Papers by Daichi Mochihashi

7 papers
Learning Adverbs with Spectral Mixture Kernels (2024.findings-acl)

Copied to clipboard

Challenge: In order for robots to collaborate with humans, it is important to share and understand their experiences through language.
Approach: They propose a hierarchical Dirichlet Process-Spectral Mixture Latent Dirichlets Allocation model which learns the relationship between human motions and adverbs by capturing frequency kernels that represent motion characteristics and shared topics of a given aadverts.
Outcome: The proposed model outperforms representative neural network models in terms of perplexity score and predicts more appropriate adverbs.
Scale-Invariant Infinite Hierarchical Topic Model (2023.findings-acl)

Copied to clipboard

Challenge: Existing hierarchical topic models yield fragmented topics with overlapping themes whose expected probability becomes exponentially smaller along the depth of the tree.
Approach: They propose a hierarchical infinite hierarchic topic model that adapts to topic creation to make expected topic probability decay considerably slower than existing models.
Outcome: The proposed model has better topic uniqueness and hierarchical diversity than existing approaches.
Analyzing Continuous Semantic Shifts with Diachronic Word Similarity Matrices (2025.coling-main)

Copied to clipboard

Challenge: Existing methods to analyze word sense proportions are insufficient for understanding semantic shifts . et al., 2018: semantic shift and its effects.
Approach: They propose a framework for how semantic shifts occur over multiple time periods by using word embeddings.
Outcome: The proposed framework can analyze semantic shifts over multiple time periods using word embeddings.
Holographic CCG Parsing (2023.acl-long)

Copied to clipboard

Challenge: Existing methods for supertagging and parsing use black-box neural architectures to implicitly model phrase structure dependencies.
Approach: They propose a method for formulating CCG as a recursive composition in a continuous vector space by using holographic embeddings as holography operator.
Outcome: The proposed method can achieve comparable performance to state-of-the-art parsing with Transformers.
How LSTM Encodes Syntax: Exploring Context Vectors and Semi-Quantization on Natural Text (2020.coling-main)

Copied to clipboard

Challenge: LSTMs are widely used to capture informative long-term syntactic dependencies, but how they are reflected in their internal vectors for natural text has not been adequately investigated.
Approach: They analyze how syntactic dependencies are reflected in LSTM's internal gates by learning a language model where syntaktic structures are implicitly given.
Outcome: The proposed model can predict whether a word is inside a phrase structure or not from a small number of components of the context-update vector.
Cross-lingual and Word-Independent Methods for Quantifying Degree of Grammaticalization (2026.eacl-long)

Copied to clipboard

Challenge: Existing methods for quantifying the degree of grammaticalization are language- and word-dependent . existing methods are language dependent and lack training data .
Approach: They propose to use Positive-Unlabeled learning or Cross-Validation-like learning to quantify degree of grammaticalization.
Outcome: The proposed method achieves high correlations to human judgments in English deverbal prepositions and Japanese nouns being grammaticalized.
Infinite SCAN: An Infinite Model of Diachronic Semantic Change (2022.emnlp-main)

Copied to clipboard

Challenge: Existing methods for capturing semantic changes using word embeddings cannot account for existence of each sense and its relative importance.
Approach: They propose a Bayesian model that can estimate the number of senses of words and their changes through time using a dynamic topic model and a logistic stick-breaking process.
Outcome: The proposed model outperforms the baseline model and investigates the semantic changes of several well-known target words using the CCOHA corpus.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations