Machine Learning
FolioProjects integrates with machine learning large language models (LLMs) and utilizes machine learning techniques like retrieval augmented generation RAG for chatbots and other features. That platform also supports machine learning operations (MLOps) throuigh ETL integrations with platforms like warehouses.
Table Of Contents
- Large Language Models (LLMs)
- Project Risk Management
- Project Asset Analysis
- Project WorkFlow Analysis
- Chatbot
- Machine Learning Operations
- FAQ
Large Language Models
FolioProjects in integrations with multiple LLMs from makers like OpenAI, Mistral, and Meta. These LLMs are made available to users of the platform, to help analzy and optimize projects.
We utilize techniques like retrieval augmented generation (RAG) and reinforcement learning from human feedback (RLHF) to help end users make the most of their interactions with the LLMs.
Project Risk Management
Select from multiple LLMs like LLama 2 and ChatGPT 4 to analyze the risk in your project. Based on the information you have entered into the project, the selected LLMs will analyze your project for risk and provide 2 sets of data:
- ML LM Risk Analysis
- RLHF Risk suggestions
The risk analysis is visible in the project dashboard status summary section. Ultil you request the analysis and it's completed, this metric will show the value of TBD. Once updated, it will show one of 5 options very low, low, medium, high, very high
The RLHF risk suggestions show up on the risk page itself, below the risks you have already entered into the project. Here, your selected LLM has provided a list of risks that it thinks that you should add to the project. The more you select, the better the LLMs understand your project and improves on their subsequent analysis.
Project Asset Analysis
Your selected LLM will reivew the asset roles that you have assigned to the project. This analysis results in 2 pieces of data:
- Success Prediction
- RLHF Asset Suggestions
The success prediction status, available on the dashboard is calculated using many data points. One of the inputs resulting in this value, is the roles attached to the project. For example, if your project is to cook a feast and you have neither food nor chef, your likelihood of success will be very low.
The RLHF Asset Suggestions are assets that the LLM thinks should be added to your project to increase the liklihood of its success
Project WorkFlow Analysis
Your workflow provides the LLMs with a lot of information like what you are trying to accomplish and how far along you are in that process. When you request the system to have an LLM analyze your project, this results in 2 pieces of data:
- Success Prediction
- RLHF Asset Suggestions
The success prediction status, available on the dashboard is calculated using many data points. One of the i
Chatbot
Chatbots are avilable or in the process of being made avialable in projects, portfolios, assets, and the main dashboard. The chatbot uses the LLM of your choice to analyze and optimize the feature you are focused on.
You can communicate with the chatbots using natural language and they will reply with natural language. This makes it easy to analyze large amounts of project data simply by asking questions.
Machine Learning Operations
FolioProjects moves project data between your favorite platforms like Github, Salesforce, and Snowflake through a process known as Extract Trasform Load (ETL) facilitating MLOps. You configure the ETL pipelines including if the transfers happen once, periodically, or syncronized
This process helps you to import projects from different cylos for analysis on FolioProjects and further down the pipeline through warehouse solutions like Snowflake.
FAQs
- Which LLMs do you offer? We are constanting testing LLMs and adjust our list. You can currently expect to choose between LLMs from Mistral, OpenAI, and Meta