KnowBot: phase 4 final report

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F20%3A43960999" target="_blank" >RIV/49777513:23520/20:43960999 - isvavai.cz</a>

Result on the web

—

DOI - Digital Object Identifier

—

Result language

angličtina

Original language name

KnowBot: phase 4 final report

Original language description

The task of the fourth phase of the project is to improve the chatbot by employing current state-of-the-art methods in information retrieval and question answering. These methods are based on Transformer architecture. Concretely we use BERT-like models to make use of pre-training on a large amount of unlabeled textual data. Special attention is paid to response speed. To optimize this, we use a two-tower model that enables us to precompute semantic representations of the whole document collection to reduce response time significantly.

Czech name

—

Czech description

—

Type

V_souhrn - Summary research report

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

N - Vyzkumna aktivita podporovana z neverejnych zdroju

Publication year

2020

Confidentiality

C - Předmět řešení projektu podléhá obchodnímu tajemství (§ 504 Občanského zákoníku), ale název projektu, cíle projektu a u ukončeného nebo zastaveného projektu zhodnocení výsledku řešení projektu (údaje P03, P04, P15, P19, P29, PN8) dodané do CEP, jsou upraveny tak, aby byly zveřejnitelné.

Number of pages

8

Place of publication

—

Publisher/client name

McKinsey &amp; Company, Inc. Prague

Version

—