3164 research outputs found
Sort by
工学的立場からの農業支援の一端
Get PDF爆発的な世界人口増加に対して日本の農業従事者は減少の一途を辿っている．そのような将来の食糧問題に立ち向かうためにスマート農業が推し進められている．本論文では，現在進行中の工学的立場からの農業支援の一端について著者らが取り組んでいる現状について述べる．スマート農業という言葉が生まれてから既に10年近くが経過し，その間多くの技術が提案されてきたが，未だ植物工場をはじめとするスマート農業には問題が山積している．それらの現状を踏まえ，我々も一歩ずつ進歩している最中のセンサやシステムの開発等の一端をご紹介して，スマート農業の一助となれば幸いである．journal articl
Impact on U.S. Diplomacy based on the Relationship between the American President and Secretary of State
Get PDFThe United States is a world superpower which exerts great influence diplomatically. However, this influence has not resulted in great diplomatic achievements so far. For example, the USA could not get worldwide cooperation to support the Iraq War in 2003. Although there are several reasons, I would like to research how the relationship between the American President and the Secretary of State has an impact on the outcome of diplomacy. C-SPAN conducted a survey in which 142 historians or professors rated each president in 10 categories including foreign relations. According to this result, the highest ranking president in terms of foreign relations after World War Two was Dwight Eisenhower and the worst was Donald Trump. Eisenhower had a good relationship with his Secretary John Foster Dulles and Trump fired Secretary Rex Tillerson, reflecting their cold relations. The historical data or documents show that the higher the ranking, the better the relationship between the President and Secretary of State. One of the reasons is that a bad relationship would make the President inaccessible to the abundant human and/or secret resources of the State Department. A bad relationship may also hamper the efficient cooperation within the National Security Council, in which the Secretary of State participates. In order to analyze the results of diplomacy of the Biden administration, it is important to examine the relationship between President Biden and Antony Blinken, the current Secretary of State.departmental bulletin pape
A Case Study on Learning Chinese as a Second Foreign Language at Kitami Institute of Technology
Get PDFIn this paper, the author conducted a questionnaire survey of two Chinese
language classes at Kitami Institute of Technology regarding the reasons for
enrolling in Chinese courses and the relationship to joining study abroad programs.
According to the results, most of the students were mainly interested in earning
credits, and only a few enrolled in the classes for other reasons. From the
perspective of the internationalization of the university, it is necessary to establish
a system in which taking a second foreign language is not merely a matter of
earning credits, but is also useful for studying abroad and finding a job. In addition,
developing diverse learning opportunities, especially in light of the merger of the
three universities in April 2022, is important.departmental bulletin pape
教育目的のすごろくとその拡張ゲームの数理に関する一考察
Get PDFすごろくは伝統的なゲームである. すごろくに確定的なイベントや利得を追加したゲーム, 行動選択や確率的な利得を追加したゲームなど, さまざまなタイプのすごろくの拡張ゲームがある. すごろくやその拡張ゲームはさまざまな教育分野で利用されている. 教育目的に基づいて効果的なゲームを作成するには, 振り出しからゴールまでにサイコロを振る回数の期待値や, 行動選択を伴う拡張ゲームにおける期待総利得の最大値などに関する数理工学に基づく検討が必要である. しかし, 数理工学に基づく検討, 特に期待総利得に関する検討はほとんど行われていない. そこで, 本研究では, 拡張ゲームにおける期待総利得の計算方法を検討し, 提案方法の有効性を計算例で示す. 従来研究ではゲームのプレイ後でないと利得設定の妥当性は確認できない. しかし, 提案方法では事前に確認できる. 本研究は基礎研究であり,今後の拡張研究が必要である.journal articl
スマートマニュファクチャリング用ビッグデータアナリティクスの開発
Get PDF北見工業大学博士 （工学）Big data means horizontally networked yet independent data systems containing a vast number of structured and unstructured datasets. Statistical and logical computational arrangements (referred to as big data analytics) must be installed to make sense of big data. Like human-cyber-physical systems, digital twins, artificial intelligence, the internet of things, and sustainability, big data and big data analytics are essential elements of the fourth industrial revolution or smart manufacturing. In this study, big data of manufacturing processes are categorized into three main issues: 1) datasets for digital twin; 2) control variable-evaluation variable-centric datasets; 3) graphical datasets. This thesis considers the problem of developing big data and big data analytics for smart manufacturing, focusing on three related issues.
Issue 1: digital twins of manufacturing phenomena are supposed to machine-learn the required knowledge using relevant datasets available in big data. Therefore, a research question is how to preprocess manufacturing phenomena-relevant datasets for using them directly in digital twins.
Issue 2: Big data and analytics require expensive resources and sophisticated computation arrangements. Thus, big data hardly benefits small and medium-sized manufacturing organizations, resulting in “big data inequality.” Consequently, a research question is how to eliminate big data inequality.
Issue 3: big data is often visualized using several two-dimensional plots (graphical dataset). These plots are then used to make a decision informally. Consequently, a research question is how to make formal decisions by computing two-dimensional plots, not numerical data.
Thus, this thesis is organized as follows.
Chapter 1 presents this thesis's background, objective, scope, and limitations. It also presents a comprehensive literature review on big data relevant to smart manufacturing.
Chapter 2 describes the proposed big data analytics framework showing all the subsystems. In this chapter, the functional requirements of the subsystems are explained. big data analytics is developed for manufacturing process-relevant decision-making. The proposed analytics consists of five integrated systems: 1) big data preparation system, 2) big data analytics exploration system, 3) data visualization system, 4) data analysis system, and 5) knowledge extraction system. The big data analytics preparation system prepares contents that exhibit the characteristics of digital manufacturing commons.
Chapter 3 deals with Issue 1. A digital twin consists of five modules (input, modeling, simulation, validation, and output modules), and big data must supply datasets for building these modules. This chapter presents a manufacturing phenomenon-related datasets preprocessing method considering the four modules of digital twins (input, modeling, simulation, and validation modules). As an example, the preprocessing of surface roughness-relevant datasets is considered.
Chapter 4 deals with Issue 2. This chapter described the developed big data analytics framework for the control variable-evaluation variable-centric dataset. This system can support user-defined ontology and automatically produces Extensible Markup Language-based datasets. The big data exploration system can extract relevant datasets prepared by the first system. The system uses keywords derived from the names of manufacturing processes, materials, and analyses- or experiments-relevant phrases (e.g., design of experiment). The third system can help visualize relevant datasets extracted by the second system using suitable methods (e.g., scatter plots and possibility distribution). The fourth system establishes relationships among the relevant control variables (variables that can be adjusted as needed) and evaluation variables (variables that measure the performance) combinations for a given situation. In addition, it quantifies the uncertainty in the relationships. The last system can extract knowledge from the outcomes of the fourth system using user-defined criteria (e.g., minimize surface roughness and maximize material removal rate). The efficacy of the proposed big data analytics is demonstrated using a case study where the goal is to determine the right states of control variables of dry electrical discharge machining for maximizing material removal rate. It is found that the proposed big data analytics is transparent and free from big data inequality.
Chapter 5 deals with Issue 3. Big data analytics is developed to compute two-dimensional plots (graphical datasets) generated from big data. The efficacy of the tool is demonstrated by applying it to assess sustainability in terms of Sustainable Development Goal 12 (responsible consumption and production). Regarding that, engineering materials' functional, economic, and environmental issues play a vital role. Accordingly, three two-dimensional plots generated from big data of engineering materials are computed using the proposed analytics. The plots refer to six criteria (strength, modulus of elasticity, cost, density, CO2 footprint, and water usage). The proposed analytics correctly rank the given materials (mild steel, aluminum alloys, and magnesium alloys).
Chapter 6 describes future research directions and discusses the implication of this study from the viewpoint of smart manufacturing.
Finally, Chapter 7 provides the concluding remarks of this thesis.doctoral thesi
Zero-shot cross-lingual transfer language selection using linguistic similarity
Get PDFWe study the selection of transfer languages for different Natural Language Processing tasks, specifically sentiment analysis, named entity recognition and dependency parsing. In order to select an optimal transfer language, we propose to utilize different linguistic similarity metrics to measure the distance between languages and make the choice of transfer language based on this information instead of relying on intuition. We demonstrate that linguistic similarity correlates with cross-lingual transfer performance for all of the proposed tasks. We also show that there is a statistically significant difference in choosing the optimal language as the transfer source instead of English. This allows us to select a more suitable transfer language which can be used to better leverage knowledge from high-resource languages in order to improve the performance of language applications lacking data. For the study, we used datasets from eight different languages from three language families.journal articl
Continuous measurement of flow direction and streamflow based on travel time principles using a triangular distribution of acoustic tomography systems
Get PDFMonitoring flood dynamics is important to provide optimal flood mitigation practices and understand their hydrological responses. Recently, applications of travel time principles have gained a growing interest in hydrological research and river engineering. In this study, flood dynamics were monitored using the fluvial acoustic tomography system (FAT), based on travel time principles. The primary objective of this study is to continuously measure the mean cross-sectional velocity, river flow direction, and river discharge using an innovative tomographic system during two flood events. In this regard, three FAT systems were placed in a gravel bed stream forming a triangle shape to measure stream velocity along two cross-sections. By investigating the magnitude of the cross-sectional mean velocity vectors and the cross-sectional areas, we proposed new equations to evaluate the expected flow direction. The performance of flow measurement by the FAT system was verified with another reference record. Importantly, we demonstrated that the minimum number of acoustic stations to determine river flow direction in unidirectional streams can be reduced to three stations which can be more practical and easier. Further, one of the novel aspects of this study is offering new guidelines to continuously estimate flow direction using a triangular distribution of tomographic systems. In general, this study presents a promising method for monitoring flow dynamics in rivers.journal articl
Applications of travel time principle to measure river flow dynamics using underwater acoustic tomography system
Get PDFAdvanced and precise acquisition of river flow in remote regions such as mountain rivers is essential for developing, correcting, and improving the current hydrological and physical models. The primary objective of this study is to shed light on the applications of travel time principle using the fluvial acoustics tomography system (FAT) as an advanced tomographic technology to evaluate river temperature, velocity, and discharge. Accordingly, a pair of the FAT system was installed on both riverbanks, then mutual transmission of acoustic signals between the stations was conducted to estimate the sound speed and thus to measure both river velocity, and flowrate. In addition, stream mean temperature was estimated using sound speed. To validate the precision of our measurement, comparisons with other data were conducted. Generally, this study reveals a promising and advanced system able to deliver accurate records of streamflow dynamics.journal articl
Unied Processing of Shape and Motion Information Using Retinal Ganglion Cell Model and Spiking Neural Network
Get PDF汎用人工知能の基盤となるモデルとして注目されるSpiking Neural Network（SNN）をマルチモーダルな情報処理タスクに適用することを目的として，視覚において空間情報（形状）および時間情報（動作）の抽出を担う網膜神経節細胞のモデルとSNNの組合せによる形状認識・動作認識の統一処理モデルを提案した．人物の様々な動作を収録した動画データセットを用いた提案モデルの学習実験により，形状認識，および動作認識の双方を同一のSNN上で実現可能であることを示した．journal articl