Papers by Ryohei Sasano

Acquiring Frame Element Knowledge with Deep Metric Learning for Semantic Frame Induction (2023.findings-acl)

Challenge:	Semantic frame induction is the task of clustering frame-evoking words according to the semantic frames they evoke.
Approach:	They propose a prompt-based method for obtaining Frame Embeddings that outputs One frame-name as a Label .
Outcome:	The proposed method outperforms existing methods on English and Japanese datasets.

Copied to clipboard

Information Extraction & Text Mining

Challenge:	Existing methods for semantic frame induction are labor intensive . a method that uses contextualized embeddings can be used to acquire frame element knowledge.
Approach:	They propose a method that applies deep metric learning to semantic frame induction tasks . they use a pre-trained language model to fine-tune frame-annotated models to perform argument clustering .
Outcome:	The proposed method achieves substantially better performance than existing methods on FrameNet.

Comparison and Combination of Sentence Embeddings Derived from Different Supervision Signals (2022.starsem-1)

Copied to clipboard

Hayato Tsukagoshi, Ryohei Sasano, Koichi Takeda

An Empirical Study on Fine-Grained Named Entity Recognition (C18-1)

Challenge:	Existing methods to derive sentence embeddings have not been well understood what properties are captured in the resulting sentences depending on the supervision signals.
Approach:	They propose to combine two types of sentence embedding methods with similar architectures and tasks to investigate their properties.
Outcome:	The proposed methods perform better on unsupervised and downstream tasks than the proposed methods on untrained STS tasks and probing tasks.

Copied to clipboard

Khai Mai, Thai-Hoang Pham, Minh Trung Nguyen, Tuan Duc Nguyen, Danushka Bollegala, Ryohei Sasano, Satoshi Sekine

Information Extraction & Text Mining

Challenge:	Named entity recognition (NER) is a well studied topic in natural language processing.
Approach:	They propose to remove the CNN layer and use dictionary and category embeddings to improve Japanese FG-NER performance.
Outcome:	The proposed method improves Japanese FG-NER F-score from 66.76% to 75.18%.

WikiSplit++: Easy Data Refinement for Split and Rephrase (2024.lrec-main)

Copied to clipboard

Hayato Tsukagoshi, Tsutomu Hirao, Makoto Morishita, Katsuki Chousa, Ryohei Sasano, Koichi Takeda

Reasoning

Challenge:	Existing text simplification methods rely on encoder-decoder models to achieve this task.
Approach:	They propose a text-to-text generation approach that applies encoder-decoder models to a large-scale dataset to improve Split and Rephrase.
Outcome:	The proposed approach improves Split and Rephrase readability and performance on large datasets, but still suffers from hallucinations and under-splitting.

Cross-lingual Linking of Automatically Constructed Frames and FrameNet (2022.lrec-1)

Copied to clipboard

Ryohei Sasano

Cross-Lingual Transfer

Challenge:	Existing semantic frame resources have been manually elaborated, but manual development is labor-intensive.
Approach:	They propose to link Japanese frames to English FrameNet by using cross-lingual word embeddings and a model that takes only the frame-evoking words into account.
Outcome:	The proposed model will facilitate the development of cross-lingual frame resources.

Verifying Claims About Metaphors with Large-Scale Automatic Metaphor Identification (2024.naacl-short)

Copied to clipboard

Kotaro Aono, Ryohei Sasano, Koichi Takeda

Sentiment Analysis

Challenge:	Existing studies on metaphors have focused on a small number of examples, whereas few studies verify claims with large corpus.
Approach:	They propose to use a large corpus to verify existing claims about verb metaphors . they apply metaphor detection to sentences extracted from Common Crawl .
Outcome:	The proposed method identifies verb metaphors with lower concreteness, imageability, familiarity and more emotional and subjective sentences.

Copied to clipboard

Hongkuan Zhang, Saku Sugawara, Akiko Aizawa, Lei Zhou, Ryohei Sasano, Koichi Takeda

Multimodality

Challenge:	Existing image captioning approaches treat image-caption pairs indistinctly without considering the differences in their learning difficulties.
Approach:	They propose a pretrained vision–language model that measures cross-modal similarity and a model that uses cross-module similarity to measure the difficulty of captioning.
Outcome:	The proposed model achieves superior performance and competitive convergence speed to baselines without incurring additional training costs.

Transformer-based Live Update Generation for Soccer Matches from Microblog Posts (2023.emnlp-main)

Copied to clipboard

Masashi Oshika, Kosuke Yamada, Ryohei Sasano, Koichi Takeda

Building a Buzzer-quiz Answering System (2023.acl-srw)

Challenge:	Existing systems to generate sports updates from tweets are not able to handle vast amounts of diverse tweets, and this remains a challenge for future studies.
Approach:	They propose to generate live updates for soccer matches from tweets using a large pre-trained language model and incorporate a classifier to control the number of updates and a mechanism to reduce redundancy of duplicate and similar updates.
Outcome:	The proposed system can generate live updates for soccer matches from tweets and achieve high performance by considering preceding updates.

Copied to clipboard

Naoya Sugiura, Kosuke Yamada, Ryohei Sasano, Koichi Takeda, Katsuhiko Toyama

Natural Language Interfaces

Challenge:	A buzzer quiz is a genre of quiz in which multiple players simultaneously listen to a quiz being read aloud and respond it by buzzing in as soon as they can predict the answer.
Approach:	They propose two types of buzzer-quiz answering systems: a system that directly generates an answer from part of a question by using an autoregressive language model and a second system that reconstructs the entire question by applying an autoreregressively language model.
Outcome:	The proposed system estimates the accuracy of the answers by using the internal scores of each model.

Semantic Frame Induction from a Real-World Corpus (2025.acl-srw)

Copied to clipboard

Shogo Tsujimoto, Kosuke Yamada, Ryohei Sasano

Automating Interlingual Homograph Recognition with Parallel Sentences (2022.findings-aacl)

Challenge:	Existing studies on semantic frame induction have demonstrated that pre-trained language models (PLMs) have led to more accurate results.
Approach:	They conduct semantic frame induction using the Colossal Clean Crawled Corpus and assess the applicability of existing frame inducing methods to real-world data.
Outcome:	The proposed methods outperform existing methods on real-world data and can induce frames corresponding to novel concepts.

Copied to clipboard

Yi Han, Ryohei Sasano, Koichi Takeda

Cross-Lingual Transfer

Challenge:	Existing methods for interlingual homograph recognition require linguistic knowledge and massive annotation work.
Approach:	They propose an automatic interlingual homograph recognition method based on cross-lingual word embedding similarity and co-occurrence of form-identical words in parallel sentences.
Outcome:	The proposed method can make accurate predictions across languages.

Improving Sentence Embeddings with Automatic Generation of Training Data Using Few-shot Examples (2024.acl-srw)

Copied to clipboard

Soma Sato, Hayato Tsukagoshi, Ryohei Sasano, Koichi Takeda

Sequential Span Classification with Neural Semi-Markov CRFs for Biomedical Abstracts (2020.findings-emnlp)

Challenge:	Decoder-based large language models (LLMs) have shown high performance on many tasks in natural language processing.
Approach:	They propose to automatically generate an NLI dataset with an LLM and use it for fine-tuning of PromptEOL.
Outcome:	The proposed model outperforms existing models on STS tasks without large manually annotated datasets.

Copied to clipboard

Kosuke Yamada, Tsutomu Hirao, Ryohei Sasano, Koichi Takeda, Masaaki Nagata

Text Classification

Challenge:	Existing methods for dividing biomedical abstracts into rhetorical segments assign a rhetorical label to each sentence while considering context in the abstract.
Approach:	They propose to use Neural Semi-Markov Conditional Random Fields to assign a rhetorical label to a span that consists of continuous sentences.
Outcome:	The proposed method achieved the best micro sentence-F1 score and the best macro span-F1.

How Do Language Models Acquire Character-Level Information? (2026.eacl-long)

Copied to clipboard

Soma Sato, Ryohei Sasano

Redundancy, Isotropy, and Intrinsic Dimensionality of Prompt-based Text Embeddings (2025.findings-acl)

Challenge:	Language models (LMs) implicitly encode character-level information, despite not being explicitly provided during training.
Approach:	They analyze how language models acquire character-level knowledge by comparing them to standard settings.
Outcome:	The results show that LMs do not treat words as opaque tokens, but instead treat them as tokens.

Copied to clipboard

Hayato Tsukagoshi, Ryohei Sasano

Leveraging Three Types of Embeddings from Masked Language Models in Idiom Token Classification (2022.starsem-1)

Challenge:	Prompt-based text embedding models generate task-specific embeddables but have thousands of dimensions . dimensionality reductions for embedded text can result in performance degradations of only the first 25% of the dimensions resulting in a very small degradation .
Approach:	They investigate how post-hoc dimensionality reduction affects performance of various tasks . they find that embeddings for classification and clustering exhibit lower intrinsic dimensionalities .
Outcome:	The proposed model generates task-specific embeddings upon receiving tailored prompts, but has thousands of dimensions and high storage costs.

Copied to clipboard

Ryosuke Takahashi, Ryohei Sasano, Koichi Takeda

DefSent: Sentence Embeddings using Definition Sentences (2021.acl-short)

Challenge:	Recent research shows that contextualized word embeddings can give promising results for idiom token classification.
Approach:	They propose to leverage contextualized word embeddings from masked language models to improve idiom token classification.
Outcome:	The proposed method improves idiom token classification for English and Japanese datasets.

Copied to clipboard

Hayato Tsukagoshi, Ryohei Sasano, Koichi Takeda

To Drop or Not to Drop? Predicting Argument Ellipsis Judgments: A Case Study in Japanese (2024.lrec-main)

Challenge:	Sentence embedding methods using natural language inference datasets are limited for limited languages due to large datasets.
Approach:	They propose a sentence embedding method that uses definition sentences from a word dictionary.
Outcome:	The proposed method performs comparably on unsupervised semantics textual similarity tasks and slightly better on SentEval tasks than methods using large NLI datasets.

Copied to clipboard

Yukiko Ishizuki, Tatsuki Kuribayashi, Yuichiroh Matsubayashi, Ryohei Sasano, Kentaro Inui

Challenge:	Speakers sometimes omit certain arguments of a predicate in a sentence; such omission is especially frequent in pro-drop languages.
Approach:	They collect large-scale human annotations of whether and why a particular argument should be omitted across over 2,000 data points in Japanese, a prototypical pro-drop language.
Outcome:	The proposed model can explain why certain arguments are omitted in Japanese, a prototypical pro-drop language.

Sentence Representations via Gaussian Embedding (2024.eacl-short)

Copied to clipboard

Shohei Yoda, Hayato Tsukagoshi, Ryohei Sasano, Koichi Takeda

Representation Learning

Challenge:	Sentence embeddings represent a sentence's meaning as a point in a vector space and primarily use symmetric measures such as the cosine similarity to measure the similarity between sentences, they cannot capture asymmetric relationships between two sentences, such as entailment and hierarchical relations.
Approach:	They propose a Gaussian-distribution-based contrastive learning framework for sentence embedding that can handle asymmetric inter-sentential relations and a similarity measure for identifying entailment relations.
Outcome:	The proposed framework performs comparable to that of previous methods on natural language inference tasks and estimates direction of entailment relations, which is difficult with point representations.

Verb Sense Clustering using Contextualized Word Representations for Semantic Frame Induction (2021.findings-acl)

Copied to clipboard

Definition Generation for Automatically Induced Semantic Frame (2024.findings-acl)

Text Clustering

Challenge:	Contextualized word representations are effective in many natural language processing tasks, but it remains unclear to what extent they can cover hand-coded semantic information such as semantic frames.
Approach:	They compare contextualized word representations with two English frame-semantic resources . they find that several contextualized representations are informative for semantic frame induction .
Outcome:	The proposed representations are useful in natural language processing tasks, but are not fully understood by the literature.

Copied to clipboard

Yi Han, Ryohei Sasano, Koichi Takeda

Semantic Frame Induction with Deep Metric Learning (2023.eacl-main)

Challenge:	Semantic frames are conceptual structures that describe specific types of situations or events.
Approach:	They propose to generate frame definitions from a set of frame-evoking words using a large language model.
Outcome:	The proposed task incorporates frame element reasoning as chain-of-thought to enhance the inclusion of correct frame elements in the generated definitions.

Copied to clipboard

Transformer-based Lexically Constrained Headline Generation (2021.emnlp-main)

Challenge:	Recent studies have shown the usefulness of contextualized word embeddings in semantic frame induction, but they are not always consistent with human intuitions about semantic frames.
Approach:	They propose a model that fine-tunes contextualized embeddings to perform semantic frame induction.
Outcome:	The proposed model improves clustering evaluation scores on FrameNet by 8 points or more.

Copied to clipboard

Kosuke Yamada, Yuta Hitomi, Hideaki Tamori, Ryohei Sasano, Naoaki Okazaki, Kentaro Inui, Koichi Takeda

Incorporating Textual Information on User Behavior for Personality Prediction (P19-2)

Challenge:	Existing automatic headline generation methods cannot include a given phrase in the generated headline.
Approach:	They propose a Transformer-based method that guarantees to include a given phrase in a generated headline.
Outcome:	The proposed method achieves ROUGE scores comparable to previous methods with Japanese news corpus.

Copied to clipboard

Development of a Medical Incident Report Corpus with Intention and Factuality Annotation (2020.lrec-1)

Sentiment Analysis

Challenge:	Recent studies have shown that textual information of user posts and user behaviors are useful for predicting the personality of social media users.
Approach:	They propose to use textual information of user behaviors to predict personality of Twitter users by taking user behaviors into account.
Outcome:	The proposed models can predict personality of users who do not post frequently, while taking user behaviors into account.

Copied to clipboard

Hongkuan Zhang, Ryohei Sasano, Koichi Takeda, Zoie Shui-Yee Wong

Challenge:	Medical incident reports are documents that record what happened in a medical incident.
Approach:	They propose to annotate medical incident reports with annotations of intention and factuality and medication entities and their relations.
Outcome:	The proposed method combines the definition of medication entities and the method to annotate the relations between entities and extracts important information from the unstructured part.

Simplifying Translations for Children: Iterative Simplification Considering Age of Acquisition with LLMs (2024.findings-acl)

Copied to clipboard

Masashi Oshika, Makoto Morishita, Tsutomu Hirao, Ryohei Sasano, Koichi Takeda

Text Generation

Challenge:	Neural machine translation (NMT) systems do not take into account the complexity of the words used to compose the translations.
Approach:	They propose a method that replaces high Age of Acquisitions words in translations with simpler words to match the user’s level.
Outcome:	The proposed method replaces high-AoA words with lower-Aa words while maintaining high BLEU and COMET scores.

Realistic Citation Count Prediction Task for Newly Published Papers (2023.findings-eacl)

Copied to clipboard

Jun Hirako, Ryohei Sasano, Koichi Takeda

Challenge:	Existing studies on citation count prediction assume that future citation counts of academic papers have not had enough time pass since publication.
Approach:	They propose to use citation counts of newly published papers as a realistic citation count prediction task and to use them to leverage the citations of papers shortly after publication.
Outcome:	The proposed methods significantly improve the performance of citation count prediction for newly published papers in a realistic setting.

Semantic Frame Induction using Masked Word Embeddings and Two-Step Clustering (2021.acl-short)

Copied to clipboard