Ekstrakcja odpowiedzi

Wprowadzenie do generatywnej AI w Snowflake

James Cha-Earley

Senior Developer Advocate, Snowflake

Dlaczego warto używać ekstrakcji odpowiedzi?

Identyfikowanie skarg
Ekstrakcja opinii o udogodnieniach
Odpowiadanie na pytania z długich dokumentów

Dokument ze znakiem zapytania otoczony emotikonami uśmiechniętych lub smutnych twarzy, symbolizujący brak wiedzy o treści recenzji

Typowy problem biznesowy

review = """
The property needs updating and management needs to hire more staff. 
The staff does everything they can but it is not enough for the number of guests. 
The rooms look old as well as everything in the bathrooms. 
Definitely better options in the hotel zone.
"""

Snowflake Cortex z pomocą!

Ekstrakcja odpowiedzi

from snowflake.cortex import extract_answer


response = extract_answer(
    from_text=review,
    question="What did the guest complain about?"
)

Wynik

print(response)

[
    {
    "answer": "Management needs to hire more staff.",
    "score": 0.8013839254
    }
]

Wyniki w zakresie od 0 do 1

Budowanie przepływu pracy

SELECT DESCRIPTION
FROM HOTELS.REVIEWS
WHERE RATING < 6
    AND DATE = '2023-08-23';

Wynik zapytania SQL

I will never book a vacation through Expedia not only did I pay over 5;000 dollars
but they false advertise the resort; in my reservation It clearly stated that
breakfast and dinner was included. When I got to the resort they informed me that 
these issues have happed before with Expedia because the resort does not offer an
all exclusive. After flighting practically all day I called Expedia and they had me
on hold for 45min and nothing got resolved. I ended up spending over 1000.00 more
just on food. I am beyond me they are not only false advertising but refused to
take any responsibility on their part for the issues they caused me and my family.  Again I will never chose Expedia to plan a vacation!!!

Zadawanie pytania

negative_review = cell1.to_pandas()['DESCRIPTION'].iloc[0]


response = extract_answer(
    from_text=negative_review,
    question="What was the primary issue that the guest mentioned?"
)

print(response)

Analiza odpowiedzi

[
    {
    "answer" : "Expedia falsely advertised that meals were included, but the
                resort did not offer an all-inclusive package.",
    "score": 0.75829312442
    }
]

Ograniczenia

Może zwracać częściowe lub niejasne frazy
Najlepiej stosować do akapitów, nie dokumentów wielotematycznych
Zawsze weryfikować krytyczne wyniki

Znak ostrzegawczy

Czas na ćwiczenia!

Wprowadzenie do generatywnej AI w Snowflake