Evolution and Prospects of Foundation Models: From Large Language Models to Large Multimodal Models

Since the 1950s,when the Turing Test was introduced,there has been notable progress in machine language intelligence.Language modeling,crucial for AI development,has evolved from statistical to neural models over the last two decades.Recently,transformer-based Pre-trained Language Models(PLM)have excelled in Natural Language Processing(NLP)tasks by leveraging large-scale training corpora.Increasing the scale of these models enhances performance significantly,introducing abilities like context learning that smaller models lack.The advancement in Large Language Models,exemplified by the development of ChatGPT,has made significant impacts both academically and industrially,capturing widespread societal interest.This survey provides an overview of the development and prospects from Large Language Models(LLM)to Large Multimodal Models(LMM).It first discusses the contributions and technological advancements of LLMs in the field of natural language processing,especially in text generation and language understanding.Then,it turns to the discussion of LMMs,which integrates various data modalities such as text,images,and sound,demonstrating advanced capabilities in understanding and generating cross-modal content,paving new pathways for the adaptability and flexibility of AI systems.Finally,the survey highlights the prospects of LMMs in terms of technological development and application potential,while also pointing out challenges in data integration,cross-modal understanding accuracy,providing a comprehensive perspective on the latest developments in this field.

Wild Gibbon Optimization Algorithm

Complex optimization problems hold broad significance across numerous fields and applications.However,as the dimensionality of such problems increases,issues like the curse of dimensionality and local optima trapping also arise.To address these challenges,this paper proposes a novel Wild Gibbon Optimization Algorithm(WGOA)based on an analysis of wild gibbon population behavior.WGOAcomprises two strategies:community search and community competition.The community search strategy facilitates information exchange between two gibbon families,generating multiple candidate solutions to enhance algorithm diversity.Meanwhile,the community competition strategy reselects leaders for the population after each iteration,thus enhancing algorithm precision.To assess the algorithm’s performance,CEC2017 and CEC2022 are chosen as test functions.In the CEC2017 test suite,WGOA secures first place in 10 functions.In the CEC2022 benchmark functions,WGOA obtained the first rank in 5 functions.The ultimate experimental findings demonstrate that theWildGibbonOptimization Algorithm outperforms others in tested functions.This underscores the strong robustness and stability of the gibbonalgorithm in tackling complex single-objective optimization problems.

Smart industrial IoT empowered crowd sensing for safety monitoring in coal mine

The crowd sensing technology can realize the sensing and computing of people,machines,and environment in smart industrial IoT-based coal mine,which provides a solution for safety monitoring through distributed intelligence optimization.However,due to the difficulty of neural network training to achieve global optimality and the fact that traditional LSTM methods do not consider the relationship between adjacent machines,the accuracy of human body position prediction and pressure value prediction is not high.To solve these problems,this paper proposes a smart industrial IoT empowered crowd sensing for safety monitoring in coal mine.First,we propose a Particle Swarm Optimization-Elman Neural Network(PE)algorithm for the mobile human position prediction.Second,we propose an ADI-LSTM neural network prediction algorithm for pressure values of machines supports in underground mines.Among them,our proposed PE algorithm has the lowest average cumulative prediction error,and the trajectory fit rate is improved by 24.1%,13.9%and 8.7%compared with Kalman filtering,Elman and Kalman plus Elman algorithms,respectively.Meanwhile,compared with single-input ARIMA,RNN,LSTM,and GRU,the RMSE values of our proposed ADI-LSTM are reduced by 36.6%,52%,32%,and 13.7%,respectively;and the MAPE values are reduced by 0.0003%,0.9482%,1.1844%,and 0.3620%,respectively.

A Literature Review of Fertility Education for High School Students

Purpose: Fertility education for high school students leads them to be able to decide when to get pregnant and give birth, and to think about their desired life plan. However, knowledge of fertility in Japan is low compared to other countries. This study aims to understand the state of fertility knowledge among high school students in Japan and to clarify issues in fertility education. Method: Ten articles published from 2013 to July 2023 were selected for review from ICHUSHI Web Ver. 5 and CiNii with the keywords “high school students,” “fertility,” and “education.” The review was limited to Japanese-language literature due to differences in fertility education and knowledge between Japan and other countries. Results/Conclusion: 60% - 90% of the high school students indicated that their source of information on fertility was school lessons. Half of the students understood egg aging and the effects of smoking and drinking on fertility, but their knowledge of menstruation, male reproductive physiology/function, and infertility was insufficient. 50% - 80% of students wanted to marry by the age of 30, and 70% - 80% wanted to have children. 80% of students who attended a lecture on preconception care considered the lesson an opportunity to think about their future life plans. 90% of school nurse teachers reported that fertility should be taught in schools, and half said that egg aging, causes of infertility, and other topics should be covered in high school. Educational caution points included the consideration of developmental stages and the relatedness of the content to what had already been taught, among other things. High school is the best time for fertility education because it gives students the opportunity to accept their own sexuality and think about their future. In addition, since many students plan to marry and have children by the age of 30, the acquisition of fertility knowledge can be an effective means of future life design.

Analysis of a POMDP Model for an Optimal Maintenance Problem with Multiple Imperfect Repairs

I consider a system whose deterioration follows a discrete-time and discrete-state Markov chain with an absorbing state. When the system is put into practice, I may select operation (wait), imperfect repair, or replacement at each discrete-time point. The true state of the system is not known when it is operated. Instead, the system is monitored after operation and some incomplete information concerned with the deterioration is obtained for decision making. Since there are multiple imperfect repairs, I can select one option from them when the imperfect repair is preferable to operation and replacement. To express this situation, I propose a POMDP model and theoretically investigate the structure of an optimal maintenance policy minimizing a total expected discounted cost for an unbounded horizon. Then two stochastic orders are used for the analysis of our problem.

Novel Biological Based Method for Robot Navigation and Localization

The capability and reliability are crucial characteristics of mobile robots while navigating in complex environments. These robots are expected to perform many useful tasks which can improve the quality of life greatly. Robot localization and decisionmaking are the most important cognitive processes during navigation. However, most of these algorithms are not efficient and are challenging tasks while robots navigate through complex environments. In this paper,we propose a biologically inspired method for robot decision-making, based on rat’s brain signals. Rodents accurately and rapidly navigate in complex spaces by localizing themselves in reference to the surrounding environmental landmarks. Firstly, we analyzed the rats’ strategies while navigating in the complex Y-maze, and recorded local field potentials(LFPs), simultaneously.The recorded LFPs were processed and different features were extracted which were used as the input in the artificial neural network(ANN) to predict the rat’s decision-making in each junction. The ANN performance was tested in a real robot and good performance is achieved. The implementation of our method on a real robot, demonstrates its abilities to imitate the rat’s decision-making and integrate the internal states with external sensors, in order to perform reliable navigation in complex maze.

Optimizing Deep Learning Parameters Using Genetic Algorithm for Object Recognition and Robot Grasping

The performance of deep learning(DL)networks has been increased by elaborating the network structures. However, the DL netowrks have many parameters, which have a lot of influence on the performance of the network. We propose a genetic algorithm(GA) based deep belief neural network(DBNN) method for robot object recognition and grasping purpose. This method optimizes the parameters of the DBNN method, such as the number of hidden units, the number of epochs, and the learning rates, which would reduce the error rate and the network training time of object recognition. After recognizing objects, the robot performs the pick-andplace operations. We build a database of six objects for experimental purpose. Experimental results demonstrate that our method outperforms on the optimized robot object recognition and grasping tasks.

Energy-Efficient Architecture and Technologies for Device to Device(D2D) Based Proximity Service

Considering that modern mobile terminals possess the capability to detect users' proximity,and offer means to directly communicate and share content with the people in close area,Device-to-Device(D2D) based Proximity Services(ProSe) have recently witnessed great development,which enable users to seek for and utilize relevant value in their physical proximity,and are capable to create numerous new mobile service opportunities.However,without a breakthrough in battery technology,the energy will be the biggest limitation for ProSe.Through incorporating the features of ProSe(D2D communication technologies,abundant built-in sensors,localization-dependent,and context-aware,etc.),this paper thoroughly investigates the energy-efficient architecture and technologies for ProSe from the following four aspects:underlying networking technology,localization,application and architecture features,context-aware and user interactions.Besides exploring specific energy-efficient schemes pertaining to each aspect,this paper offers a perspective for research and applications.In brief,through classifying,summarizing and optimizing the multiple efforts on studying,modeling and reducing energy consumption for ProSe on mobile devices,the paper would provide guide for developers to build energy-efficient ProSe.

Study on Performance Evaluation of Small Axial Fan

In order to establish the design methodology of a small axial fan, the axial fan with impeller diameter of 36 mm was designed, manufactured and tested. Especially, in order to investigate the influence of difference in blade cord length and blade thickness on performance characteristics, the performance characteristics obtained by the designed axial fans with difference blade shapes were examined. Also, by using CFD, the same flow field as the experiment was visualized. It was found that the lift of the blade was increased and the performance was improved in high flow rate region by thinning of the blade thickness and by extending the blade chord length.

Target Entities of Total Quality Management Based on the New TQM and Three-Dimensional Unification Value Models

In recent years, the method of TQM (Total Quality Management) is widely recognized. However, contents of target entities for assessment are various, and it is very difficult to define the whole scope of TQM. On the other hand, it is very important to define the whole target entities of quality management about the TQM because lack of important target entities may cause significant risk of loss in future. Furthermore, target entities of TQM should correspond to needs and priority of an objective requirement and that should be based on the consideration of basic principle of quality management. In the previous study, we have proposed the “framework of new TQM” of assessment for a total quality management of organizations based on the original concept of “TQM matrix”. On the other hand, we have proposed the view point of “Three-Dimensional Unification Value Models” for an evaluation of system product. Therefore, in this paper, we propose the target entities of whole assessment of quality management of organization totally based on the consideration of “New TQM framework” and the view point of “Three-Dimensional Unification Value Models”. Also, this paper proposes the result of verification based on the result of comparison between proposed target entities of assessment and American Malcolm Baldrige Prize for Performance Excellence.

Statistical Analysis of Process Monitoring Data for Software Process Improvement and Its Application

Software projects influenced by many human factors generate various risks. In order to develop highly quality software, it is important to respond to these risks reasonably and promptly. In addition, it is not easy for project managers to deal with these risks completely. Therefore, it is essential to manage the process quality by promoting activities of process monitoring and design quality assessment. In this paper, we discuss statistical data analysis for actual project management activities in process monitoring and design quality assessment, and analyze the effects for these software process improvement quantitatively by applying the methods of multivariate analysis. Then, we show how process factors affect the management measures of QCD (Quality, Cost, Delivery) by applying the multiple regression analyses to observed process monitoring data. Further, we quantitatively evaluate the effect by performing design quality assessment based on the principal component analysis and the factor analysis. As a result of analysis, we show that the design quality assessment activities are so effective for software process improvement. Further, based on the result of quantitative project assessment, we discuss the usefulness of process monitoring progress assessment by using a software reliability growth model. This result may enable us to give a useful quantitative measure of product release determination.

Negative Parity States in ^(39)Cl Configured by Crossing Major Shell Orbits

Traditional "magic numbers" were once regarded as immutable throughout the nuclear chart. However, unexpected changes were found for unstable nuclei around N = 20. With both proton and neutron numbers around the magic number of 20, the neutron-rich ^(39)Cl isotope provides a good test case for the study of the quantumstate evolution across the major shell. In the present work, the negative parity states in ^(39)Cl are investigated through the β decay spectroscopy of 39 S. Newly observed γ transitions together with a new state are assigned into the level scheme of ^(39)Cl. The spin parity of 5/2^- for the lowest negative parity state in ^(39)Cl is reconfirmed using the combined γ transition information. These systematic observations of the negative parity states in ^(39)Cl allow a comprehensive comparison with the theoretical descriptions. The lowest 5/2^- state in ^(39)Cl remains exotic in terms of comparisons with existing theoretical calculations and with the neighboring isotopes having similar single-particle configurations. Further experimental and theoretical investigations are suggested.

Resolution of Resource Contentions in the CCPM-MPL Using Simulated Annealing and Genetic Algorithm

This research aims to plan a “good-enough” schedule with leveling of resource contentions. We use the existing critical chain project management-max-plus linear framework. Critical chain project management is known as a technique used to both shorten the makespan and observe the due date under limited resources;the max-plus linear representation is an approach for modeling discrete event systems as production systems and project scheduling. If a contention arises within a single resource, we must resolve it by appending precedence relations. Thus, the resolution framework is reduced to a combinatorial optimization. If we aim to obtain the exact optimal solution, the maximum computation time is longer than 10 hours for 20 jobs. We thus experiment with Simulated Annealing (SA) and Genetic Algorithm (GA) to obtain an approximate solution within a practical time. Comparing the two methods, the former was beneficial in computation time, whereas the latter was better in terms of the performance of the solution. If the number of tasks is 50, the solution using SA is better than that using GA.

Modified Ramsey Rule, Optimal Carbon Tax and Economic Growth

In contrast to the overlapping-generations model, it is allowable to discount the future utility in a dynasty model without the ethical difficulty related to intergenerational conflicts. Much precedent research uses Ramsey-type optimal growth theory in order to estimate the social discount rate. However, one must note that almost all the formulations neglect the existence of negative intertemporal externalities. This problem is vital when one analyzes the global warming problem mainly caused by the excess concentration of carbon dioxide (CO2). This is because an adjoining effect of capital accumulation exists besides the improvement of product capacity, which is reflected in the rate of interest (or equivalently, the marginal productivity of capital). That is, one cannot neglect a negative externality to the future productivity that originates from the excess emissions of CO2. Accordingly, following the optimal growth theory, the effective social discount rate should be heightened by a proportional carbon tax to suppress future excess consumption/ emissions than in the case of the existing analyses, which exclude such an intertemporal external diseconomy.

Theoretical analysis of nonlinear vibration characteristics of gear pair with shafts

The circumferential vibration of a gear pair is a parametric excitation caused by nonlinear tooth stiffness,which fluctuates with meshing.In addition,the vibration characteristics of the gear pair become complicated owing to the tooth profile error and backlash.It is considered that the circumferential vibration of the gear pair is affected by the torsional vibration of the shafts.It is important to understand quantitatively the vibration characteristics of the gear system considering the shafts.Therefore,the purpose of this research was to clarify the nonlinear vibration characteristics of a gear pair considering the influence of the shafts using theoretical methods.To achieve this objective,calculations were performed using equations of motion in which the circumferential vibration of the gear pair and the torsional vibration of the shafts were coupled.The nonlinear tooth stiffness was represented by a sine wave.The influence of tooth separation was considered by defining a nonlinear function using backlash and the tooth profile error.For the numerical calculations,both stable and unstable periodic solutions were obtained by using the shooting method.The effect of the shafts on the gear system vibration were clarified by comparing the results in the cases in which the shaft was not considered,one shaft was considered,and both shafts were considered.

Analysis of Dose Calculation Accuracy in Cone Beam Computed Tomography with Various Amount of Scattered Photon Contamination

Cone-beam computed tomography (CBCT) images have inaccurate CT numbers because of scattered photons. Thus, quantitative analysis of scattered photons that affect an electron density (ED) curve and calculated doses may be effective information to achieve CBCT-based radiation treatment planning. We quantitatively evaluated the effect of scattered photons on the accuracy of dose calculations from a lung image. The Monte Carlo method was used to calculate CBCT projection data, and we made two calibration curves for conditions with or without scattered photons. Moreover, we applied cupping artifact correction and evaluated the effects on image uniformity and dose calculation accuracy. Dose deviations were compared with those of conventional CT in conventional and volumetric intensity modulated arc therapy (VMAT) planning by using γ analysis and dose volume histogram (DVH) analysis. We found that cupping artifacts contaminated the scattered photons, and the γ analysis showed that the dose distribution was most decreased for a scattered photon ratio of 40%. Cupping artifact correction significantly improved image uniformity;therefore, ED curves were near ideal, and the pass rate results were significantly higher than those associated with the scattered photon effect in 65.1% and 78.4% without correction, 99.5% and 97.7% with correction, in conventional and VMAT planning, respectively. In the DVH analysis, all organ dose indexes were reduced in the scattered photon images, but dose index error rates with cupping artifact correction were improved within approximately 10%. CBCT image quality was strongly affected by scattered photons, and the dose calculation accuracy based on the CBCT image was improved by removing cupping artifacts caused by the scattered photons.

Analysis of Influence of Price to Customer Satisfaction Based on the Prediction Models

For the purpose of improvement of the sales, confirming the influence of price to the sales and customer satisfaction of the product is important. The most suitable price should be determined from the view point of customers, and it is extremely important to implement a high quality product corresponding to the real need of customers. It may have close relationship between cost and an expense to implement the individual inherent attribute of system product. Also, it may have close relationship between production cost and price of product. For the purpose of improvement of the customer satisfaction for quality of system product, the method of quantitative quality requirement and evaluation based on the ISO/IEC9126 quality model that includes six quality characteristics is widely recognized. In the precedent study, I have introduced the requirements definition method for the quality of system product based on the system quality model defined in ISO/IEC9126 and proposed the effectiveness of it statistically. In the previous study, I have also confirmed the relationship between inherent attributes of the product and quantitative result of the measured value of total customer satisfaction from the view point of six quality characteristics statistically. I performed the development of the prediction model to estimate the total customer satisfaction for the system product from the view point of inherent attribute of the product. And, I have proposed the effectiveness of application of the estimated prediction model and possibility of improvement of the total customer satisfaction of a system product. Based on the result of previous study, in this paper, I propose the result of investigation of influence of price to customer satisfaction, and the possibility of application of estimated prediction model for improvement of the total customer satisfaction of system product based on the price of product. Also, based on the result of investigation of relationship among price and inherent attributes of product, I propose the possibility of application of estimated prediction model and improvement of the price of system product from the view point of inherent attributes of product.

Design-Stage Prediction of Project Correctness or Incorrectness

The success of a software development project requires the early objective determination of the project’s correctness or incorrectness and the identification of the most effective solution for project management. However, few studies have been conducted on the reliable quantitative early judgment of correctness or incorrectness. In recent years, the collection and accumulation of actual attribute data from Japanese domestic software development projects have been conducted by the Software Engineering Centre of the Information-Technology Promotion Agency of Japan. In a previous article, we proposed a precise definition of project correctness or incorrectness and identified the important factors in successful projects;we also proposed a quantitative decision-making method for judging project correctness or incorrectness objectively and quantitatively on the basis of discriminant analysis using project completion attribute data. On the basis of the previous results, we propose a quantitative decision-making technique for the early judging of project correctness or incorrectness based on the attribute data of design stage as early stage of development.

Theoretical Analysis and Simulation of Jacking Procedure of Pantadome System

In order to obtain the principle of Pantadome lifting process and make theoretical foundation for practical applications, the core idea of Pantadome was introduced, which is to make a structure become a mechanism by temporarily removing some members during the process of construction.The Abstract motion model was built. By determining the change of the coordinates of the hinge joint and that of each point of the structure, simulative analysis of the mechanical motion of Pantadome was realized. Then general program that simulates the lifting process of Pantadome was developed based on AutoCAD environment by Auto Lisp language. By completing the theoretical analysis of the lifting process of Pantadome, three-dimensional simulation of the lifting process of Pantadome was realized. And it is successfully applied to bidding work of practical engineering.

Reduction and Analysis of a Max-Plus Linear System to a Constraint Satisfaction Problem for Mixed Integer Programming

This research develops a solution method for project scheduling represented by a max-plus-linear (MPL) form. Max-plus-linear representation is an approach to model and analyze a class of discrete-event systems, in which the behavior of a target system is represented by linear equations in max-plus algebra. Several types of MPL equations can be reduced to a constraint satisfaction problem (CSP) for mixed integer programming. The resulting formulation is flexible and easy-to-use for project scheduling;for example, we can obtain the earliest output times, latest task-starting times, and latest input times using an MPL form. We also develop a key method for identifying critical tasks under the framework of CSP. The developed methods are validated through a numerical example.