Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386789981> ?p ?o ?g. }
Showing items 1 to 57 of
57
with 100 items per page.
- W4386789981 abstract "Recurrent Neural Networks (RNNs) have shown an incredible performance in supervised machine learning tasks such as Natural Language Processing (NLP). However, theoretical understanding of RNNs' performances in NLP are still limited due to intrinsically complex non-linear computations of RNNs. This thesis explores a class of RNNs called Recurrent Arithmetic Circuits (RACs), possessing a dual mathematical representation as a Matrix Product State (MPS) widely used in many-body quantum physics. The duality allows us to compute the entanglement entropy of an MPS, which can be used as a proxy for information propagation in the dual neural networks, to phenomenologically explain the RNNs-based model prediction accuracy's behaviors in NLP. We found that the entanglement entropy saturates when the accuracy saturates in the fixed word embedding case. The unfixed word embedding experiments also reveal that the entanglement entropy of the RACs is decaying as the word embedding becomes more meaningful, as reflected by the behaviors of cosine similarity between word embeddings. This thesis sheds light on more transparent and explainable behaviors of RNNs-based machine learning in NLP, using tools from many-body quantum physics." @default.
- W4386789981 created "2023-09-16" @default.
- W4386789981 creator A5005716394 @default.
- W4386789981 date "2023-09-15" @default.
- W4386789981 modified "2023-09-26" @default.
- W4386789981 title "Towards explainable sentiment analysis for written reviews via quantum tensor network states" @default.
- W4386789981 doi "https://doi.org/10.58837/chula.the.2021.318" @default.
- W4386789981 hasPublicationYear "2023" @default.
- W4386789981 type Work @default.
- W4386789981 citedByCount "0" @default.
- W4386789981 crossrefType "dissertation" @default.
- W4386789981 hasAuthorship W4386789981A5005716394 @default.
- W4386789981 hasBestOaLocation W43867899811 @default.
- W4386789981 hasConcept C121040770 @default.
- W4386789981 hasConcept C121332964 @default.
- W4386789981 hasConcept C132525143 @default.
- W4386789981 hasConcept C147168706 @default.
- W4386789981 hasConcept C154945302 @default.
- W4386789981 hasConcept C204321447 @default.
- W4386789981 hasConcept C2777462759 @default.
- W4386789981 hasConcept C41008148 @default.
- W4386789981 hasConcept C41608201 @default.
- W4386789981 hasConcept C50644808 @default.
- W4386789981 hasConcept C62520636 @default.
- W4386789981 hasConcept C66402592 @default.
- W4386789981 hasConcept C80444323 @default.
- W4386789981 hasConcept C84114770 @default.
- W4386789981 hasConceptScore W4386789981C121040770 @default.
- W4386789981 hasConceptScore W4386789981C121332964 @default.
- W4386789981 hasConceptScore W4386789981C132525143 @default.
- W4386789981 hasConceptScore W4386789981C147168706 @default.
- W4386789981 hasConceptScore W4386789981C154945302 @default.
- W4386789981 hasConceptScore W4386789981C204321447 @default.
- W4386789981 hasConceptScore W4386789981C2777462759 @default.
- W4386789981 hasConceptScore W4386789981C41008148 @default.
- W4386789981 hasConceptScore W4386789981C41608201 @default.
- W4386789981 hasConceptScore W4386789981C50644808 @default.
- W4386789981 hasConceptScore W4386789981C62520636 @default.
- W4386789981 hasConceptScore W4386789981C66402592 @default.
- W4386789981 hasConceptScore W4386789981C80444323 @default.
- W4386789981 hasConceptScore W4386789981C84114770 @default.
- W4386789981 hasLocation W43867899811 @default.
- W4386789981 hasOpenAccess W4386789981 @default.
- W4386789981 hasPrimaryLocation W43867899811 @default.
- W4386789981 hasRelatedWork W2335882425 @default.
- W4386789981 hasRelatedWork W2745862583 @default.
- W4386789981 hasRelatedWork W2896411932 @default.
- W4386789981 hasRelatedWork W3033291431 @default.
- W4386789981 hasRelatedWork W3048659618 @default.
- W4386789981 hasRelatedWork W3080191145 @default.
- W4386789981 hasRelatedWork W3170430489 @default.
- W4386789981 hasRelatedWork W4210823838 @default.
- W4386789981 hasRelatedWork W4312422125 @default.
- W4386789981 hasRelatedWork W4317564474 @default.
- W4386789981 isParatext "false" @default.
- W4386789981 isRetracted "false" @default.
- W4386789981 workType "dissertation" @default.