The main goal of this paper was a creation of musical score for the composition from the 18th century Domine ad Ad juvandum in A minor and C major composed by Šimon Brixi. Another goal was finding as much information as possible related to this composition and summarize everything possible that was written previously. Separate parts were transcribed into one partiture. Nevertheless there were difficulties like mistakes in sheet music (corrected according to harmony of other voices), missing beats (replaced by identical parts of score), differences in period station. Hard times also bring on differences of period notation. The resulting musical scores can be directly used for music production, which is schemed for the next year. The second part was researching new or never-published information related to these compositions. Here I describe Š. Brixi´s life. I managed to read out the cause of Šimon Brixi death in a chronicle, which was not in the publication I found. Brixi died of tuberculosis like his son. Transcriptors, J.V.Flaška and J.J.Hoffman, are also mentioned. Lastly the information about sheet music itself and its historical context were described. Another discovery was an identification of watermark on sheet music paper. Assuming the sheet music is truly transcribed in 1742 the Kolinec paper mill watermark on the sheet music paper is the oldest evidence of its use. In this paper I summarized the historical context of the baroque era including social life, art and culture.I mentioned the most important events that were crucial for historical development in Bohemia and especially in the Klatovy region. This context can be useful for today’s musicians who want to understand the historical background and express the composition with full respect to the time of its creation.

Face pose estimation has many possible applications, ranging from driver attention measurement systems to applications in the metaverse, which this project will be focused on. Rather than using a more traditional landmark-to-pose method where the head pose is estimated via keypoints, our method trains a simple convolutional neural network, using the dataset 300W_LP, where the images are simply inputted into the network. The model is fitted with three fully connected layers that are linked to the each of the three Euler angles (yaw, pitch, and roll), alongside multiple loss functions, which improve the robustness of the network.

To investigate the reproduction success and natural recruitment of several Cyprinidae fish and Claridae fish species in the Allemanskraal Dam. The purpose of the project included investigating whether each individual fish species studied has a successful 2020/2021 spawning season in comparison with each other. Sections of the seine net were measured along with a distance of 10 along the shoreline. The ends of the seine nets were attached to one foot and the top of the net was held by hand. Both volunteers moved in unison while covering the 10m. The volunteer in the “deep end” moved towards the shoreline creating a semi-circle while the other volunteer remained stationary. The two ends of the net were then pulled onto the shore and the fish were collected. The results found that the Labeo Umbratus and Cyprinus carpio had the most successful spawning seasons with the highest recorded numbers. These high numbers of the Labeo Umbratus can be due to the fact that the species lays a large number of eggs. The high numbers of the Cyprinus carpio is due to the lower numbers of the other fish species as previous studies have shown that the species negatively impacts the environment which could in turn negatively impact the other fish species. The Claridae gariepinus and Labeobarbus aeneusas were the lowest. The low numbers of the Labeobarbus aneusas may be due to their slow growth and late maturity rate. The Labeo capensis had an average number relative to the other species and this is due to the fact that during the sampling period the dam was at 100% capacity as this is essential for the survival of the juvenile fish. The hypothesis was accepted as the Labeo Umbratus, Cyprinus carpio and Labeo capensis all have a successful spawning season. However, due to the size of the Cyprinus carpio, they would be most suited for a small scale fishery.

目前水冷散熱系統普遍應用在電腦CPU降溫中，雖然水的「比熱容」比空氣及大部分介質都高，但因水冷散熱系統內冷卻液的熱量，最終仍靠風扇送到機殼外，故CPU之最低溫度仍存在一個臨界值。本研究旨在對於「一體式」與「分離式」電腦水冷散熱系統及「熱電致（製）冷晶片」（Thermoelectric Cooling Module）結合進行模組開發設計，將此兩類相關元件搭配結合，以突破傳統水冷式散熱所無法降達的溫度。本研究將「致（製）冷晶片」之致冷端及水冷系統作結合，利用致冷端作為吸收CPU主要熱量，結果發現：與單純只利用風扇將熱量帶走的方式相比，本研究所開發之『第一代』一體式散熱模組與『第二代』分離式散熱模組皆成功地將頂級CPU之工作溫度再壓低，使電腦工作效率維持在最佳範圍。

Currently, no method can completely eliminate the human immunodeficiency virus (HIV) in an infected person. HIV employs an accessory protein called Nef that forms a complex with cellular AP-1, preventing detection of HIV-infected cells. Lovastatin has been recently identified to inhibit the formation of said Nef-AP-1 complex, but its effective concentration is remarked to be far higher than other Nef inhibitors. This study aims to develop a modified lovastatin molecule exhibiting higher binding affinity to the HIV-1 Nef protein than lovastatin in silico. Modified lovastatin molecules based on the interaction map of lovastatin with Nef were modeled, and flexible ligand-flexible receptor docking to the Nef binding site was performed using AutoDock Vina. Residues within the Nef binding site identified by Liu et al. (2019) to be crucial (Glu-63, Val-66, Phe-68, Asp-108, Leu-112, Tyr-115) were set as flexible. Fragment-based drug design was utilized to append molecular fragments to lovastatin in order to maximize its interactions with said crucial residues. From the fragment-based approach, molecule F4 ((1S,3S)‐8‐{2‐[(2R,4R)‐4‐chloro‐6‐oxooxan‐2‐yl]ethyl}‐3‐(hydroxymethyl)‐7‐methyl‐1,2,3,4‐tetrahydronaphthalen‐1‐yl 4‐aminobenzoate) exhibited a binding affinity of -9.0 kcal/mole, and its estimated IC50 ranges between 0.25-0.51 μM which is at least 7.5 times lower than the reported IC50 of lovastatin from literature. This study presents insights on the key modifications to improve lovastatin as an HIV-1 Nef inhibitor and pertinent information about the Nef binding site for future drug development studies.

本實驗主要在研究各項因子對氫氣火箭的影響，我們查了許多影片和資料，發現許多氫氣火箭實驗都沒有記錄各項因子的數據及比較，在歷屆科展的火箭實驗中，將氫氣作為燃料的作品也寥寥無幾，所以決定嘗試這個題目作為本實驗的研究主題。 我們的研究使用電解製造氫氣與氧氣，進行燃料、造型實驗，並用電腦計算出飛行高度，希望能找出最佳飛行高度因子。我們發現各種因子的飛升高度與質量、質心、燃料與造型有很大的關聯。 希望能藉由這些實驗讓人們更認識氫氣火箭，也能藉由氫氣火箭的飛行實驗讓大家更認識、更熟悉火箭之中的科學原理，讓大家對科學更有興趣。

With the introduction of the transformer architecture by Vaswani et al. (2017), contemporary Text Generation Models (TGMs) have shown incredible capabilities in generating neural text that, for humans, is nearly indistinguishable from human text (Radford et al., 2019; Zellers et al., 2019; Keskar et al., 2019). Although TGMs have many potential positive uses in writing, entertainment and software development (Solaiman et al., 2019), there is also a significant threat of these models being misused by malicious actors to generate fake news (Uchendu et al., 2020; Zellers et al., 2019), fake product reviews (Adelani et al., 2020), or extremist content (McGuffie & Newhouse, 2020). TGMs like GPT-2 generate text based on a given prompt, which limits the degree of control over the topic and sentiment of the neural text (Radford et al., 2019). However, other TGMs like GROVER and CTRL allow for greater control of the content and style of generated text, which increases its potential for misuse by malicious actors (Zellers et al., 2019; Keskar et al., 2019). Additionally, many state-of-the-art pre-trained TGMs are available freely online and can be deployed by low-skilled individuals with minimal resources (Solaiman et al., 2019). There is therefore an immediate and substantial need to develop methods that can detect misuse of TGMs on vulnerable platforms like social media or e-commerce websites. Several methods have been explored in detecting neural text. Gehrmann et al. (2019) developed the GLTR tool which highlights distributional differences in GPT-2 generated text and human text, and assists humans in identifying a piece of neural text. The other approach is to formulate the problem as a classification task to distinguish between neural text and human text and train a classifier model (henceforth a ‘detector’). Simple linear classifiers on TF-IDF vectors or topology of attention maps have also achieved moderate performance (Solaiman et al., 2019; Kushnareva et al., 2021). Zellers et al. (2019) propose a detector of GROVER generated text based on a linear classifier on top of the GROVER model and argue that the best TGMs are also the best detectors. However, later results by Uchendu et al. (2020) and Solaiman et al. (2019) show that this claim does not hold true for all TGMs. Consistent through most research thus far is that fine-tuning the BERT or RoBERTa language model for the detection task achieves state-of-the-art performance (Radford et al., 2019; Uchendu et al., 2020; Adelani et al., 2020; Fagni et al., 2021). I will therefore be focussing on attacks against a fine-tuned RoBERTa model. Although extensive research has been conducted on detecting generated text, there is a significant lack of research in adversarial attacks against such detectors (Jawahar et al., 2020). However, the present research that does exist preliminarily suggests that neural text detectors are not robust, meaning that the output can change drastically even for small changes in the text input and thus that these detectors are vulnerable to adversarial attacks (Wolff, 2020). In this paper, I extend on Wolff’s (2020) work on adversarial attacks on neural text detectors by proposing a series of attacks designed to counter detectors as well as an algorithm to optimally select for these attacks without compromising on the fluency of generated text. I do this with reference to a fine-tuned RoBERTa detector and on two datasets: (1) the GPT-2 WebText dataset (Radford et al., 2019) and (2) the Tweepfake dataset (Fagni et al., 2021). Additionally, I experiment with possible defences against these attacks, including (1) using count-based features, (2) stylometric features and (3) adversarial training.

目的：在台灣，胰臟癌雖非癌症發生數最高，但85%為末期，五年存活率小於5%，有癌王之稱。如何早期偵測胰臟癌風險，一直是醫療上的重大議題。本研究運用時間矩陣搭配深度學習進行大量變數之胰臟癌風險偵測。 方法：健保資料庫收集國人疾病、用藥等結構化資料，可藉此反映每人健康狀態。將三年內診斷與藥物碼轉為時間矩陣，以卷積神經網路訓練，訓練組與測試組比例為9:1。 結果：共計案例組1095名及對照組10950名，訓練後測試組表現之AUROC(area under the receiver operating characteristiccurve)為0.937，六十歲以上及以下AUROC分別為0.846及0.897。 結論：結論：能將診斷、藥物、時間轉成矩陣，是以疾病軌跡預測胰臟癌風險，且能找到新特徵，未來搭配健康存摺，為低成本、快速、大量的胰臟癌數位快篩。

本研究利用合成生物學技術開發的創新基因組篩檢平台PRISM(v2.0)，採用合成轉錄因子(crispr-TF)辨認DNA但不進行剪切，透過修改CRISPR/Cas的PAM-interacting (PI) domain序列打破其辨認限制，由原本NGG序列變成可以結合NAG的序列，提升crispr-TF在高通量基因組的篩選能力，期待未來可以做出辨認NNN序列，以此技術初步成功建構Chk2/Rad53所參與的基因網絡，讓失去Chk2/Rad53功能的酵母菌能在DNA損傷的情況生存。Chk2/Rad53蛋白磷酸酶在DNA修復的訊息傳遞中調控龐大的基因網絡，其中潛在大量不同調控路徑聯結的組合性，過去的研究都無法有效地分析且重建完整的DNA修復機制。未來將進一步探討Chk2/Rad53調控的下游基因組，並以此新技術和研究應用於複雜腫瘤與老化疾病的預防及治療。

A variant of the coronavirus (2019-nCoV or COVID-19) was reported in Wuhan, China on December 31, 2019, spreading rapidly around the world. Owing to a plan to reactivate the economy, the Mexican government, requested to implement protective measures to enter establishments with confined spaces: wear a mask, provide alcohol-based hand rub and the measurement of body temperature, allowing problem when not have a fever; however, these measures cause long waiting lines, causing contagion risk. To support this problem, applied and experimental research was used, generating as a result, the creation of a robot that moves autonomously thanks to a line follower sensor. Dispenses alcohol-based hand rub and determines the temperature by means of an infrared sensor, checking that the distance is adequate, with an ultrasonic sensor, using a buzzer that emits a sound of half a second when it performs a normal detection, but, if it is equal to or greater os 38 ° C, emits a sound for 10 seconds asking the person to leave the line and go to a medical service, helping to reduce problems infections in lines

本研究源於 2022 年數學雜誌《Crux Mathematicorum》的一道四邊形動態幾何問題，我們先將此問題設定為三角形，利用綜合幾何方法給出了兩種構圖條件下的三角形外心軌跡皆為圓弧，並且發現兩種圓弧的變換關係，也給出豐富有趣的性質。值得一提的是，分別對三角形的三個頂點輪換進行第一種構圖得出三個圓弧，這些圓弧恰可組合成三角形的九點圓，這是有趣的發現！回到原始問題的四邊形，我們構造了兩個三角形，透過巧妙轉換頂角與直徑圓變換而給出外心軌跡所在圓弧的兩個定點而解決此問題。最後推廣至鄰邊連線時，我們用雙射對應觀點簡潔刻劃了軌跡圓弧。重要的是，本研究處理四邊形的手法與三角形的是一致的，意味著證明手法具有推廣性。

According to CNN Indonesia 2020, the demand for e-Commerce in Indonesia has nearly doubled during this pandemic. This surge in demand calls for a time-efficient method for warehouse order-picking. One approach to achieve that goal is by incorporating automation in their warehouse systems. Globally, the market of warehouse robotics is expected to reach 12.6 billion USD by 2027 (Data Bridge Market Research, 2020). In this research, the warehouse system studied would utilize AMR (Autonomous Mobile Robots) to lift and deliver movable shelf units to the packing station where workers are at. This research designed a heuristic algorithm called A.N.T.s (Algorithm for Navigating Traffic System) to conduct task assigning and pathfinding for AMR in the automated warehouse. The warehouse layout was drawn as a two-dimensional map in grids. When an order is placed, A.N.T.s would assign the task to a robot that would require the least amount of time to reach the target shelf. A.N.T.s then conducted pathfinding heuristically using Manhattan Distance. A.N.T.s would help the robot to navigate its way to the target shelf unit, lift the shelf and bring it to the designated packing station. A.N.T.s algorithm was tested in various warehouse layouts and with a varying number of AMRs. Comparison against the commonly used Djikstra’s algorithm was also conducted (Shaikh and Dhale, 2013). Results show that the proposed A.N.T.s algorithm could execute 100 orders in a 27x23 layout with five robots 9.96 times faster than Dijkstra with no collisions. The algorithm is also shown to be able to help assign tasks to robots and help them find short paths to navigate their ways to the shelf units and packing stations. A.N.T.s could navigate traffic to avoid deadlocks and collisions in the warehouse with the aid of lanes and directions.