Learning is a different technique from unsupervised learning or supervised learning because it does not require a supervised input/output pair. The number of corrections is also less, so it is a highly efficient technique.

Now, the understanding of reinforcement learning is incomplete without knowing about Markov Decision Process (MDP). MDP is involved with each state that has been presented in the results of the environment, derived from the state previously there. The information which composes both states is gathered and transferred to the decision process. The task of the chosen agent is to maximize the awards. The MDP optimizes the actions and helps construct the optimal policy.

For developing the MDP, you need to follow the Q-Learning Algorithm, which is an extremely important part of data science and machine learning.

What is Q-Learning Algorithm?

The process of Q-Learning is important for understanding the data from scratch. It involves defining the parameters, choosing the actions from the current state and also choosing the actions from the previous state and then developing a Q-table for maximizing the results or output rewards.

The 4 steps that are involved in Q-Learning:

Initializing parameters – The RL (reinforcement learning) model learns the set of actions that the agent requires in the state, environment and time.
Identifying current state – The model stores the prior records for optimal action definition for maximizing the results. For acting in the present state, the state needs to be identified and perform an action combination for it.
Choosing the optimal action set and gaining the relevant experience – A Q-table is generated from the data with a set of specific states and actions, and the weight of this data is calculated for updating the Q-Table to the following step.
Updating Q-table rewards and next state determination – After the relevant experience is gained and agents start getting environmental records. The reward amplitude helps to present the subsequent step.

In case the Q-table size is huge, then the generation of the model is a time-consuming process. This situation requires Deep Q-learning.

Hopefully, this write-up has provided an outline of Deep Q-Learning and its related concepts. If you wish to learn more about such topics, then keep a tab on the blog section of the E2E Networks website.

Reference Links

https://analyticsindiamag.com/comprehensive-guide-to-deep-q-learning-for-data-science-enthusiasts/

https://medium.com/@jereminuerofficial/a-comprehensive-guide-to-deep-q-learning-8aeed632f52f

October 13, 2022

GAUDI: A Neural Architect for Immersive 3D Scene Generation

The evolution of artificial intelligence in the past decade has been staggering, and now the focus is shifting towards AI and ML systems to understand and generate 3D spaces. As a result, there has been extensive research on manipulating 3D generative models. In this regard, Apple’s AI and ML scientists have developed GAUDI, a method specifically for this job.

An introduction to GAUDI

The GAUDI 3D immersive technique founders named it after the famous architect Antoni Gaudi. This AI model takes the help of a camera pose decoder, which enables it to guess the possible camera angles of a scene. Hence, the decoder then makes it possible to predict the 3D canvas from almost every angle.

What does GAUDI do?

GAUDI can perform multiple functions –

The extensions of these generative models have a tremendous effect on ML and computer vision. Pragmatically, such models are highly useful. They are applied in model-based reinforcement learning and planning world models, SLAM is s, or 3D content creation.
Generative modelling for 3D objects has been used for generating scenes using graf, pigan, and gsn, which incorporate a GAN (Generative Adversarial Network). The generator codes radiance fields exclusively. Using the 3D space in the scene along with the camera pose generates the 3D image from that point. This point has a density scalar and RGB value for that specific point in 3D space. This can be done from a 2D camera view. It does this by imposing 3D datasets on those 2D shots. It isolates various objects and scenes and combines them to render a new scene altogether.
GAUDI also removes GANs pathologies like mode collapse and improved GAN.
GAUDI also uses this to train data on a canonical coordinate system. You can compare it by looking at the trajectory of the scenes.

How is GAUDI applied to the content?

The steps of application for GAUDI have been given below:

Each trajectory is created, which consists of a sequence of posed images (These images are from a 3D scene) encoded into a latent representation. This representation which has a radiance field or what we refer to as the 3D scene and the camera path is created in a disentangled way. The results are interpreted as free parameters. The problem is optimized by and formulation of a reconstruction objective.
This simple training process is then scaled to trajectories, thousands of them creating a large number of views. The model samples the radiance fields totally from the previous distribution that the model has learned.
The scenes are thus synthesized by interpolation within the hidden space.
The scaling of 3D scenes generates many scenes that contain thousands of images. During training, there is no issue related to canonical orientation or mode collapse.
A novel de-noising optimization technique is used to find hidden representations that collaborate in modelling the camera poses and the radiance field to create multiple datasets with state-of-the-art performance in generating 3D scenes by building a setup that uses images and text.

To conclude, GAUDI has more capabilities and can also be used for sampling various images and video datasets. Furthermore, this will make a foray into AR (augmented reality) and VR (virtual reality). With GAUDI in hand, the sky is only the limit in the field of media creation. So, if you enjoy reading about the latest development in the field of AI and ML, then keep a tab on the blog section of the E2E Networks website.

‍

Reference Links

https://www.researchgate.net/publication/362323995_GAUDI_A_Neural_Architect_for_Immersive_3D_Scene_Generation

https://www.technology.org/2022/07/31/gaudi-a-neural-architect-for-immersive-3d-scene-generation/

https://www.patentlyapple.com/2022/08/apple-has-unveiled-gaudi-a-neural-architect-for-immersive-3d-scene-generation.html

Build on the most powerful infrastructure cloud

Contact Sales Get Started for Free

A vector illustration of a tech city using latest cloud technologies & infrastructure

Implementing Advanced RAG Techniques with Elasticsearch Vector DB

Table of Contents

Introduction

GitHub

Let’s Code

Setting Up the Environment

Data Preparation

Auto-Merging

Data Visualization

Storing the Data

Fetching the Results

Reranking

Initiating the LLM

Final Output

Conclusion

References

A Comparative Analysis of H200 vs. H100 vs. A100 vs. L40S vs. L4 GPUs

A Guide to Building Llama 3.1 RAG Applications with TIR AI Studio

Step-by-Step Guide to Deploying and Using Llama 3 LLMs on TIR AI Studio

Urgent Security Advisory: CUPS Vulnerabilities

Guide to Image Generation Using FLUX.1-dev-gguf with ComfyUI

Step-by-Step Guide to Fine-Tune Flux.1 with AI Toolkit and Generate Images for Ecommerce

A Complete Guide To Customer Acquisition For Startups

Image-based 3D Object Reconstruction State-of-the-Art and trends in the Deep Learning Era

The Main Objective of the 3D Object Reconstruction

A Comprehensive Guide To Deep Q-Learning For Data Science Enthusiasts

GAUDI: A Neural Architect for Immersive 3D Scene Generation

Build on the most powerful infrastructure cloud