電腦科學與資訊工程

背景：預測分子性質如溶解度、毒性及熔沸點對於基礎科學至關重要。然而，實驗測量這些性質耗時且昂貴，因此本研究使用多種機器學習模型藉由調整變相來準確預測熔、沸點。 方法：本研究使用超過一萬筆數據及兩種類型的機器學習方法：淺度與深度學習。淺度學習由 PyCaret實現，並以Mordred作為分子描述器；深度學習使用圖神經網路，包括（CMPNN和GCN），並調整隱藏層參數。 結果：CMPNN在目前嘗試的模型中表現最佳。發現影響沸點預測的關鍵特徵是piPC1，與鍵級相關；熔點則是AATS0d，與σ電子的 Moreau-Broto自相關有關。 結論：CMPNN模型在沸點與熔點預測中均表現最佳。沸點中深度學習模型優於淺度學習模型（p<0.05）。此外，使用SHAP成功找出piPC1和AATS0d對最關鍵。本研究不僅得出了高準確性的模型，還發現了影響分子性質的關鍵特徵，且可擴展至其他預測。

發展安全且可靠的身份辨識技術是當今的重要議題，而指靜脈因其高安全性及難以偽造特性成為我們的主題。本研究提出一種基於Transformer模型架構的指靜脈辨識模型稱為GLA-FD，旨在解決現有技術對指靜脈影像特徵表示與提取的局限性。透過開發特徵解耦與重建模組(FDRM)，模型能夠有效區分指靜脈的背景資訊與紋理特徵，並將其重新組合以提升辨識準確度。此外，本研究開發的全域-局部注意力模組(GLAM)能同時捕捉影像的全域與局部特徵，進一步強化模型對指靜脈特徵的理解。GLA-FD在FV-USM、PLUSVein-FV3、MMCBNU-6000、UTFVP、NUPT-FPV 資料集中的正確辨識率(CIR)達到100%、98.47%、99.75%、96.11%、99.82%，展現卓越的穩定性與泛化能力。此外，本模型在處理不同年齡層、國籍與影像模糊度的資料下，仍能保持高辨識準確度，顯示其在需要高安全性辨識的應用場景中具備廣泛的實用性。

Throughout our generations, a traditional system has been implemented for registering student attendance, in which the teacher is responsible for carrying out said activity, investing an average time of 15 to 20 minutes, which are part of the time of class. The objective of this project is to optimize this process, thus achieving effective class times, promoting the use of digital tools and innovation in teaching practice, in addition to generating security and confidence in tutors through the use of a service of message, which will notify the student's attendance in real time. Through a survey of the teaching staff of the CONALEP 338 Córdoba campus, it was detected that each teacher has academic loads equivalent to 3 to 5 modules per day, with an average of more than 40 students assigned to each module. Based on this information, the use of technological tools will be promoted and this process of teaching practice will be innovated with zero costs.

Our vision aims to raise the quality of life and we had planned smart cities from scratch but what about the current cities the residents of Riyadh suffer from extreme traffic and spend hours circling the block searching for a open park which wastes time money and is bad for the environment. 30-50% of traffic is causing by not being able to park and due to Riyadh lack of proper city planning and radid increase in inhabitants especially after allowing women to drive and as the car being the main way of transportation finding a open park could be a nightmare for some. We have approached this problem from the technological perspective by developing a free application for Riyadhs inhabitants that's main goal is to navigate each driver from their current location to the best open park possible in the shortest time possible but what dirstinguishes us from similar apps in the literature is that we provide the time of departure for each park as well as the ability to book suck parks even if it is ahead of time via a interactive live map. The technology's that we used are the cellar censor that will track the users location and the ultrasonic sensor to track the occupancy of the parking in case the driver doesn't have the app but in which case will case will not be able to provide booking features. We have struggled in the lack of expertise and experience and in motivating the drivers to input correct data about there time of departure we also didn't have enough time to validate our project For future work we will validate our project and we plan on making the detection of the time of departure automatic as well as vobering all kinds of parks. We plan on expanding the scope of target users to include institutes as well because with time the app will have collected enough data to help institutions provide better parking such as ruch hours parking scope percentage of booked parking etc we also plan on benefiting more from the cellular secsor to link data with the persons phone like certain access to private parks like disabled parking or home parking or private hotel offices parking etc

Stroke is a very common disease, almost a national disease. In terms of stroke frequency, 匈牙利 ranks second in the world. Every year, 40-50 thousand people become paralyzed or permanently injured as a result of cerebrovascular disorders. This number is three to four times higher than in developed countries. Almost every Hungarian family is affected! Of course, saving the life of someone who has a stroke is the most important thing, but rehabilitation is also very important, since only with the help of a physiotherapist will the patient be able to live a full life.

台灣少子化、老齡化問題迫切，醫療資源入不敷出，使得遠距醫療成為潛在的解決方案與趨勢。然而，遠距醫療的身份驗證安全性尚未完善。指靜脈辨識有非接觸、體內生物特徵等特性，其在衛生性和高安全性的優勢在醫學相關領域與醫療院所備受關注。若能夠開發出遠距醫療可用的指靜脈辨識，將有機會為遠距醫療產業的安全性貢獻一份心力。本研究主要分為兩階段：其一旨在優化指靜脈辨識技術，利用輕量化 CNN 指靜脈辨識模型，結合 Mini-RoI 技術， 使用 FV-USM 以及 PLUSVein-FV3 兩個資料集訓練我們開發的 FVeinLite 指靜脈辨識模型，並使用不同的 epoch 值訓練出最好的模型。而我們訓練的模型相較於其他指靜脈技術具有高辨識正確率、參數量更少、運算速度快等優勢。其二，我們將模型結合自製的低成本嵌入式裝置， 並製作 API 與並使用模擬的病患資料完整打造一個可使用於遠距醫療及醫療院所的指靜脈身分辨識系統。

本研究探討了利用生成式人工智慧技術為漫畫創作帶來新可能性。在當今競爭激烈的漫畫產業中，創作者們需要不斷創新以吸引觀眾，而創作引人入勝的漫畫需要豐富的想像力和劇情結構。本研究希望能協助創作者製作草稿，並探索與AI當朋友的新型創作模式。在生成方面，提出了將漫畫劇情提取、劇情預測以及圖片生成三個步驟的生成流程，並使用了多種模型和技術，如 YOLO模型用於漫畫人臉檢測、文字生成模型用於劇情預測、LoRA技術用於模型微調等，為解決人物生成不連續的問題，我們也提出一種基於特徵提取與融合的解決辦法。本研究提供了一個全面的方案，旨在利用人工智慧技術幫助漫畫創作者創作出簡單的草稿。

Medication errors in patients are a global problem. They can negatively affect patients and be costly for hospitals and medical clinics. In 2021, a 28-year-old man with heart problems was admitted to a hospital in Porto Alegre. Due to a pharmacy error and insufficient monitoring in the administration, he received a dose 10 times higher than prescribed. This caused serious and probably irreversible damage to the patient. Reading the news and following the case in the media has encouraged research in scientific databases, searching for information and data on medication errors, as well as emerging technologies to reduce the occurrence of adverse medication events. Based on the findings of an English study that proved that errors occur at the drug prescription stage, the first stage of this research focused on drug dosage errors. The aim of this study is to develop an application based on artificial intelligence that can recognise these errors and help prevent them. The application uses a neural network to analyse prescriptions and warn of possible cases of incorrect dosage. The computer program was developed using a neural network and the drug dosage error recognition system using Python and Keras. The system was trained with 10 drugs and correct and incorrect dosage cases. A graphical interface was created to input and display new case data. Neural networks with different configurations were tested to obtain high accuracy with the training and validation data. A confusion matrix was used to assess the accuracy of the network for cases not used for training. The accuracy was approximately 96%, but problems were found in certain intervals. The errors are due to the need for more training, higher processing capacity and a cloud server. The results of the first stage of the research indicate the feasibility of using a neural network to recognise medication dosage errors and thus preventing the associated risks. Such a method could prevent cases like the one in Porto Alegre. Future studies could incorporate more types of drugs, allergies, drug interactions, pre-existing illnesses and other relevant factors into the system.

Wibrazz is a wearable communication tool that allows the teacher, the therapist, the parent to communicate information to the child remotely using the device. Haptic (vibrationbased) feedback is becoming increasingly important in everyday life. A vibrating device that transmits information through clothing can help people with disabilities who have no or limited sensory use to live an integrated life in society without barriers.

本研究以 Minecraft 為例，探討沙盒類遊戲式學習平台系統伺服器架設的節能效率，旨在透過動態調整伺服器數量降低總CPU使用率，提升伺服器的管理效能和能源使用效率。隨著線上遊戲的普及，伺服器的營運管理變得越來越複雜，如何在滿足玩家需求並同時降低能源消耗成為一個重要議題。本研究將分析伺服器資源使用狀況，特別是在玩家活動量高低波動的情境下，透過管理策略的調整，探討其對節能效率的影響。 研究透過實證數據的收集與回歸分析，建立一套可應用於 Minecraft 伺服器的節能動態調整系統，並探討動態調整的具體效率。研究結果發現隨著玩家人數增加，越接近系統負載上限，節能效果會越來越不明顯，以本次研究的伺服器來講玩家人數到達35人以後就無法再減少伺服器數量。