InfraNodus is an AI-powered text analysis and visualization tool that helps you to:
1. Analyze any text;
2. Develop your ideas;
3. Study an existing discourse;
You can identify the main concepts and topics within, reveal the relations between the ideas, and use the AI to generate interesting research questions and facts that can help you develop the discourse further.
Here is a basic workflow that you can use to understand how InfraNodus works and to start using it to generate new ideas.
You can watch the video to understand how it works or follow our step-by-step guide below.
Step 1: Create a New Graph or Import Existing Data
Go to the Apps and click "Create a New Graph". You can also do that using the + Add New Graph link in the main menu:
You can also import existing data. For example, your Evernote notes, Google search results, Twitter discourse, current news (via RSS feeds), Amazon product reviews, and much more:
You can also use the specialized apps that allow you to create mind maps, word clouds, or use the live AI ideation system:
Step 2: Choose Your Workflow
Once you open a new graph, you will see a dialog that will help you choose the right workflow based on your objective:
- Analyze a Text — use this option if you have an already existing text or text files that you want to analyze.
- Develop an Idea — use this option if you have an abstract of an idea or a research question that you would like to develop further using the live graph feedback and the AI
- Explore a Topic — use this option if you want to start with a keyword or a topic to better understand the context around it using the AI, Google search results, Wikipedia or Google Scholar.
💡 If you would like to Analyze a Text, please, proceed to the following tutorial:
Step 3: Develop an Idea / Explore a Topic
For the sake of this demo, we assume that you selected Explore a Topic option. It works similarly to the other option, Develop an Idea, if you choose the "Live AI Generator" mode once you get in:
In the example above, we added "machine learning" into the search box, choose the "Live AI Ideator" and then click Proceed.
We could also add that exact same phrase into the "Develop an Idea" dialogue, and the resulting InfraNodus behavior and workflow would be similar to what we demonstrate below.
Step 4: Visualize the Content Added as a Network Graph
InfraNodus will visualize the idea you added on the graph. It will then propose you to select the nodes on the graph and choose the insight engine that you want to use to generate more ideas.
You can also add your own text to the graph using the +add text editor panel on the left:
Step 5: Add More Text Content
Use the + add text button at the bottom left to add your own ideas or the AI insight panel to generate the related ideas. You can also use both.
If you decide to add your text, simply:
- open the text editor,
- write your idea,
- choose a tag (e.g. "ideas") — this is optional
- click "save" to add it to the graph
Once you add this text, it will be visualized in a graph, the main topics and the key concepts will be detected, you will also be able to use the advanced analytics and the AI insight generator to come up with new ideas.
This is a truly iterative process that is well-suited for adding free associations about a certain topic. Don't focus too much on coherency at this point.
If you have an already prepared text, you can copy and paste it into the editor box and then use our Graph Exploration Workflow to explore the ideas within.
Step 6: Using the AI Insight Helper to Generate New Ideas
Let's assume that you decided not to write your own text, but to add some ideas using the AI first, to build your graph.
Select the nodes, click Proceed, and you will see the following options:
- AI Questions — generates the research questions using GPT-3 AI related to the concepts that you selected;
- AI Facts — generates interesting facts using GPT-3 AI that relate to the concepts you selected
- AI Ideas — generates innovative ideas using GPT-3 AI
- Google — will show you the top Google search results for the concepts selected
- Context — will show how the concepts selected are used in your target language typically
- Wikipedia — will get the results from Wikipedia on the topics selected
For example, we will click AI Facts, and get the following results:
Click "More Facts" to generate more facts
Click "Save to Graph" if you like the idea and want it saved into the graph. When you click this button, it will be saved with a "gpt3" tag, so you can later separate it from your own ideas (marked with "ideas" tag, for example).
Step 7: Explore the Graph
Here's what the graph will look like after you add a few ideas (either generated by the AI or your own) into InfraNodus:
The words are the nodes and their co-occurrences are the connections between them. Once we build a graph in this way, we can use powerful algorithms from graph theory and network analysis to detect the most influential concepts, the topical groups, and the gaps between them.
Here's a brief explanation of how you can read the graph. Learn more in How to Read and Interpret Text Network Graphs article
a. The Most Influential Concepts
The main concepts are shown bigger on the graph. The measure we use to rank them is based on betweenness centrality measure that detects the nodes that connect distinct groups of nodes together. In our case, these are the concepts learn, learning and models
b. Topical Groups
The topical groups are the nodes that are placed closer to each other on the graph and have a specific color. We use the community detection algorithm used in network science to detect those clusters of nodes in combination with Force Atlas layout algorithm to visualize their location.
For example, in the graph above we see a green topical group on the right that talks about how AI models can be sexist and racist if they are exposed to biased data.
Both the most Influential Concepts and the Topical Groups is also shown in the Analytics Panel on the right:
This gives us a pretty good overview of the existing discourse and shows which topics it consists of.
We can develop it further by generating more facts and adding more content.
Step 8: Add More Ideas
Let's zoom into the main topic "learn sexist exposed" to elaborate in a bit more detail on that subject.
You can add your own ideas using the text editor at the bottom left or use the AI insight generator.
For the purposes of this tutorial, let's use the AI to help us develop this discourse further. Select the nodes "biased" and "data", and then click "Back" in the AI Insight Panel and choose the AI Facts model:
The system will generate several facts. Add them into the graph (click the "save to graph" button).
After a few iterations, we will see that the graph is expanded and contains much more information on the concept of "bias" than before:
Step 9: Explore the Structural Gaps in Ideas
Let us now explore the structural gaps in the current discourse. Structural gaps appear between the topical groups that could be connected but are not. Bridging those gaps can help us generate more ideas.
Deselect all the selected nodes, then click "back" in the AI Insight panel or Analytics > Insight to show the structural gap in this discourse:
We can then use the AI Insight Panel to generate new ideas that will bridge those gaps or use the text editor to write our own ideas that would introduce a connection between these two topical groups.
Step 10: Bridge the Structural Gaps
In our case, we are focusing on a link between the concept of "belief" and "bias" — what if the bias in machine applications is a reflection of human capacity to believe. A link between artificial intelligence and religion, maybe?
We can add our idea using the text editor and choose the "idea" tag, so we can separate it from the AI-generated content later:
We will add this idea manually and tag it with "sidetracking" as we're moving away from our main topic. Once it's added into the graph, it will link those two groups of concepts in a new way:
We can continue this thread and elaborate further on the link between the religion, belief, machine learning, and bias, turning our text into a philosophical direction.
Step 11: Explore the Graph, Reiterate
Once we are done, we can use the basic Graph Exploration Workflow to develop this text further.