This paper proposes a system that supports watching Mahjong games for novices. Two factors make watching Mahjong games difficult for novices: it is difficult to know what combination of tiles the players are aiming for, and it is difficult to instantly grasp what tiles each player has and what tiles they discard. The system predicts and visualizes the combination of tiles at the goal that each player is aiming for and also visualizes the status of each player’s holding and discarded tiles side by side in the horizontal line. In the evaluation experiment, participants watched Mahjong games through the proposed system. Utterances during the experiment were analyzed based on Think-aloud protocols. For comparison, we prepared a comparison system that reproduced a typical Mahjong game, had the participants watch the game similarly, and analyzed the content of their speech. Participants watched two games, and the proposed system increased utterances indicating an understanding of the game’s situation in the second time more than that in the first time. This indicates that the proposed system can help the spectator understand the game’s contents and watch the Mahjong game.
In recent years, the development of large-scale language models (LLM) has dramatically improved text generation performance. However, general-purpose LLMs have the problem that they do not always perform optimally in tasks in specific fields. The approach of fine-tuning with data from specific fields is commonly used to address this problem, but collecting training data from limited fields is difficult. In particular, data related to human values have problems such as difficulty in annotation, insufficient amount of data, and large variability. In addition, the training and inference of LLMs is expensive. This study focused on the task of human values estimation and tested the effectiveness of an approach that uses an LLM to generate pseudo-data and trains a small-scale model on that data. In the experiment, we augmented a 3,870-items human values dataset with 10 categories to four times its original size with the generated pseudo-data. The training dataset with pseudo-data increased human values estimation accuracy by 17
Important bug report comments can help developers efficiently fix bugs. This paper proposes a bug report comment ranking model called BRCR based on deep learning models. This paper also proposes a pseudo-labeling approach to address the few-shot problem. In the experiments, bug reports of three GitHub software projects are used to evaluate the proposed BRCR ranking model. The experimental results show that BRCR with LSTM and pseudo-labeled data can effectively outperform other ranking models.
This paper proposes a new evaluation index for evaluating human fallibility in shogi. By analyzing the differences between the decision-making processes of AI and humans, we found that while humans rely on intuition and decisions made within a limited timeframe, AI often makes different decisions because it carries out exhaustive searches. In this study, we developed a new evaluation index that extracts features prone to human error using a policy network trained based on professional game records. This evaluation model uses logistic regression to predict the probability of making a mistake in the endgame. We have shown that the proposed indicator is effective through testing with test data. In the future, we plan to construct similar evaluation indicators for the opening and middle game phases and verify their effectiveness.
Given the surge in digital communication, the spread of fraud messages across various communication platforms poses an increasingly significant threat to the public due to their accessibility and ubiquity. Existing fraud detection methodologies predominantly focus on recognized fraud tactics, thus diminishing their effectiveness against unfamiliar or evolving deceptive strategies, a challenge known as domain adaptation. To address this issue, we introduce a novel approach, Graph-Based Syntactic Structure Generation, capable of autonomously learning common syntactic structures within fraud messages without the need for predefined fraud patterns. This method leverages co-occurrence graphs and graph auto-encoder techniques to extract connection state features of each token from extensive texts, thereby acquiring global information essential for identifying complex syntactic structures indicative of fraud messages. Further, we employ an attention-based model for syntactic structure transformation of texts, converting each token into word, part of speech (POS), or # (fixed valued) to facilitate effective syntactic pattern recognition. This approach not only enhances the model’s understanding of textual syntactic structures but also enriches semantic information for subsequent large language model analysis. Experiments conducted on datasets sourced from Cofacts and the 165 anti-fraud platform demonstrate our method’s robust recognition performance on cross-domain datasets, even with limited training data, surpassing traditional methods’ limitations when confronting newly emerged fraud tactics.
The Industrial Internet of Things (IIoT) is crucial for smart manufacturing and automation but is highly vulnerable to cyber-attacks due to inadequate security mechanisms and diverse communication protocols. Attackers can exploit weak IIoT devices to infiltrate entire networks, posing significant challenges in detecting malicious traffic within heterogeneous IIoT environments. This study introduces a hybrid model, composed of BERT and LSTM network for anomaly detection in IIoT networks. Given that the effectiveness of such models depends on the quality of the training data, this research investigates the effects of data imbalance and granularity, offering solutions to these issues. Experimental results show that the proposed method effectively identifies malicious traffic and outperforms other reference models in detection efficiency.
Social media platform frequently experiences incidents where accounts are compromised. These incidents involve attempts to misappropriate personal information, deceive others into trusting fraudulent URLs, or manipulate public sentiment by exploiting stolen accounts. Little research has been done on compromised account detection. Some approaches adopt an incremental approach at message level by comparing incoming messages against established normal behavior model for detection. The other approach utilizes supervised algorithms at the account level by binary classification into normal and compromised accounts. This paper aims to investigate compromised account detection over social media. We proposed the segmented-based supervised approach by leveraging time series change point detection from statistical community. The change point detection algorithm attempts to discover the suspicious compromised intervals. Based on the change point detection, a new feature, change point feature is proposed to capture user’s behavior changing pattern. Moreover, the interest feature based on a user’s activities in his/her interested groups and the polarity feature that captures the user’s polarity based on like/dislike behaviors are proposed. Experiments conducted on both real and synthetic datasets, demonstrate the superior accuracy of the proposed method.
Patents are essential for securing market leadership and protecting intellectual property, but traditional valuation methods often fail to fully capture their strategic value. This paper introduces a causality-driven framework for patent valuation that integrates domain expertise with advanced language models in a structured interview-like process. By analyzing semiconductor patents filed between 1997 and 2007-a pivotal period of technological transformation-our approach employs Directed Acyclic Graphs (DAGs) and Structural Equation Models (SEMs) to reveal causal relationships, providing precise value attribution. This integration enhances the analysis of complex patent data, offering a powerful tool for aligning technological innovation with market and legal strategies.
With the advancement of computational power, large language models (LLMs) have rapidly developed in various fields. However, predicting stock price fluctuations remains a significant challenge, mainly due to the following two aspects: To address these challenges, we propose a new framework. We independently trained two models, TechGPT and SentiGPT, to analyze stock price data and text data from community platforms, respectively. By combining the outputs of TechGPT and SentiGPT, we developed a comprehensive model named IntegraGPT. During the training data collection process, we used In-Context Learning to require multiple large language models to generate reasoning rationales, avoiding reliance on a single large language model for reasoning. This approach addresses the issues of insufficient interpretability and model prediction bias.
The intrinsic value of metals provides a stable hedge against inflation; however, various macroeconomic factors, such as inflation, interest rates, and money supply, can have a profound impact on metal prices. This paper proposes a deep learning model to predict the prices of non-ferrous and precious metals, given their importance as components in investment portfolios. Positional encoding was applied to encode the daily insights of the sparse macroeconomic data, and a dynamic correlation encoder was used to model asset relationships without predefined information. In experiments, the proposed framework outperformed state-of-the-art methods in terms of Mean Absolute Error (MAE).
We propose optimization methods to accept immediate tasks in automated warehouses. While Multi-Agent Pickup and Delivery (MAPD) problems for the automated warehouses suppose that all delivery tasks are given at planning phases, there are several situations where the system needs to accept immediate and emergent tasks in real-time in practical applications. We conducted four methods, Optimal Timing Search, Optimal Vertical Location Search, Optimal Horizontal Location Search, and Variance Minimization to assign an immediate task into already-planned schedules of MAPD. We demonstrate that the proposed methods can mitigate costs increases associated with accepting immediate tasks on the system.
This study analyzed players’ decision-making process in Werewolf and the success factors of persuasion in a 5-player Werewolf. In Experiment 1, we focused on the villager player and analyzed the process of deciding who to vote for based on what other players say. We proposed a decision-making model in which logical and economic rationality are prioritized, and when they still cannot decide, “empathy” becomes the deciding factor. Experiment 2 focused on persuasion situations and examined how successful persuasion is. The results confirmed that persuasion using logical persuasion materials is successful. This supported the model’s validity in Experiment 1 that logical rationality is the overriding criterion for decision-making. However, we could not observe the success factors when both persuaders had logical material. Future research will investigate the aspects of successful persuasion more deeply through complex persuasion situations and games with more proficient players.
With the continuous growth of electricity demand during peak times, demand response (DR) programs have emerged as an effective tool for load management. However, incentive-based DR programs often face issues of instability and uncertainty. Traditional methods typically involved direct control of user equipment to reduce electricity use, but this approach is not feasible for households without controllable devices. As such, this paper discusses how to evaluate the existing instability and uncertainty of incentive-based DR by only providing load reduction incentives and signals to households and proposes an algorithm to find feasible plans. We propose an innovative 2-phase Monte Carlo Simulation Genetic Algorithm (2P-MCSGA) designed to maximize the likelihood of achieving load reduction within a specific budget. Our approach integrates Monte Carlo simulation with genetic algorithms, employing constraint relaxation and gene repair techniques to enhance solution feasibility and algorithm efficiency. This method ensures both the practicality and effectiveness of the DR plan, offering a novel solution to the challenges of incentive-based demand response.
This paper proposes a method for creating a dashboard, assuming the task of the visual analysis on rating matrices used for information recommendation. Recommender systems have been developed to help people find relevant information. One of the most representative recommendation algorithms is collaborative filtering (CF), which utilizes rating matrices consisting of user-item interactions. However, CF does not explicitly take into account the type of person interacting with items. The proposed dashboard is composed of information that is expected to be useful for understanding user profiles, which are extracted from a rating matrix. It includes visualizations such as heatmaps and histograms that display information about users and items. By using it, analysts can find trends of user preferences and identify users with distinctive preferences. Furthermore, the developers of recommendation services can obtain valuable information for generating recommendation explanations and considering new systems. To show the effectiveness of the proposed method, this paper describes a case study assuming the users using the proposed dashboard. One of the authors uses the dashboard as if he were the user of the given rating history, and tries to find those with similar preferences by examining highly/poorly rated items and evaluation strictness. The result shows that similar users can be found by using the proposed dashboard, which demonstrates the effectiveness of the proposed method.
The objective of this research is to enable machines to simulate the gameplay of specific Go players and to obtain additional information during matches with opponents through AI recognition of Go move styles. To achieve this, two models were trained separately: a specific move policy model and a playing style recognition model. The specific player move policy model is designed to mimic the playing moves of a particular Go player. Since the game records of specific players usually consist of only a few hundred to a few thousand games, which is insufficient for training directly, we used transfer learning to train this model to solve this problem. Meanwhile, the playing style recognition model can identify the style of each move, categorizing it as fighting, balanced, or field type. We included the surrounding board conditions of the move in a 3×3 to 7×7 area as a feature map, achieving an accuracy of 83.2
A shift schedule is absolutely necessary to manage the work of each employee. However, manually constructing the schedule taken account of employee preference can impose a large burden on a constructor. In private tutoring school, a shift schedule should make arrangement of the requests of students in addition to teachers’ preference, being necessary to construct simultaneously a shift schedule for each teacher and a timetable for each student. In this paper, we have proposed a two-phase optimization method to find such schedules: shift scheduling using memetic algorithms and timetabling using simulated annealing. We have demonstrated that the proposed method, in particular memetic algorithms in the first phase, can be more effective for constructing a shift schedule and timetable that are taken account of both requests of teachers and students than the previous method using simple genetic algorithms and simulated annealing. In addition, we have developed an interface that allows users to interactively check and make minor modifications to the automatically created shift schedule. This user interface can be useful for the user to instantly check where constraints have been violated and assist in making corrections by the user.
In human communication, how we say things is just as important as what we say. This “how” is called paralanguage, e.g., the pitch of our voice, how we stress words, how fast we speak, and the rhythm of our speech. These elements help us understand what someone means, even if they say only one-word like “ha” In our study, we focus on how people use and understand the word “ha” in different situations as an extreme challenge for paralanguage recognition. We applied a game to collect data on how people say “ha” in eight different contexts. When people could see and hear others saying “ha,” they guessed the correct context about 63
Microsectioning is a destructive testing method extensively employed in the flexible printed circuit board (FPC) fabrication industry to assess the structural integrity and quality of FPCs. FPCs are essential components in various electronic devices due to their flexibility, lightweight nature, and ability to fit into complex shapes and spaces. A cross-section, or microsection, involves obtaining a thin slice of the FPC to expose its internal structure. During cross-section analysis, operators manually measure the thickness of FPC components, such as copper layers, OSC layers, and FSL layers. However, this manual process can lead to inconsistencies and difficulties in establishing standardized measurement procedures. To address these challenges, we propose an “AI-based Microsection Measurement Framework using ComfyUI Workflow” for FPC. This framework comprises five key modules: the target detection module, the image preprocessing and augmentation module, the AI model building and fine-tuning module, the measurement algorithm development module, and the ComfyUI visualization module. The measurement algorithm uses predicted masks from the AI model to perform precise measurements, while the visualization module plots these results directly onto the original image for easy review. In addition, we evaluate the proposed framework on two microsection types. Our experiments demonstrate that the measurement accuracy reaches an error margin of 0 pixels. Compared to the existing method, we provide a unified, faster, and lower labor cost measurement framework.
This paper proposes a method for recommending views to be added to dashboards using Knowledge Graphs (KGs) and association rules. With the increasing importance of information visualization driven by digitalization, dashboards have become essential tools in various fields. However, creating dashboards remains a challenging task that requires experience and expertise. As a basic technology for supporting dashboard creation, this paper proposes a method to predict appropriate views to be added to a dashboard under creation. Three methods are proposed, based on knowledge graph embedding (KGE), association rules, and a combination of both, and evaluated with a dataset created from actual dashboard data.
This study aims to determine the influence of comments posted online about news on viewers. Opinions posted by news readers via the web often influence the formation of other readers’ views. In this paper, we investigate the impressions of news articles under three conditions: (a) presenting a news article only, (b) presenting a news article and comments posted in the comment section of the article by readers, and (c) presenting a news article, comments posted by readers, and the readers’ previous comments for other news articles. The result revealed that (1) readers’ comments do not affect the impression of the original news article but are more influenced by the perspective of whether the respondent is familiar with the news article, (2) The impression of the comments that viewers agreed on tend to emphasize sincerity, responsibility, and credibility, regardless of the news article’s genre, and (3) presenting readers’ past comments may influence viewers’ evaluation of the information and lead them to form different opinions.
