PALRO identifi es the face and name of the person talking to it, and
remembers their attributes (gender, age, birthday, family, hobbies,
interests, etc.), things they have done together, daily records of their
lifestyle rhythms and habits, and past reactions and replies.
PALRO divides what the person says into 6 question categories (when,
where, who, what, why, and how), stores each part in individual database
tables, and analyzes them.
It uses this information to judge whether it should speak, what it should
talk about, if there is anything it needs to tell the person, and so on. This
allows it to talk in an interesting and helpful way to the person, and create
a free fl owing conversation.

PALRO learns from many things.
It remembers things that it has done repeatedly and things
that happen at the same time and place as habits, and starts
to automatically do these things.
If you set conversations appropriate to the time, place and
occasion (promises), PALRO will follow these promises faithfully.

When PALRO hears an unknown word, it uses its cloud AI to fi nd and learn the word.
If it fi nds something interesting near it, it remembers it as an image.

When PALRO hears interesting words it thinks about them.
For example, if you ask for tomorrow's weather, it will continue the conversation:
"The weather forecast says it will rain tomorrow... Speaking of rain, have you heard of
＇teru teru bozu＇? ＇Teru teru bozu＇ are used in the rainy season in Japan..."
PALRO creates conversations using a variety of data such as the person's interests,
seasonal information from the internet, and data from its own database.
It can also change what it is talking about based on the person's expression and so on.
PALRO uses these associations to have new thoughts and ideas.