Introducing Classifier Bots

I have added more Machine Learning bots. They are of labelled learning types. 

These are Machine Learning bots that analyse inputs (texts and images) and classify them into preset range of values (labels). The default range contains values from -3 (extreme negative), to 0 (neutral), to 3 (extreme positive).  Already I found so many applications about how they can be used.

		Examples of scenarios
Value	Description	Survey	Product Review	Political Leaning
-3	Strongly negative	Strongly disagree	Strongly dislike	Extreme left-wing
-2	Negative	Disagree	Dislike	Left-wing
-1	Somewhat negative	Somewhat disagree	Some improvement	Somewhat left-wing
0	Neutral	Neither agree or disagree	No comment	Neutral
1	Somewhat positive	Somewhat agree	Ok	Somewhat right-wing
2	Positive	Agree	Like	Right-wing
3	Strongly Positive	Strongly agree	Strong favourable	Extreme right-wing

How good is that to have these bots working for us? The good thing is that, with the current underlying Sidekick's framework, these bots can be created quickly. Some examples quickly came to mind were:

• ✨ Positivity Analyser bot. This bot analyses inputs and classifies whether they are negative or positive.
• 🎭 Political Leaning Analyser bot. This bot identifies whether inputs are leaning toward left-wing or right-wing.
• 🌏 Climate Change Analyser bot. This bot analyses inputs and gauges whether they reflect negatively or positively toward the issues of climate changes, global warming, etc.
• 🧿 AI Acceptance Analyser bot. This bot identifies whether inputs sound negatively or positively toward the application of AI.

How will these bots be used?

Immediately, one sample usage that occurred to me was to use them internally - to analyse the inputs that are feeding to Sidekick server. It was quite interesting to see movements on these gauges here.

These capabilities were released in Bootcamp build 14 (12 March 2024). At time, these bots were under training and the APIs to access to them are readily available. Please let me know in the comment section if you are interested in accessing them.