arXiv subjects

Takahiro Omi

Publications and source records attributed to Takahiro Omi.

9 recordsLinked to original sources

arXiv2026

Aggregate vs. Personalized Judges in Business Idea Evaluation: Evidence from Expert Disagreement

Evaluating LLM-generated business ideas is often harder to scale than generating them. Unlike standard NLP benchmarks, business idea evaluation relies on multi-dimensional criteria such as feasibility, novelty, differentiation, user need, and market size, and expert judgments often disagree. This paper studies a methodological question raised by such disagreement: should an automatic judge approximate an aggregate consensus, or model evaluators individually? We introduce PBIG-DATA, a dataset of approximately 3,000 individual scores across 300 patent-grounded product ideas, provided by domain experts on six business-oriented dimensions: specificity, technical validity, innovativeness, competitive advantage, need validity, and market size. Analyses show substantial expert disagreement on fine-grained ordinal scores, while agreement is higher under coarse selection, suggesting structured heterogeneity rather than random noise. We then compare three judge configurations: a rubric-only zero-shot judge, an aggregate judge conditioned on mixed evaluator histories, and a personalized judge conditioned on the target evaluator's scoring history. Across dimensions and model sizes, personalized judges align more closely with the corresponding evaluator than aggregate judges, and evaluator agreement correlates with similarity of judge-generated reasoning only under personalized conditioning. These results indicate that pooled labels can be a fragile target in pluralistic evaluation settings and motivate evaluator-conditioned judge designs for business idea assessment.

cs.CL↗

arXiv2025

Exploring Design of Multi-Agent LLM Dialogues for Research Ideation

Large language models (LLMs) are increasingly used to support creative tasks such as research idea generation. While recent work has shown that structured dialogues between LLMs can improve the novelty and feasibility of generated ideas, the optimal design of such interactions remains unclear. In this study, we conduct a comprehensive analysis of multi-agent LLM dialogues for scientific ideation. We compare different configurations of agent roles, number of agents, and dialogue depth to understand how these factors influence the novelty and feasibility of generated ideas. Our experimental setup includes settings where one agent generates ideas and another critiques them, enabling iterative improvement. Our results show that enlarging the agent cohort, deepening the interaction depth, and broadening agent persona heterogeneity each enrich the diversity of generated ideas. Moreover, specifically increasing critic-side diversity within the ideation-critique-revision loop further boosts the feasibility of the final proposals. Our findings offer practical guidelines for building effective multi-agent LLM systems for scientific ideation. Our code is available at https://github.com/g6000/MultiAgent-Research-Ideator.

cs.CL↗

arXiv2025

Enhancing Large Vision-Language Models with Layout Modality for Table Question Answering on Japanese Annual Securities Reports

With recent advancements in Large Language Models (LLMs) and growing interest in retrieval-augmented generation (RAG), the ability to understand table structures has become increasingly important. This is especially critical in financial domains such as securities reports, where highly accurate question answering (QA) over tables is required. However, tables exist in various formats-including HTML, images, and plain text-making it difficult to preserve and extract structural information. Therefore, multimodal LLMs are essential for robust and general-purpose table understanding. Despite their promise, current Large Vision-Language Models (LVLMs), which are major representatives of multimodal LLMs, still face challenges in accurately understanding characters and their spatial relationships within documents. In this study, we propose a method to enhance LVLM-based table understanding by incorporating in-table textual content and layout features. Experimental results demonstrate that these auxiliary modalities significantly improve performance, enabling robust interpretation of complex document layouts without relying on explicitly structured input formats.

cs.CL↗

arXiv2024

Pretraining and Updates of Domain-Specific LLM: A Case Study in the Japanese Business Domain

The development of Large Language Models (LLMs) in various languages has been advancing, but the combination of non-English languages with domain-specific contexts remains underexplored. This paper presents our findings from training and evaluating a Japanese business domain-specific LLM designed to better understand business-related documents, such as the news on current affairs, technical reports, and patents. Additionally, LLMs in this domain require regular updates to incorporate the most recent knowledge. Therefore, we also report our findings from the first experiments and evaluations involving updates to this LLM using the latest article data, which is an important problem setting that has not been addressed in previous research. From our experiments on a newly created benchmark dataset for question answering in the target domain, we found that (1) our pretrained model improves QA accuracy without losing general knowledge, and (2) a proper mixture of the latest and older texts in the training data for the update is necessary. Our pretrained model and business domain benchmark are publicly available to support further studies.

cs.CL↗

arXiv2023

Training Generative Question-Answering on Synthetic Data Obtained from an Instruct-tuned Model

This paper presents a simple and cost-effective method for synthesizing data to train question-answering systems. For training, fine-tuning GPT models is a common practice in resource-rich languages like English, however, it becomes challenging for non-English languages due to the scarcity of sufficient question-answer (QA) pairs. Existing approaches use question and answer generators trained on human-authored QA pairs, which involves substantial human expenses. In contrast, we use an instruct-tuned model to generate QA pairs in a zero-shot or few-shot manner. We conduct experiments to compare various strategies for obtaining QA pairs from the instruct-tuned model. The results demonstrate that a model trained on our proposed synthetic data achieves comparable performance to a model trained on manually curated datasets, without incurring human costs.

cs.CL↗

arXiv2020

Fully Neural Network based Model for General Temporal Point Processes

A temporal point process is a mathematical model for a time series of discrete events, which covers various applications. Recently, recurrent neural network (RNN) based models have been developed for point processes and have been found effective. RNN based models usually assume a specific functional form for the time course of the intensity function of a point process (e.g., exponentially decreasing or increasing with the time since the most recent event). However, such an assumption can restrict the expressive power of the model. We herein propose a novel RNN based model in which the time course of the intensity function is represented in a general manner. In our approach, we first model the integral of the intensity function using a feedforward neural network and then obtain the intensity function as its derivative. This approach enables us to both obtain a flexible model of the intensity function and exactly evaluate the log-likelihood function, which contains the integral of the intensity function, without any numerical approximations. Our model achieves competitive or superior performances compared to the previous state-of-the-art methods for both synthetic and real datasets.

cs.LG↗

arXiv2017

Hawkes process model with a time-dependent background rate and its application to high-frequency financial data

A Hawkes process model with a time-varying background rate is developed for analyzing the high-frequency financial data. In our model, the logarithm of the background rate is modeled by a linear model with a relatively large number of variable-width basis functions, and the parameters are estimated by a Bayesian method. Our model can capture not only the slow time-variation, such as in the intraday seasonality, but also the rapid one, which follows a macroeconomic news announcement. By analyzing the tick data of the Nikkei 225 mini, we find that (i) our model is better fitted to the data than the Hawkes models with a constant background rate or a slowly varying background rate, which have been commonly used in the field of quantitative finance; (ii) the improvement in the goodness-of-fit to the data by our model is significant especially for sessions where considerable fluctuation of the background rate is present; and (iii) our model is statistically consistent with the data. The branching ratio, which quantifies the level of the endogeneity of markets, estimated by our model is 0.41, suggesting the relative importance of exogenous factors in the market dynamics. We also demonstrate that it is critically important to appropriately model the time-dependent background rate for the branching ratio estimation.

q-fin.ST↗

arXiv2007

Can distributed delays perfectly stabilize dynamical networks?

Signal transmission delays tend to destabilize dynamical networks leading to oscillation, but their dispersion contributes oppositely toward stabilization. We analyze an integro-differential equation that describes the collective dynamics of a neural network with distributed signal delays. With the gamma distributed delays less dispersed than exponential distribution, the system exhibits reentrant phenomena, in which the stability is once lost but then recovered as the mean delay is increased. With delays dispersed more highly than exponential, the system never destabilizes.

cond-mat.dis-nn↗

arXiv2007

Reverberating activity in a neural network with distributed signal transmission delays

It is known that an identical delay in all transmission lines can destabilize macroscopic stationarity of a neural network, causing oscillation or chaos. We analyze the collective dynamics of a network whose intra-transmission delays are distributed in time. Here, a neuron is modeled as a discrete-time threshold element that responds in an all-or-nothing manner to a linear sum of signals that arrive after delays assigned to individual transmission lines. Even though transmission delays are distributed in time, a whole network exhibits a single collective oscillation with a period close to the average transmission delay. The collective oscillation can not only be a simple alternation of the consecutive firing and resting, but also nontrivially sequenced series of firing and resting, reverberating in a certain period of time. Moreover, the system dynamics can be made quasiperiodic or chaotic by changing the distribution of delays.

cond-mat.dis-nn↗