PT117717A - METHOD FOR PROCESSING NATURAL LANGUAGE USING MULTIPLE MODELS BASED ON CONTEXT LEVEL DOMAIN - Google Patents

METHOD FOR PROCESSING NATURAL LANGUAGE USING MULTIPLE MODELS BASED ON CONTEXT LEVEL DOMAIN

Info

Publication number
PT117717A
PT117717A PT117717A PT11771721A PT117717A PT 117717 A PT117717 A PT 117717A PT 117717 A PT117717 A PT 117717A PT 11771721 A PT11771721 A PT 11771721A PT 117717 A PT117717 A PT 117717A
Authority
PT
Portugal
Prior art keywords
context
natural language
models based
level domain
multiple models
Prior art date
Application number
PT117717A
Other languages
Portuguese (pt)
Inventor
Silva Furão Sara
Nuno De Sousa Rocha Pedro
Antonio Simões Almeida José
Original Assignee
Altice Labs S A
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Altice Labs S A filed Critical Altice Labs S A
Priority to PT117717A priority Critical patent/PT117717A/en
Priority to PCT/IB2022/062132 priority patent/WO2023126730A1/en
Publication of PT117717A publication Critical patent/PT117717A/en

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis
    • G06F40/35—Discourse or dialogue representation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities

Abstract

A PRESENTE INVENÇÃO PROVÊ UMA SOLUÇÃO DE PROCESSAMENTO DE LINGUAGEM NATURAL COM BASE NUM MÉTODO QUE UTILIZA PEQUENOS MODELOS DE DADOS TREINADOS QUE SÃO CRIADOS POR UM PROCESSO DE PARTIÇÃO SOBRE AS INTENÇÕES DE UM PROGRAMA (100). CADA MODELO DE CONTEXTO DE FLUXO CONVERSACIONAL É ATIVADO COM BASE NAS INFORMAÇÕES DE CONTEXTO DA SESSÃO, SELECIONANDO O MAIS ADEQUADO PARA DAR A MELHOR RESPOSTA, DADO O CONTEXTO DA CONVERSA (101). ESTE MÉTODO PERMITE A UTILIZAÇÃO DE VÁRIOS MODELOS DE CONTEXTO DE FLUXO CONVERSACIONAL PREVIAMENTE CRIADOS, AO INVÉS DE UM ÚNICO MODELO. O MÉTODO APLICA HEURÍSTICAS (101) SOBRE OS RESULTADOS DE UM MECANISMO DE COMPREENSÃO DE LINGUAGEM NATURAL, APROVEITANDO A PARTIÇÃO DO CONJUNTO DE DADOS PARA REDUZIR O ALTO ACOPLAMENTO ENTRE AS INTENÇÕES E, CONSEQUENTEMENTE, A AMBIGUIDADE NA ESCOLHA DA MELHOR RESPOSTA PARA CADA ENUNCIADO.THE PRESENT INVENTION PROVIDES A NATURAL LANGUAGE PROCESSING SOLUTION BASED ON A METHOD THAT USES SMALL TRAINED DATA MODELS THAT ARE CREATED BY A PARTITION PROCESS ON THE INTENTIONS OF A PROGRAM (100). EACH MODEL OF CONTEXT OF CONVERSATIONAL FLOW IS ACTIVATED BASED ON INFORMATION FROM THE CONTEXT OF THE SESSION, SELECTING THE MOST APPROPRIATE TO GIVE THE BEST RESPONSE, GIVEN THE CONTEXT OF THE CONVERSATION (101). THIS METHOD ALLOWS THE USE OF SEVERAL MODELS OF PREVIOUSLY CREATED CONVERSATIONAL FLOW CONTEXT, INSTEAD OF A SINGLE MODEL. THE METHOD APPLYS HEURISTICS (101) ON THE RESULTS OF A NATURAL LANGUAGE UNDERSTANDING MECHANISM, TAKING ADVANTAGE OF THE PARTITION OF THE DATA SET TO REDUCE THE HIGH COUPLING BETWEEN THE INTENTIONS AND, CONSEQUENTLY, THE ambiguity IN THE CHOICE OF THE BEST ANSWER FOR EACH STATEMENT.

PT117717A 2021-12-30 2021-12-30 METHOD FOR PROCESSING NATURAL LANGUAGE USING MULTIPLE MODELS BASED ON CONTEXT LEVEL DOMAIN PT117717A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PT117717A PT117717A (en) 2021-12-30 2021-12-30 METHOD FOR PROCESSING NATURAL LANGUAGE USING MULTIPLE MODELS BASED ON CONTEXT LEVEL DOMAIN
PCT/IB2022/062132 WO2023126730A1 (en) 2021-12-30 2022-12-13 Method for processing natural language using multiple models based on context level domain

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PT117717A PT117717A (en) 2021-12-30 2021-12-30 METHOD FOR PROCESSING NATURAL LANGUAGE USING MULTIPLE MODELS BASED ON CONTEXT LEVEL DOMAIN

Publications (1)

Publication Number Publication Date
PT117717A true PT117717A (en) 2023-06-30

Family

ID=84901633

Family Applications (1)

Application Number Title Priority Date Filing Date
PT117717A PT117717A (en) 2021-12-30 2021-12-30 METHOD FOR PROCESSING NATURAL LANGUAGE USING MULTIPLE MODELS BASED ON CONTEXT LEVEL DOMAIN

Country Status (2)

Country Link
PT (1) PT117717A (en)
WO (1) WO2023126730A1 (en)

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10713441B2 (en) * 2018-03-23 2020-07-14 Servicenow, Inc. Hybrid learning system for natural language intent extraction from a dialog utterance
WO2021137068A1 (en) * 2019-12-30 2021-07-08 Altice Labs, S.A. Method for natural language processing based on multi-model segments
PT116967B (en) * 2020-12-23 2023-01-30 Altice Labs S A METHOD FOR SEGMENT-BASED NATURAL LANGUAGE PROCESSING OF MULTIPLE MODELS USING A HYBRID PLN

Also Published As

Publication number Publication date
WO2023126730A1 (en) 2023-07-06

Similar Documents

Publication Publication Date Title
Milostivaya et al. Achieving Communicative Equivalence: Space-time Text Organization Peculiarities in Stream of Consciousness Novels of James Joyce in German and Russian Translations.
Huggard Wh-words in Hittite: A study in syntax-semantics and syntax-phonology interfaces
Kawahara Geminate devoicing in Japanese loanwords: Theoretical and experimental investigations
Oba et al. In-contextual bias suppression for large language models
Moessner Early Middle English Syntax
PT117717A (en) METHOD FOR PROCESSING NATURAL LANGUAGE USING MULTIPLE MODELS BASED ON CONTEXT LEVEL DOMAIN
PT116967A (en) METHOD FOR NATURAL LANGUAGE PROCESSING BASED ON MULTIPLE MODEL SEGMENTS USING A HYBRID PLN
Searle Liturgy and metaphor
Rahayu Comparison of honorific language in Javanese and Japanese speech community
Everett A closer look at a supposedly anumeric language
Kaisse The domain of stress assignment: Word-boundedness and frequent collocation
McPherson The role of music in documenting phonological grammar: Two case studies from West Africa
Tessier U‐shaped development in error‐driven child phonology
Anderson Proto-Munda Prosody, Morphotactics and Morphosyntax in South Asian and Austroasiatic Contexts
Law The negation mou5 in Guangdong Yue
O'Malley Catholic church history: One hundred years of the discipline
Tezcan A Stylistic analysis of A Rose for Emily by William Faulkner and its Turkish translation
Lowenstamm The no straddling effect and its interpretation: A formal property of Chaha 2nd feminine singular formation
Cook Double-consonant base verbs in Korean
Fitzsimmons et al. An analysis of the Split-Phonology Hypothesis in Michif: Sociolinguistic and phonological perspectives
Ettlinger Variation as a window into opacity
Bondarenko et al. Voice onset time (VOT) of L3 Spanish/ptk/by multilingual heritage speakers of Ukrainian and Polish
Pater Non-convergence in the GLA and variation in the CDA
Bolduc et al. Claude Bernard’s non reception of Darwinism
Buckley Foot alignment in Spanish secondary stress

Legal Events

Date Code Title Description
BB1A Laying open of patent application

Effective date: 20230224