EconomicsAI

Introduction

This project -- named EconomicsAI -- is for Large Language Model (LLM) derived explanations of economics topics.

EconomicsAI uses LLM Retrieval Augmented Generation (RAG). The RAG vector databases are created over economic articles, blog posts, or interview transcripts.

The initial vectors databases are created using the transcripts of interviews or discussions from the YouTube channels:

• "The Greater Eurasia Podcast", [GDc1]

• "Dialogue works", [NAc1]

• "Robinson's Podcast", [REc1]

Remark: Currently, most of the text data (≈95%) is from interviews with Prof. Richard D. Wolff and Prof. Michael Hudson.

Web interface

EconomicsAI is available through the Web interface RAG-evaluations.

Workflows

The following flowchart shows the vector databases creation and their utilization for producing RAG-based LLM answers of user queries.

flowchart TD
    subgraph LocalVDB[Local Folder]
        direction LR
        A(Vector Database 1)
        B(Vector Database 2)
        C(Vector Database N)
    end

    subgraph Creation
        ID
        SD
        EV
        CD
    end

    subgraph RAG[The RAG]
        D 
        E
        CC
        T 
        GP   
    end

    ID[Ingest document collection]
    SD[Split Documents]
    EV[Get LLM Embedding Vectors]
    CD[Create Vector Database]
    ID --> SD --> EV --> CD

    EV <-.-> LLMs
    
    CD -.- CArray[[CArray<br>representation]]

    CD -.-> |export| LocalVDB

    LocalVDB -.- JSON[[JSON<br>representation]]

    LocalVDB -.-> |import|D[Ingest Vector Database]
 
    D -.- CArray
    D --> E
    E[Get query] --> CC[Query vector embedding]
    CC --> T[Find nearest neighbors]
    T --> GP[Get augmented LLM response]
    T -.- |nearest neighbors<br>distance function|CArray
    

    subgraph LLMs
        OpenAI{{OpenAI}}
        Gemini{{Gemini}}
        MistralAI{{MistralAI}}
        LLaMA{{LLaMA}}
    end

Loading

---

References

Articles

[AA1] Anton Antonov, "Outlier detection in a list of numbers", (2022), RakuForPrediction at WordPress.

Packages

[AAp1] Anton Antonov, WWW::OpenAI Raku package, (2023), GitHub/antononcube.

[AAp2] Anton Antonov, WWW::LLaMA Raku package, (2024), GitHub/antononcube.

[AAp3] Anton Antonov, LLM::Functions Raku package, (2023-2024), GitHub/antononcube.

[AAp4] Anton Antonov, LLM::Prompts Raku package, (2023-2024), GitHub/antononcube.

[AAp5] Anton Antonov, ML::FindTextualAnswer Raku package, (2023-2024), GitHub/antononcube.

[AAp6] Anton Antonov, Math::Nearest Raku package, (2024), GitHub/antononcube.

[AAp7] Anton Antonov, Math::DistanceFunctions Raku package, (2024), GitHub/antononcube.

[AAp8] Anton Antonov, Statistics::OutlierIdentifiers Raku package, (2022), GitHub/antononcube.

Videos, video channels

[AAv1] Anton Antonov, "Raku RAG demo", (2024), YouTube/@AAA4Prediction.

[GDc1] Glenn Diesen, "The Greater Eurasia Podcast", (2011-2025).

[NAc1] Nima Alkhorshid, "Dialogue works", (2021-2025).

[REc1] Robinson Erhard, "Robinson's Podcast", (2022-2025).