arXiv subjects

Shogo Okada

Publications and source records attributed to Shogo Okada.

15 recordsLinked to original sources

arXiv2026

Cross-lingual Functional Vectors for Emotion Detection in Large Language Models

Function vectors (FVs) have recently emerged as a promising mechanism for steering the behavior of large language models (LLMs) by injecting task-specific latent direction representations derived from in-context demonstrations. While prior studies have shown that FVs can recover task behavior in structured in-context learning settings, their effectiveness on semantically complex tasks and their ability to generalize across languages remain underexplored. We investigate the cross-lingual transferability of FVs using multilingual multi-label emotion recognition as a challenging semantic classification benchmark. Specifically, we examine whether FVs extracted from a source language can steer task behavior in another language under both standard clean and perturbed zero-shot settings without providing demonstrations during inference. Across diverse cross-lingual settings, applying FVs substantially improves performance, suggesting that FVs capture language-agnostic, task-relevant signals rather than purely language-specific lexical patterns, and highlighting their potential as a lightweight and transferable mechanism for multilingual task adaptation. We observe that each LLM exhibits a relatively stable optimal range of attention heads for constructing effective FVs, and the pattern remains consistent across languages. In addition, FVs can partially replicate the task-steering effects of standard few-shot in-context learning while avoiding the computational overhead of processing multiple demonstrations, making them effective for large-scale practical applications. Our code is available at https://github.com/yingjie7/cross_lingual_fvs.

cs.CL↗

arXiv2026

Multimodal Rapport Estimation in Real-World HRI

Evaluating interaction quality in real-world HRI is an important challenge. If interaction quality can be estimated reliably, the results can be used to improve dialogue strategies and ultimately enable robots to adapt their behavior autonomously. However, existing automatic evaluation methods have been developed primarily in controlled laboratory settings, and it remains unclear whether they can be directly applied to real-world environments, where users are free to disengage and multi-party participation may arise naturally. In this study, we investigate the automatic estimation of third-party-rated rapport scores using 62 sessions of multimodal recordings collected in a Japanese drugstore. We compare zero-shot LLMs, pretrained text, audio, and visual models, and their prediction-level fusion. The results show that, in real-world HRI, zero-shot LLMs achieve strong performance, while audio and visual models tend to provide complementary information. In particular, Gemini 2.5 Flash performs strongly as a single model, and a fusion model combining Gemini (text) with HuBERT and V-JEPA performs best overall. Further analyses showed that estimation performance varied across interaction-duration and group-size conditions. These findings suggest that rapport estimation in real-world HRI requires evaluation and model design that account for contextual variability beyond that assumed in laboratory settings.

cs.HC↗

arXiv2026

Speech Signals Complement LLMs for Predicting Interpersonal Attraction in Speed Dating

Large language models (LLMs) can predict interpersonal attraction from conversation transcripts, but it remains unclear what a speech predictor can add beyond transcript-only LLM prediction. Using Japanese speed-dating conversations, we combine predictions from a transcript-only LLM and a supervised speech predictor to estimate participants' reported liking of their partners. We show that speech can complement transcript-only LLM prediction, but that this complementarity is conditional rather than universal. Combining the two predictions significantly improves pairwise ranking accuracy over the transcript-only LLM alone in all evaluated conditions. By contrast, gains in per-participant Pearson $r$ vary across conversation rounds and rating directions, with none significant after correction. Retrospectively, these $r$ gains are concentrated among participants for whom the speech predictor is more accurate. Speech can therefore retain predictive value even when an LLM predicts attraction from transcripts. The relevant question is not simply whether speech helps, but where its complementarity emerges.

cs.CL↗

arXiv2026

XPASS-Vis: A Dataset for Cross-Domain Personalized Image Aesthetic Assessment

Personalized image aesthetic assessment (PIAA) seeks to model, at the individual level, the subjective nature of aesthetic judgments toward artworks and photographs. Aesthetic preference is known to be both deeply personal and partially consistent across visual domains. Yet existing PIAA datasets and methods are largely confined to a single domain, or provide too few samples per annotator within each domain to enable personalization across domains. Consequently, the cross-domain generalization of personalized aesthetic preferences remains largely unexplored. To address this gap, we introduce XPASS-Vis, the first dataset explicitly designed for cross-domain PIAA. XPASS-Vis comprises 6,526 stimuli from three visual domains -- art, fashion, and landscape -- rated by 129 annotators, yielding 87,836 user-stimulus interactions, each annotated with an overall aesthetic score and nine aesthetic-emotion ratings. Notably, each annotator rated more than 200 stimuli per domain, providing sufficient per-domain coverage to support personalization both within and across domains. Moreover, we establish baseline models for cross-domain PIAA under unsupervised domain adaptation (UDA), where a model trained on a labeled source domain is transferred to an unlabeled target domain. A systematic evaluation of representative UDA approaches shows that the best-performing method recovers approximately 60\% (Spearman's $\rho$ = .28) of the supervised upper bound under a fully unsupervised setting. This provides encouraging evidence that personalized aesthetic preferences are, to a meaningful extent, transferable across visual domains. At the same time, a substantial gap remains, highlighting the need for PIAA-specific adaptation strategies. XPASS-Vis and the accompanying baselines provide a foundation for future research on cross-domain PIAA. All datasets and code will be made publicly available upon acceptance.

cs.CV↗

arXiv2025

Technical Report of Nomi Team in the Environmental Sound Deepfake Detection Challenge 2026

This paper presents our work for the ICASSP 2026 Environmental Sound Deepfake Detection (ESDD) Challenge. The challenge is based on the large-scale EnvSDD dataset that consists of various synthetic environmental sounds. We focus on addressing the complexities of unseen generators and low-resource black-box scenarios by proposing an audio-text cross-attention model. Experiments with individual and combined text-audio models demonstrate competitive EER improvements over the challenge baseline (BEATs+AASIST model).

cs.SD↗

arXiv2024

WorldCuisines: A Massive-Scale Benchmark for Multilingual and Multicultural Visual Question Answering on Global Cuisines

Vision Language Models (VLMs) often struggle with culture-specific knowledge, particularly in languages other than English and in underrepresented cultural contexts. To evaluate their understanding of such knowledge, we introduce WorldCuisines, a massive-scale benchmark for multilingual and multicultural, visually grounded language understanding. This benchmark includes a visual question answering (VQA) dataset with text-image pairs across 30 languages and dialects, spanning 9 language families and featuring over 1 million data points, making it the largest multicultural VQA benchmark to date. It includes tasks for identifying dish names and their origins. We provide evaluation datasets in two sizes (12k and 60k instances) alongside a training dataset (1 million instances). Our findings show that while VLMs perform better with correct location context, they struggle with adversarial contexts and predicting specific regional cuisines and languages. To support future research, we release a knowledge base with annotated food entries and images along with the VQA data.

cs.CL↗

arXiv2024

Investigating Role of Big Five Personality Traits in Audio-Visual Rapport Estimation

Automatic rapport estimation in social interactions is a central component of affective computing. Recent reports have shown that the estimation performance of rapport in initial interactions can be improved by using the participant's personality traits as the model's input. In this study, we investigate whether this findings applies to interactions between friends by developing rapport estimation models that utilize nonverbal cues (audio and facial expressions) as inputs. Our experimental results show that adding Big Five features (BFFs) to nonverbal features can improve the estimation performance of self-reported rapport in dyadic interactions between friends. Next, we demystify how BFFs improve the estimation performance of rapport through a comparative analysis between models with and without BFFs. We decompose rapport ratings into perceiver effects (people's tendency to rate other people), target effects (people's tendency to be rated by other people), and relationship effects (people's unique ratings for a specific person) using the social relations model. We then analyze the extent to which BFFs contribute to capturing each effect. Our analysis demonstrates that the perceiver's and the target's BFFs lead estimation models to capture the perceiver and the target effects, respectively. Furthermore, our experimental results indicate that the combinations of facial expression features and BFFs achieve best estimation performances not only in estimating rapport ratings, but also in estimating three effects. Our study is the first step toward understanding why personality-aware estimation models of interpersonal perception accomplish high estimation performance.

cs.HC↗

arXiv2024

Multilingual Dyadic Interaction Corpus NoXi+J: Toward Understanding Asian-European Non-verbal Cultural Characteristics and their Influences on Engagement

Non-verbal behavior is a central challenge in understanding the dynamics of a conversation and the affective states between interlocutors arising from the interaction. Although psychological research has demonstrated that non-verbal behaviors vary across cultures, limited computational analysis has been conducted to clarify these differences and assess their impact on engagement recognition. To gain a greater understanding of engagement and non-verbal behaviors among a wide range of cultures and language spheres, in this study we conduct a multilingual computational analysis of non-verbal features and investigate their role in engagement and engagement prediction. To achieve this goal, we first expanded the NoXi dataset, which contains interaction data from participants living in France, Germany, and the United Kingdom, by collecting session data of dyadic conversations in Japanese and Chinese, resulting in the enhanced dataset NoXi+J. Next, we extracted multimodal non-verbal features, including speech acoustics, facial expressions, backchanneling and gestures, via various pattern recognition techniques and algorithms. Then, we conducted a statistical analysis of listening behaviors and backchannel patterns to identify culturally dependent and independent features in each language and common features among multiple languages. These features were also correlated with the engagement shown by the interlocutors. Finally, we analyzed the influence of cultural differences in the input features of LSTM models trained to predict engagement for five language datasets. A SHAP analysis combined with transfer learning confirmed a considerable correlation between the importance of input features for a language set and the significant cultural characteristics analyzed.

cs.CL↗

arXiv2024

BERP: A Blind Estimator of Room Parameters for Single-Channel Noisy Speech Signals

Room acoustical parameters (RAPs), room geometrical parameters (RGPs) and instantaneous occupancy level are essential metrics for parameterizing the room acoustical characteristics (RACs) of a sound field around a listener's local environment, offering comprehensive indications for various applications. Current blind estimation methods either fail to cover a broad range of real-world acoustic environments in the context of real background noise or estimate only a few RAPs and RGPs from noisy single-channel speech signals. In addition, they are limited in their ability to estimate the instantaneous occupancy level. In this paper, we propose a new universal blind estimation framework called the blind estimator of room parameters (BERP) to estimate RAPs, RGPs and occupancy level via a unified methodology. It consists of two modules: a unified room feature encoder that combines attention mechanisms with convolutional layers to learn common features across room parameters, and multiple separate parametric predictors for continuous estimation of each parameter in parallel. The combination of attention and convolutions enables the model to capture acoustic features locally and globally from speech, yielding more robust and multitask generalizable common features. Separate predictors allow the model to independently optimize for each room parameter to reduce task learning conflict and improve per-task performance. This estimation framework enables universal and efficient estimation of room parameters while maintaining satisfactory performance. To evaluate the effectiveness of the proposed framework, we compile a task-specific dataset from several publicly available datasets, including synthetic and real reverberant recordings. The results reveal that BERP achieves state-of-the-art (SOTA) performance and excellent adaptability to real-world scenarios. The code and weights are available on GitHub.

eess.AS↗

arXiv2023

FedCPC: An Effective Federated Contrastive Learning Method for Privacy Preserving Early-Stage Alzheimer's Speech Detection

The early-stage Alzheimer's disease (AD) detection has been considered an important field of medical studies. Like traditional machine learning methods, speech-based automatic detection also suffers from data privacy risks because the data of specific patients are exclusive to each medical institution. A common practice is to use federated learning to protect the patients' data privacy. However, its distributed learning process also causes performance reduction. To alleviate this problem while protecting user privacy, we propose a federated contrastive pre-training (FedCPC) performed before federated training for AD speech detection, which can learn a better representation from raw data and enables different clients to share data in the pre-training and training stages. Experimental results demonstrate that the proposed methods can achieve satisfactory performance while preserving data privacy.

eess.AS↗

arXiv2023

Estimating Driver Personality Traits from On-Road Driving Data

This paper focuses on the estimation of a driver's psychological characteristics using driving data for driving assistance systems. Driving assistance systems that support drivers by adapting individual psychological characteristics can provide appropriate feedback and prevent traffic accidents. As a first step toward implementing such adaptive assistance systems, this research aims to develop a model to estimate drivers' psychological characteristics, such as cognitive function, psychological driving style, and workload sensitivity, from on-road driving behavioral data using machine learning and deep learning techniques. We also investigated the relationship between driving behavior and various cognitive functions, including the Trail Making Test (TMT) and Useful Field of View (UFOV) test, through regression modeling. The proposed method focuses on road type information and captures various durations of time-series data observed from driving behaviors. First, we segment the driving time-series data into two road types, namely, arterial roads and intersections, to consider driving situations. Second, we further segment data into many sequences of various durations. Third, statistics are calculated from each sequence. Finally, these statistics are used as input features of machine learning models to estimate psychological characteristics. The experimental results show that our model can estimate a driver's cognitive function, namely, the TMT~(B) and UFOV test scores, with Pearson correlation coefficients $r$ of 0.579 and 0.708, respectively. Some characteristics, such as psychological driving style and workload sensitivity, are estimated with high accuracy, but whether various duration segmentation improves accuracy depends on the characteristics, and it is not effective for all characteristics.

cs.LG↗

arXiv2023

Interpretable multimodal sentiment analysis based on textual modality descriptions by using large-scale language models

Multimodal sentiment analysis is an important area for understanding the user's internal states. Deep learning methods were effective, but the problem of poor interpretability has gradually gained attention. Previous works have attempted to use attention weights or vector distributions to provide interpretability. However, their explanations were not intuitive and can be influenced by different trained models. This study proposed a novel approach to provide interpretability by converting nonverbal modalities into text descriptions and by using large-scale language models for sentiment predictions. This provides an intuitive approach to directly interpret what models depend on with respect to making decisions from input texts, thus significantly improving interpretability. Specifically, we convert descriptions based on two feature patterns for the audio modality and discrete action units for the facial modality. Experimental results on two sentiment analysis tasks demonstrated that the proposed approach maintained, or even improved effectiveness for sentiment analysis compared to baselines using conventional features, with the highest improvement of 2.49% on the F1 score. The results also showed that multimodal descriptions have similar characteristics on fusing modalities as those of conventional fusion methods. The results demonstrated that the proposed approach is interpretable and effective for multimodal sentiment analysis.

cs.CL↗

arXiv2022

Personality-adapted multimodal dialogue system

This paper describes a personality-adaptive multimodal dialogue system developed for the Dialogue Robot Competition 2022. To realize a dialogue system that adapts the dialogue strategy to individual users, it is necessary to consider the user's nonverbal information and personality. In this competition, we built a prototype of a user-adaptive dialogue system that estimates user personality during dialogue. Pretrained DNN models are used to estimate user personalities annotated as Big Five scores. This model is embedded in a dialogue system to estimate user personality from face images during the dialogue. We proposed a method for dialogue management that changed the dialogue flow based on the estimated personality characteristics and confirmed that the system works in a real environment in the preliminary round of this competition. Furthermore, we implemented specific modules to enhance the multimodal dialogue experience of the user, including personality assessment, controlling facial expressions and movements of the android, and dialogue management to explain the attractiveness of sightseeing spots. The aim of dialogue based on personality assessment is to reduce the nervousness of users, and it acts as an ice breaker. The android's facial expressions and movements are necessary for a more natural android conversation. Since the task of this competition was to promote the appeal of sightseeing spots and to recommend an appropriate sightseeing spot, the dialogue process for how to explain the attractiveness of the spot is important. All results of the subjective evaluation by users were better than those of the baseline and other systems developed for this competition. The proposed dialogue system ranked first in both "Impression Rating" and "Effectiveness of Android Recommendations". According to the total evaluation in the competition, the proposed system was ranked first overall.

cs.HC↗

arXiv2022

IDPS Signature Classification with a Reject Option and the Incorporation of Expert Knowledge

As the importance of intrusion detection and prevention systems (IDPSs) increases, great costs are incurred to manage the signatures that are generated by malicious communication pattern files. Experts in network security need to classify signatures by importance for an IDPS to work. We propose and evaluate a machine learning signature classification model with a reject option (RO) to reduce the cost of setting up an IDPS. To train the proposed model, it is essential to design features that are effective for signature classification. Experts classify signatures with predefined if-then rules. An if-then rule returns a label of low, medium, high, or unknown importance based on keyword matching of the elements in the signature. Therefore, we first design two types of features, symbolic features (SFs) and keyword features (KFs), which are used in keyword matching for the if-then rules. Next, we design web information and message features (WMFs) to capture the properties of signatures that do not match the if-then rules. The WMFs are extracted as term frequency-inverse document frequency (TF-IDF) features of the message text in the signatures. The features are obtained by web scraping from the referenced external attack identification systems described in the signature. Because failure needs to be minimized in the classification of IDPS signatures, as in the medical field, we consider introducing a RO in our proposed model. The effectiveness of the proposed classification model is evaluated in experiments with two real datasets composed of signatures labeled by experts: a dataset that can be classified with if-then rules and a dataset with elements that do not match an if-then rule. In the experiment, the proposed model is evaluated. In both cases, the combined SFs and WMFs performed better than the combined SFs and KFs. In addition, we also performed feature analysis.

cs.CR↗

arXiv2008

Inclusive transverse mass analysis for squark and gluino mass determination

We propose an inclusive analysis of a stransverse mass (m_{T2}) using a hemisphere method for supersymmetry studies at the LHC . The hemisphere method is an algorithm to group collinear and high p_T particles and jets, assuming that there are two of such groups in a event. The m_{T2} is defined as a function of the unknown LSP mass, two hemisphere momenta, and missing transverse momentum. The kinematical end point of the m_{T2} distribution provides information on the squark and gluino masses. We perform a Monte Carlo simulation to study the inclusive m_{T2} distribution at the LHC. We show that the end point of the inclusive m_{T2} distribution has a cusp-like structure around the true LSP mass. The knowledge of the expected kinematical behavior near the end point for true events is important to establish the end point of the inclusive distribution. We find that the inclusive analysis is useful to obtain the information on the heaviest of the squark/gluino.

hep-ph↗