Skip to main content

Practical Example - Deep Learning

  1. If you don't have an account yet - register.

  2. Login to botx console - You should be redirected to the dashboard page

  3. In the top left corner, click on the hamburger menu - This should open a mega menu with 19 menu options

  4. Click on the "Data Studio" - You should be redirected to the Data Studio page. The loading bar should be animated, and after loading is finished, you should see "Databases" and "File Manager" tabs. The loading can take 10-20 seconds.

  5. In the "Databases" tab, click on the "+" button to create a new Database - You should be redirected to a "Setup Wizard" page

  6. Type in any name and click "Next." - After brief loading, the page should update to step 2, "Create Model." If you don't have a sufficient subscription, you should not be allowed to carry on, and instead, you should see an error.

  7. Type in any name, leave other settings as default, and click "Next." - After brief loading, the page should update to step 3, "Create Columns." If you don't have a sufficient subscription, you should not be allowed to carry on, and instead, you should see an error.

  8. Click on the button with three dots - You should now see a textbox with "Paste columns from text..."

  9. Paste all of the following into the textbox and click the "+" button - All 11 fields should now be displayed in a data grid. You should see "Type" as "Text" for all and "Name" precisely as we entered them. Each field should have a delete button next to it.

  10. Click on the delete icon next to the "colcheck" field, and when prompted, click "Yes." - The Colcheck field should now be deleted and no longer visible.

  11. Change the "age" column to "number" and the rest of them to "list" - Types should change accordingly

  12. Click on the new column to add 2x new columns - Two new empty columns should be added with the type "Text"

  13. Into the top one, type in highbp, and into the one below, type in BMI - The names should change accordingly

  14. Click on the next button on the bottom of the page - After loading (20-30 sec), you will be shown a success page with all the information that you've added

  15. Click the "Enter Database" button - You should be redirected to the database's model. The database will be empty, as we haven't imported any information in yet.

  16. In the left panel, right-click on the model under your database - You should see a menu appear showing "Details," "Edit Schema," "Delete"

  17. Click on "Details" and change the name - The model name should change accordingly

  18. In the left panel, right-click on the model again and choose "Edit Schema" - A Schema Editor dialog should open

  19. Hover over and click on the data type and name - The values should not be editable, and the hover should show a popup confirming this.

  20. For the first column (highbp), change the display name to High Blood Pressure - The display name should adjust accordingly.

  21. For the last column (diabetes_binary), change the display name to Has Diabetes. - The display name should change accordingly.

  22. Click on the "Save" button - After a short load, the dialog should auto-close.

  23. Stay on the same page (model view) and click the "New Entry" button - A dialog should popup with all the pre-defined columns. Only High Blood Pressure and Has Diabetes? columns should show their display names. The rest of the columns should show their original column names.

  24. Close the dialog and click on the "Import CSV" button - A dialog should open up with the option to upload files.

  25. Choose File and upload the file attached in this step - A dialog should now show all the detected columns from the uploaded file.

  26. Click on the "Infer Assignment" button to automatically assign the uploaded file's columns to the model's columns - Each detected column (from .csv) should now have the equivalent of the target column (from the model). Autofill should be on for all the "list" type columns. All columns should have the "Use" option checked.

  27. Click on the "Import" button. After briefly loading, the data should be imported. The dialog should autoclose, and you should see the first page of the results. A total of 4419 items should be imported.

  28. Click on the pencil icon on any entries - A dialog should pop up with the values corresponding to that entry.

  29. In the top left corner, click on the hamburger menu and select "Add New Bot." - You should be redirected to the "New Bot" page

  30. Add any name and choose the "AI Agent" bot type - Options should change accordingly.

  31. Click on the "Create" button - The page should show a success message showing you the "Set Robot" and "Create Robot" options. If you don't have a sufficient subscription -> the robot should not be created, showing you an error message.

  32. Click on the "Set robot" - You should be redirected to the robot page with a popup showing you "Start from scratch" or "Create from template"

  33. Choose "Create from Template" and click on the "Deep Learning" option - 6 nodes should be auto-loaded into the diagram. You should see the nodes precisely as they are in the screenshot attached.

  34. In the Database node, choose the database and model that you've created previously - The select boxes should be updated accordingly

  35. Click on the settings icon in the top right corner of the node - A side panel should show up with advanced settings for the node. All the columns from the model should be shown in the panel.

  36. Ensure that all the columns are visible in the node's advanced settings (12x) - All 12x columns should be visible.

  37. Click on the select box next to the diabetes_binary and change the "Input" to "Predict." - The value of the special box should change immediately.

  38. Click on the "Deep Learning" node - Advanced settings of the node should show up in the side panel.

  39. Click on the "Deep Learning Editor" button - A "Neural Network Editor" dialog should appear.

  40. Click on the "Auto Suggest Params" button - After brief loading, 3x layers should be added to the left side of the dialog, and a neural network with the connections should be shown on the right side.

  41. Click on the "Save" button in the dialog - The "Neural Network Editor" dialog should close.

  42. Click on the "Save" button next to the "Publish" button above the diagram - A confirmation toast should show up on the bottom of the page, saying "Successfully saved."

  43. Turn on the robot by clicking on the "Online" switch above the "Publish" button - After brief loading, the robot should change from Offline to Starting to Idle.

  44. In the "Deep Learning" node, click the "Start Training" button - The robot should immediately start training, changing its state from Idle to Training. After a short while, the robot should be done with training and shown again as Idle.

  45. In the "Input Database" node, click on the advanced settings and then "Edit data in DCS." - You should be redirected to the Data Studio page right into the databases table view.

  46. In the left treeview panel, right-click on your table and click on the "Edit Schema" - Schema Editor dialog should show up.

  47. Click on the "New Field" button and using the six dots - drag it all the way down - A new empty field should be added and dragged all the way down below the diabetes_binary field.

  48. Change its type to "Button" and add any name - Type and name are changed accordingly.

  49. Click on the settings icon of the field - A new dialog should show up with advanced settings of the specific field.

  50. Change the button caption to "Predict" - The caption changes accordingly

  51. Choose color, variant, and size to whatever you want - The values should change accordingly.

  52. Click on the + button next to the "Sequence of actions." - Two new select boxes should show up. The first select box should be "AI Agent."

  53. Change the second select box to the name of your AI Agent from previous steps - Another two new select boxes should appear afterward.

  54. Change the third select box to "Serialize Fields." The select box should change accordingly.

  55. Change the fourth select box by clicking "Select All" - All columns should be selected.

  56. In the list of columns in the fourth select box, scroll all the way down, uncheck the button and diabetes_binary columns, and click OK - The fourth select box should say "11 selected".

  57. Click on the "Save" button and then again on the "Save" button - The changes should now be saved, and the dialog should auto-close.

  58. Click on the pencil icon next to the any of the entries in the table - Dialog should show up. We should now see the "Predict" button that we added in the Scheme Editor.

  59. For the "Has Diabetes?" field, remember its value and then remove it by clicking on the "x" button. The value should now be removed, and instead, the placeholder text "Select a value..." should be shown.

  60. Click on the "Predict" button. - The correct values should now be automatically added to the "Has Diabetes?" field.

  61. Return to your AI Agent and turn it off by clicking on the "Online" switch - The robot should successfully go from Idle to Offline.