Swinburne University of Technology

Amir Aryani

Integrating Large Language Models (LLMs) such as GPT into organizations’ data workflows is a complex process with various challenges. These obstacles include but are not limited to technical, operational, ethical, and legal dimensions, each presenting hurdles that organisations must navigate to harness the full potential of LLMs effectively.

What are the challenges in integrating LLMs into organisations' data workflows?

I asked Generative AI Models about their context window. Their response was intriguing. The context window for a large language model (LLM) like OpenAI’s GPT refers to the maximum amount of text the model can consider at any one time when generating a response. This includes both the prompt provided by the user and the model’s generated text.

What is your Context Window?

Authors: Hui Yin, Amir Aryani  As we discussed in our previous article “A Brief Introduction to Retrieval Augmented Generation (RAG)”, RAG is an artificial intelligence framework that incorporates the latest reliable external knowledge and aims to improve the quality of responses generated by pre-trained language models (PLM). Initially, it was designed to improve the performance of knowledge-intensive NLP tasks (Lewis et al., 2020). As

Navigating the Present: Exploring Practical Horizons of Retrieval-Augmented Generation (RAG)

At the time of writing this post (November 19, 2023), there are 401,014 models and transformers listed on Huggingface.co. This number is a testament to the success and impact of the open-source model in the development of AI technologies. However, the challenge at hand is understanding the available models and keeping up to date with the latest developments.

Estimated time: 12 Years to Review all 400k Models on Hugging Face

Prompt engineering is a technique used to effectively communicate with large language models (LLM) like GPT-3 or GPT-4 and get the desired output. Here there are applications of 8 distinct prompt techniques to interact with Mistral AI and find out how we can prompt effectively to learn about Foundation Models.
<strong>
 Zero-Shot Learning
</strong>
: This involves giving the AI a task without any prior examples.

8 Types of Prompt Engineering

There is a plethora of articles and predictions about the future development of AI. Reading through them suggests three main areas of high-impact changes that would transform our daily lives: Healthcare, Automation, and Manufacturing. The impact of AI in these industries is profound and multifaceted, with each sector experiencing unique transformations. Some of these ideas are described in this NVIDIA article by Cliff Edwards.

Revolutionizing Wellness and Wheels: AI’s Transformative Role in Health and Automotive Industries

Procrastination, a common adversary in our daily lives, often leads to a backlog of tasks and ambitions. It’s a silent thief, stealthily stealing opportunities and time. You might have a stack of books on your bookshelf, longing to be read. Perhaps there’s a new language you’ve wanted to learn, or piles of paperwork cluttering your office that you’ve been meaning to organize.

Maximizing Productivity in 10-Minute Bursts

Authors: Hui Yin , Amir Aryani  With the increasing application of large language models in various scenarios, people realize that these models are not omnipotent. When generating dialogues (Shuster et al., 2021), the models often produce hallucinations, leading to inaccurate answers.

A Brief Introduction to Retrieval Augmented Generation (RAG)

Stories by Amir Aryani on Medium

Authors: [Hui Yin](https://www.linkedin.com/in/hui-yin-784538213/),
[Amir Aryani](https://www.linkedin.com/in/amiraryani/)

> RAG is an AI framework enhancing pre-trained language models with
> up-to-date, reliable external knowledge.

As we discussed in our previous article "[A Brief Introduction to
Retrieval Augmented Generation
(RAG)](https://medium.com/@amiraryani/a-brief-introduction-to-retrieval-augmented-generation-rag-4bd6e50da532)",
RAG is an artificial intelligence framework that incorporates the latest
reliable external knowledge and aims to improve the quality of responses
generated by pre-trained language models (PLM). Initially, it was
designed to improve the performance of knowledge-intensive NLP tasks
([Lewis et al., 2020](https://arxiv.org/abs/2005.11401)). As more and
more people notice the advantages of RAG, it has also been applied to a
broader range of scenarios. We conducted a rapid review of RAG-related
publications and found there are mainly five categories of use cases:
Medical, Knowledge Graph, Code Comment, Chatbot, Review, and Image
processing.

<figure>
<img
src="https://cdn-images-1.medium.com/max/1024/1*HGpI_s3nvDIyHRE2_yE5kQ.png" />
<figcaption>Created by Dr Hui Yin, Jan 2024. CC-BY License
for reuse.</figcaption>
</figure>

### **Chatbot**

Chatbots have gained widespread recognition and use by the public. It
acts as an intermediary, translating user queries or requests written in
natural language and communicating with the underlying LLM to generate
responses.

The researchers incorporated RAG into the following tasks to improve the
performance of the chatbot, allowing it to provide more accurate
responses.

Question Answering ([Mao et al.,
2020](https://arxiv.org/abs/2009.08553), [Ju et al.,
2022](https://doi.org/10.48550/arXiv.2210.02933), [Shamane et al.,
2023](https://doi.org/10.1162/tacl_a_00530)) requires access to external
knowledge bases, documents, or structured data to provide accurate and
informative responses.

Text summarization ([Cai et al.,
2022](https://doi.org/10.1145/3477495.3532682), [Hofstätter et al.,
2023](https://doi.org/10.1145/3539618.3591687)), which requires an
understanding of the content and context with the assistance of external
knowledge to generate a concise and informative summary of longer text.

In certain scenarios, dialogue systems can also benefit from RAG
technology, allowing them to handle specific tasks more efficiently
([Shuster et al., 2021](https://doi.org/10.48550/arXiv.2104.07567),
[Thule et al., 2021](https://doi.org/10.48550/arXiv.2102.04643)).

### **Code Comment and Summarization**

Code comments are crucial as they provide explanatory information about
the purpose, functionality, and usage of the code, clear and informative
comments help improve code understanding.

[Yu et al., 2022](https://doi.org/10.48550/arXiv.2206.13325) proposed
BashExplainer which aims to automatically generate code comments for
Bash scripts to enhance code comprehension, especially for developers
who may struggle with understanding Bash code due to its unique
characteristics.

[Liu et al., 2021](https://doi.org/10.48550/arXiv.2006.05405) introduced
a novel retrieval-augmented mechanism for code summarization, aiming to
bridge the language gap between source code and natural language
summaries.

### **Image Processing**

Image processing covers three topics, text-to-image, image-to-text and
image synthesis.

Text-to-image is the generation of visual content (images) from textual
descriptions.

[Chen et al., 2022](https://doi.org/10.48550/arXiv.2209.14491) designed
Re-Image to address challenges in generating images for uncommon
entities. By leveraging a multi-modal knowledge base and a retrieval
step, the model significantly improves fidelity, particularly for rare
entities.

Image-to-text is the reverse of text-to-image, it converts visual
information(images) into textual descriptions, such as generating a
caption or description.

[Sarto et al., 2022](https://doi.org/10.48550/arXiv.2207.13162) proposed
a Retrieval-Augmented Transformer for Image Captioning, which combines a
knowledge retriever, differentiable encoder, and kNN-augmented attention
layer, and enhances caption quality by using explicit external memory.

Image synthesis is very different from the above two items. It creates
new images or modifies existing images, such as generating realistic
images of non-existent objects or scenes, often using generative models
such as generative adversarial networks (GAN). You can notice that some
medium posts contain very stunning images at the top of the post page
and are labelled "Generated by AI". So far, we know of only one work
that uses RAG for image captioning. Blattmann et al. proposed a
Retrieval-Augmented Diffusion Model as a semi-parametric approach for
generative image synthesis.

### **Review**

RAG enhances review generation by incorporating external knowledge, such
as product specifications and customer feedback. The retrieval mechanism
ensures that generated reviews are more informative, contextually
relevant, and aligned with the product or service characteristics,
resulting in nuanced and helpful content compared to traditional
generative models. Several works were found in this research topic, such
as [Kim et al., 2020](https://aclanthology.org/2020.coling-main.207).

### **Medical**

RAG technology has been widely used in Medical area, such as the
Summarisation of electronic health records (EHRs) ([Saba et al.,
2024](https://doi.org/10.48550/arXiv.2401.01469), [Thompson et al.,
2023](https://doi.org/10.48550/arXiv.2312.06457)), Evidence-based
medicine (EBM) ([Vaid et al.,
2024](https://doi.org/10.48550/arXiv.2401.02851)). In fact, it is
obvious that the data processed by the models in these works is text
data related to medical content, which is still an NLP task in nature,
and the external data used in this example specifically belongs to the
medical field.

### **Knowledge Graph**

Integration of Knowledge Graph in RAG is mostly done to improve the
accuracy and performance of the RAG Pipelines. This is mainly a
methodology rather than a use case. However, given the substantial
community support, it has transformed to be an important characteristic
of RAG pipelines.

The knowledge graph is a structured representation of knowledge that
captures entities, their attributes, and their relationships in a
graph-like format. It is structured data that is easy to retrieve. In
Retrieval-Augmented Generation (RAG), leveraging information from
knowledge graphs like Wikipedia can assist in generating more
contextually relevant and accurate natural language text, especially for
open-domain tasks. In addition, the RAG structure uses other external
resources depending on the specific task, such as web scraping,
retrieval of databases, existing text corpora, and external APIS. Below
are some survey papers on RAG with in-depth discussions on external
sources and knowledge graphs, [Wei et al.,
2021](https://doi.org/10.48550/arXiv.2110.08455), [Agrawal et al.,
2023](https://doi.org/10.48550/arXiv.2311.07914), [Pan et
al., 2023](https://doi.org/10.48550/arXiv.2306.08302).

### **Acknowledgement**

This research was supported by the Australian Government through the
Australian Research Council's Industrial Transformation Training Centre
for Information Resilience (CIRES) project number IC200100022.

### References

Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N.,
Küttler, H., Lewis, M., Yih, W.-t., Rocktäschel, T., Riedel, S., &
Kiela, D. (2020b). Retrieval-Augmented Generation for
Knowledge-Intensive NLP Tasks. Neural Information Processing Systems 33
(NeurIPS 2020). <https://arxiv.org/abs/2005.11401>

Mao, Y., He, P., Liu, X., Shen, Y., Gao, J., Han, J., & Chen, W. (2020).
Generation-Augmented Retrieval for Open-domain Question Answering.
*CoRR*, abs/2009.08553. <https://arxiv.org/abs/2009.08553>

Ju, M., Yu, W., Zhao, T., Zhang, C., & Ye, Y. (2022). Grape: Knowledge
Graph Enhanced Passage Reader for Open-domain Question
Answering.<https://doi.org/10.48550/arXiv.2210.02933>

Shamane Siriwardhana, Rivindu Weerasekera, Elliott Wen, Tharindu
Kaluarachchi, Rajib Rana, Suranga Nanayakkara; Improving the Domain
Adaptation of Retrieval Augmented Generation (RAG) Models for Open
Domain Question Answering. *Transactions of the Association for
Computational Linguistics* 2023.11
1--17.<https://doi.org/10.1162/tacl_a_00530>

Cai, D., Wang, Y., Liu, L., & Shi, S. (2022). Recent advances in
retrieval-augmented text generation. In \*Proceedings of the 45th
International ACM SIGIR Conference on Research and Development in
Information Retrieval\* (pp. 3417--3419)
[https://doi.org/10.1145/3477495.3532682](https://doi.org/10.1145/3477495.3532682%5D(https://doi.org/10.1145/3477495.3532682))

Hofstätter, S., Chen, J., Raman, K., & Zamani, H. (2023). Fid-light:
Efficient and effective retrieval-augmented text generation. In
\*Proceedings of the 46th International ACM SIGIR Conference on Research
and Development in Information Retrieval\* (pp. 1437--1447)
[https://doi.org/10.1145/3539618.3591687](https://doi.org/10.1145/3539618.3591687%5D(https://doi.org/10.1145/3539618.3591687))

Shuster, K., Poff, S., Chen, M., Kiela, D., & Weston, J. (2021).
Retrieval augmentation reduces hallucination in conversation. \*arXiv
preprint arXiv:2104.07567\*. <https://doi.org/10.48550/arXiv.2104.07567>

Thulke, D., Daheim, N., Dugast, C., & Ney, H. (2021). Efficient
Retrieval Augmented Generation from Unstructured Knowledge for
Task-Oriented Dialog. \*arXiv preprint\* arXiv:2102.04643.
[https://doi.org/10.48550/arXiv.2102.04643](https://doi.org/10.48550/arXiv.2102.04643%5D(https://doi.org/10.48550/arXiv.2102.04643))

Yu, C., Yang, G., Chen, X., Liu, K., & Zhou, Y. (2022). BashExplainer:
Retrieval-Augmented Bash Code Comment Generation based on Fine-tuned
CodeBERT. <https://doi.org/10.48550/arXiv.2206.13325>

Liu, S., Chen, Y., Xie, X., Siow, J., & Liu, Y. (2021).
Retrieval-Augmented Generation for Code Summarization via Hybrid GNN.
<https://doi.org/10.48550/arXiv.2006.05405>

Chen, W., Hu, H., Saharia, C., & Cohen, W. W. (2022). Re-Imagen:
Retrieval-Augmented Text-to-Image
Generator.<https://doi.org/10.48550/arXiv.2209.14491>

Sarto, S., Cornia, M., Baraldi, L., & Cucchiara, R. (2022).
Retrieval-Augmented Transformer for Image
Captioning.<https://doi.org/10.48550/arXiv.2207.13162>

Blattmann, A., Rombach, R., Oktay, K., Müller, J., & Ommer, B. (2022).
Retrieval-augmented diffusion models. \*Advances in Neural Information
Processing Systems, 35\*, 15309--15324.

Kim, J., Choi, S., Amplayo, R. K., & Hwang, S. (2020).
Retrieval-augmented controllable review generation. In Proceedings of
the 28th International Conference on Computational Linguistics (pp.
2284--2295).<https://aclanthology.org/2020.coling-main.207>

Saba, W., Wendelken, S., & Shanahan, J. (2024). Question-Answering Based
Summarization of Electronic Health Records using Retrieval Augmented
Generation.<https://doi.org/10.48550/arXiv.2401.01469>

Vaid, A., Lampert, J., Lee, J., Sawant, A., Apakama, D., Sakhuja,
A., ... Nadkarni, G. (2024). Generative Large Language Models are
autonomous practitioners of evidence-based medicine.
<https://doi.org/10.48550/arXiv.2401.02851>

Thompson, W. E., Vidmar, D. M., De Freitas, J. K., Pfeifer, J. M.,
Fornwalt, B. K., Chen, R., ... Miotto, R. (2023). Large Language Models
with Retrieval-Augmented Generation for Zero-Shot Disease
Phenotyping.<https://doi.org/10.48550/arXiv.2312.06457>

Wei, X., Wang, S., Zhang, D., Bhatia, P., & Arnold, A. (2021). Knowledge
Enhanced Pretrained Language Models: A Comprehensive
Survey.<https://doi.org/10.48550/arXiv.2110.08455>

Agrawal, G., Kumarage, T., Alghami, Z., & Liu, H. (2023). Can Knowledge
Graphs Reduce Hallucinations in LLMs?: A Survey.
<https://doi.org/10.48550/arXiv.2311.07914>

Pan, S., Luo, L., Wang, Y., Chen, C., Wang, J., & Wu, X. (2023).
Unifying Large Language Models and Knowledge Graphs: A
Roadmap.<https://doi.org/10.48550/arXiv.2306.08302>

![](https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=111ad032653d){width="1"
height="1"}


<figure>
<img
src="https://cdn-images-1.medium.com/max/1024/1*MYULJWGydnlht7zzLUtThg.png" />
<figcaption>Created with DALLE On 18 Jan 2024</figcaption>
</figure>

Authors: [Hui Yin](https://www.linkedin.com/in/hui-yin-784538213/) ,
[Amir Aryani](https://www.linkedin.com/in/amiraryani/)

With the increasing application of large language models in various
scenarios, people realize that these models are not omnipotent. When
generating dialogues (Shuster et al., 2021), the models often produce
hallucinations, leading to inaccurate answers. Despite these large
language models storing impressive knowledge within their parameters of
a neural network, these hallucinated facts are frozen at the point of
model training, making it infeasible to retrain the model. If we can
bootstrap the model correctly, it can still be a powerful tool, and here
comes the Retrieval Augmented Generator (RAG).

Retrieval-augmented generation (RAG) is an AI framework designed to
enhance the quality of responses generated by large language models
(LLM). It leverages external knowledge sources, typically knowledge
graphs or other knowledge bases, such as Wikipedia, to augment the
generation capabilities of these models. This integration enables the
model to incorporate additional external information during text
generation, thereby improving its accuracy and applicability while
ensuring access to the latest reliable information.

<figure>
<img
src="https://cdn-images-1.medium.com/max/1024/1*vvYZftOCIMiBWZxZGuA89A.png" />
<figcaption>RAG Architecture Model</figcaption>
</figure>

The retrieval-augmented generator (RAG) was initially introduced in
"REALM: Retrieval-Augmented Language Model Pretraining" by Guu et al.,
(2020), where they discuss the utilization of dense retrievers to
enhance language model pre-training. In REALM, the retriever accesses
knowledge from a textual knowledge corpus, (e.g., all of Wikipedia), and
then the signal from the generator backpropagates through the retriever.
They verified the effectiveness of the proposed Retrieval-Augmented
Language Model pre-training (REALM) on the challenging Open-domain
Question Answering (Open-QA) task. The results demonstrated its
superiority over three benchmarks by a substantial margin (4--16%
absolute accuracy), and it also showcased improvements in
interpretability and modularity.

<figure>
<img
src="https://cdn-images-1.medium.com/max/1024/1*7GvNzgFWvViZTouSoR6S9w.png" />
<figcaption>REALM Structure</figcaption>
</figure>

Structure of REALM: BERT is employed to encode queries and documents
separately, and [Maximum Inner Product Search
(MIPS)](https://en.wikipedia.org/wiki/Maximum_inner-product_search) is
utilized to calculate the inner product between query embeddings and
document embeddings, identifying the most relevant documents.
Subsequently, the concatenated query and selected documents are input
into the knowledge-augmented encoder (BERT pre-training) as new input,
and the final result is generated.

Another remarkable work is "Retrieval-Augmented Generation for
Knowledge-Intensive NLP Tasks" by Lewis et al., (2020). They constructed
a pre-trained retriever consisting of a Query encoder and a Document
index. For example, [Maximum Inner Product Search
(MIPS)](https://en.wikipedia.org/wiki/Maximum_inner-product_search) was
employed to identify the top-K documents. The generator component of the
system was a pre-trained
[seq2seq](https://en.wikipedia.org/wiki/Seq2seq) model. The entire model
was fine-tuned end-to-end. In their approach for final predictions, they
treated k (the number of documents selected by the retriever) as a
latent variable. This allowed them to control the number of outputs
predicted by the [seq2seq](https://en.wikipedia.org/wiki/Seq2seq) model.

<figure>
<img
src="https://cdn-images-1.medium.com/max/1024/1*5_w3_Gctd8hClmYkrqNpwg.png" />
<figcaption>Maximum Inner Product Search (MIPS)</figcaption>
</figure>

Overview of the structure: The retriever combines QueryEncoder and
Document Index, and the generator consists of a pre-trained
[seq2seq](https://en.wikipedia.org/wiki/Seq2seq) model fine-tuned
end-to-end. Maximum Inner Product Search (MIPS) is utilized to find the
top-k similar documents. For the final predictions, 'k' acts as a latent
variable, and marginalises over
[seq2seq](https://en.wikipedia.org/wiki/Seq2seq) predictions given
different documents.

The above two works have inspired more researchers to study RAG. Based
on the basic structure shown in the first figure, specific
implementation methods are shown in other works. These existing works
all demonstrate RAG's contribution to LLM in the following aspects:

- Utilisation of the latest external knowledge,
- Flexibility and adaptability,
- Quality improvement,
- Interpretability,
- Modularity, and
- Model size reduction.

From an application perspective, the original RAG was oriented to NLP
tasks such as Open-QA (Mao et al., 2020) and dialogue (Thulke et al.,
2021) and is now also used for image synthesis (Blattmann et al., 2022)
and text-to-image (Chen et al., 2022).

RAG possesses many advantages; however, it's crucial to acknowledge that
leveraging an external document database to enhance the generation
process involves computing the similarity between the query embedding
and external document embeddings, followed by selecting the top-k
documents. This introduces substantial computational challenges. As the
size of the external document database increases, the computational cost
of calculating similarity and determining the top-k documents grows
significantly. Achieving scalability, especially when processing
numerous documents, demands efficient algorithms and parallel processing
capabilities to make the retrieval process feasible within a reasonable
timeframe. Additionally, storing and accessing large external databases
can necessitate significant memory resources. For applications requiring
real-time responses, the time taken for retrieval becomes a critical
factor, emphasizing the need to balance computational complexity with
response time. Hence, delving into optimization and approximation
methods or incorporating specialized hardware emerges as promising
avenues for future research.

As an auxiliary to LLM, it is foreseeable that more people will devote
their energy to exploring various applications of RAG.

**Acknowledgement**

This work was supported by the Australian Government through the
Australian Research Council's Industrial Transformation Training Centre
for Information Resilience (CIRES) project number IC200100022.

Shuster, K., Poff, S., Chen, M., Kiela, D., & Weston, J. (2021).
Retrieval augmentation reduces hallucination in conversation.
<https://arxiv.org/abs/2104.07567>

Guu, K., Lee, K., Tung, Z., Pasupat, P., & Chang, M. (2020). Retrieval
Augmented Language Model Pre-Training. In *Proceedings of the 37th
International Conference on Machine Learning, PMLR* (Vol. 119, pp.
3929--3938). [](https://arxiv.org/abs/2005.11401) Available from
<https://proceedings.mlr.press/v119/guu20a.html>

Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N.,
Küttler, H., Lewis, M., Yih, W.-t., Rocktäschel, T., Riedel, S., &
Kiela, D. (2020b). Retrieval-Augmented Generation for
Knowledge-Intensive NLP Tasks. Neural Information Processing Systems 33
(NeurIPS 2020). <https://arxiv.org/abs/2005.11401>

Mao, Y., He, P., Liu, X., Shen, Y., Gao, J., Han, J., & Chen, W. (2020).
Generation-Augmented Retrieval for Open-domain Question Answering.
<https://doi.org/10.48550/arXiv.2009.08553>

Thulke, D., Daheim, N., Dugast, C., & Ney, H. (2021). Efficient
Retrieval Augmented Generation from Unstructured Knowledge for
Task-Oriented Dialog. Accepted by DSTC9 Workshop at AAAI-2021.
<https://doi.org/10.48550/arXiv.2102.04643>

Blattmann, A., Rombach, R., Oktay, K., Müller, J., & Ommer, B. (2022).
Retrieval-Augmented Diffusion Models. In *Advances in Neural Information
Processing Systems 35 (NeurIPS 2022) .*
<https://doi.org/10.48550/arXiv.2204.11824>

Chen, W., Hu, H., Chen, X., Verga, P., & Cohen, W. W. (2022). MuRAG:
Multimodal Retrieval-Augmented Generator for Open Question Answering
over Images and Text. *Accepted to EMNLP 2022 main conference*.
<https://doi.org/10.48550/arXiv.2210.02928>

![](https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4bd6e50da532){width="1"
height="1"}
