[重要通知]112學年度第1學期資工系大學部專題【系統整合與實作二】成果報告暨海報展(專題競賽)分組順序、Q&A資訊通知
[重要通知]112學年度第1學期資工系大學部專題【系統整合與實作二】成果報告暨海報展(專題競賽)分組順序、Q&A資訊通知
本學期「資工系大學部專題【系統整合與實作二】成果報告暨海報展(專題競賽)」經助教彙整各位檔案繳交情形、並由各群組之評審師長審閱書面報告後，12/06(三)成果報告的分組順序與Q&A通知請參閱以下報告時程與名單。
------------------------成果報告解說日流程------------------
成果報告解說日：12/06(三)13:30~16:45
地點：台達館6樓會議室(A601/B613/C614/D615/E617/F629)
流程：
|
時間 |
時程 |
說明 |
|
12:30-13:30 |
各組同學簡報檔案載入工作人員所準備之筆電及測試 |
同學亦可自行準備報告筆電跟相關設備 但請務必先到達報告場地，提前測試 |
|
13:30-14:50 |
1~8組報告 |
每組10分鐘(包含Q&A) |
|
14:50-15:10 |
休息時間 |
包含前面若有延誤的緩衝時間 |
|
15:10-16:00 |
9~13組報告 |
每組10分鐘(包含Q&A) |
|
16:00-16:15 |
休息時間 |
包含前面若有延誤的緩衝時間 |
|
16:15-16:45 |
14~16組報告 |
每組10分鐘(包含Q&A) |
|
16:45~ |
活動結束 |
|
|
請評審委員繳回負責群組之所有評分表 |
分組順序及Q&A之問題內容：若評審有新增問題則會在12/1(五)12:00前陸續更新
(更新日期：2023/11/28 13:02)
| 分組報告組別：A組(報告地點：台達601會議室) | ||||||||
| 報告順序 | 專題題目 | 專題學生姓名 | Q&A之問題 | |||||
| 1 | 2 | 3 | 4 | 5 | 6 | |||
| A-01 13:30-13:40 |
Estimate Volume of a Pineapple Using Intel Realsense | 林聖哲 | 吳聲宏 | |||||
| A-02 13:40-13:50 |
TinyML 在 Arduino Nano 33 BLE Sense 之應用：回收物圖像辨識與分類 | 吳竹婷 | 曾永茵 | |||||
| A-03 13:50-14:00 |
DriveEnv-NeRF: Exploring NeRF-Based Environment for RL Agent Training and Real-World Performance Validation in Autonomous Driving | 許佳綺 | 申牧義 | 黃鈺臻 | 侯皓予 | |||
| A-04 14:00-14:10 |
實現在延展實境(XR)頭盔上運行第一人稱手勢辨識模型 | 江佳臻 | 張芯瑜 | |||||
| A-05 14:10-14:20 |
Multi-view Layout Estimation | 楊煦 | ||||||
| A-06 14:20-14:30 |
Low-Cost Recommendation system with contextual logistic regression | 夏宇澄 | 邱煒甯 | 許瀚杰 | 楊輝 | 許至誠 | ||
| A-07 14:30-14:40 |
Extended Studies on Smart Grid Scheduling | 蘇芮筠 | ||||||
| A-08 14:40-14:50 |
幽默辨識模型與ChatGPT之應用 | 李風廷 | ||||||
| A-09 15:10-15:20 |
WER: Maximizing Parallelism of Irregular Graph Applications ThroughGPU Warp EqualizeR | 黃恩明 | ||||||
| A-10 15:20-15:30 |
張量分解在模型壓縮的應用 | 謝承彧 | ||||||
| A-11 15:30-15:40 |
基於HumanNeRF的3D人體模型建立和實時動作捕捉技術 | 沈家同 | 陳禹辰 | 唐聖翔 | 呂廷洋 | 葉明淳 | ||
| A-12 15:40-15:50 |
基於度量學習之室內全景影像分群與類別預測 | 張庭瑋 | ||||||
| A-13 15:50-16:00 |
人物2D轉3D快速建模 | 楊智明 | 葉宥忻 | |||||
| A-14 16:15-16:25 |
基於籃球轉播視角之球員軌跡及戰術分析 | 劉元愷 | ||||||
| A-15 16:25-16:35 |
利用機器學習建立肝癌手術選擇之分數系統 | 何習與 | ||||||
| 分組報告組別：B組(報告地點：台達613會議室) | ||||||||
| B-01 13:30-13:40 |
生成式對抗網路加入雜訊後在不利環境下對於模型萃取之應用 | 蔡明念 | 陳彥亨 | 1. It's known that generators are much more difficult to train than discriminators. 2. 你們的模型是用來生成對聯邦學習的攻擊的嗎? 為什麼要生成攻擊? |
||||
| B-02 13:40-13:50 |
Memory-augmented Reinforcement Learning in 3D Space | 王懷鴻 | 1.Pretty good. 2.yaw是什麼? 圖4看起來沒比baseline好? |
|||||
| B-03 13:50-14:00 |
Optimal Message Dissemination Scheduling under Hierarchical Trust Relationship | 林宜樺 | 林亭汝 | 顏曼娣 | 1. (1)Did you copy your report from a thesis? (2)The five algorithms in section 4 look equally good in terms of their complexity. You should discuss their relative merits and deficiencies. 2. How is the data about the tree given? Can you get size of sub tree in O(1) time? Lowerbound? |
|||
| B-04 14:00-14:10 |
Improving the Fairness of Artificial Intelligence in Chest X-Rays Interpretation Using Attention Mechanism | 陳昭汝 | 王筱君 | 郭晏昀 | 1. Not clear to me the meaning of this research. To me, it is not bad for AI models to have bias. 2. AUC? why are you doing learning with and without race label? |
|||
| B-05 14:10-14:20 |
演算法實作 | 段欣妤 | 1. Two weaknesses. (1) The focus is to implement an existing algorithm, rather than an innovative algorithm. (2) Experiments show that the existing algorithm contains errors. Overall, the result is dis-appointing. 2. 支架問題是甚麼? 執行結果的數據? |
|||||
| B-06 14:20-14:30 |
無人機群飛與精準定位 | 葉思永 | 黎彥谷 | 1. What is your definition of swarm flight? It can be very loose that a group of UAVs take off and land roughly at the same time. There is no constrain on their distances or flight patterns. The definition can be strict, in the sense that UAVs must follow a specific pattern, like the air shows on double 10 holidys. With a definition of swarm flghts, one can then evaluate the success of your project. 2. Any idea to make the drones closer? More responsive? |
||||
| B-07 14:30-14:40 |
臺指期程式化交易與漲跌預測的機器學習分析 | 張柏清 | 郭彥廷 | 何苡歆 | 謝喆安 | 1. Report writing has a lot room for improvement. Should try other machine learning methods. LSTM has been studied by other researcher. You should try other methods. 2. 30 分 K?跟資工有關的研究有點少? |
||
| B-08 14:40-14:50 |
Extended fake coin problems | 周彥君 | 1. (1) Not able to understand your description of the group size. Is it part of the problem description, or part of the method? (2) Fonts for symbols should be consistent. For example, sometime you use Roman font for symbol n, and sometime you use Italic fonts. 2. Related works? Do the fake coins have same weight? each "pair"? What's a pair? |
|||||
| B-09 15:10-15:20 |
A Survey of Contemporary Quantum Network Simulators: Benchmarking and Applicability Analysis of Select Simulators | 阮柏諭 | 1. (1) The report is well written. (2) This project fulfills good education purposes, but has relatively low innovation. 2. What does your quantum network actually do? If you only use QKD then it is just a classical network with more security |
|||||
| B-10 15:20-15:30 |
Visual Navigation with Topological Map | 許媄香 | 1. (1) The report is probably well written for experts in computer vision. But, for people outside of this area, it is not clear enough. (2) Reenforcement Learning (RL) methods typically have integer vectors for states and actions. In your problem, the input is a (3D?) image of the environment. It seems that you convert an image into a graph. But still, it seems that you need to convert a graph into a vector of probably real numbers. You need to give more information on RLs in order for outsiders to understand. 2. Plan to implement on a real robot? |
|||||
| B-11 15:30-15:40 |
congestion control for web real time communication | 黃卉綺 | 1. (1) The English in the report is poor. You should consider improving it using chatGpt. (2) Improvments on SCReAM are based on two existing methods. The contribution is minimum. 2. window-based actuation? explain make a better presentation of fig 4 |
|||||
| B-12 15:40-15:50 |
圖神經網路與推薦系統之探討 | 林祐禾 | 陳聖和 | 1. (1)You should first explain what a recommender system is and why a recommender system uses graphs. (2) Is it a negative result? It seems to indicate that you have not chosen a good method in the first place. 2. 你們怎麼把CL的結果放進UltraGCN的? |
||||
| B-13 15:50-16:00 |
Latency Minimization in Metaverse Applications: A Unified Approach to Computation, Communication, Cache and High Resolution in MEC | 范玉花 | 1. The report looks like a fairly complete research paper. It makes me wonder if you are the only participant of this research work. Is Algorithm 1 on page 7 a heuristic solution for the optimization problem in (6)-(10)? If so, how different is it between the optimal solution and the heuristic solution given by Algorithm 1? 2. plan to implement on physical devices? |
|||||
| B-14 16:15-16:25 |
協助結合大型語言模型、辭典知識庫的寫作改進系統 | 余婷真 | 薛伊婷 | 1. Are there automated way to resolve confusable words? In your method, a human is called in to resolve confusable words. 2. 你們做的是quiz 為什麼比較對象是文法糾錯? |
||||
| B-15 16:25-16:35 |
Mitigating Cold Start Latency within DAG-Based Serverless Architectures | 陳凱揚 | 張伃姍 | 1. (1)Why does a workflow of an FaaS system follow a DAG path without loops? Is it a practical assumption? (2) The results (between the branch prediction and the baseline method) are quite close. 2. explain serverless computing and relation to DAG |
||||
| B-16 16:35-16:45 |
基於桌球落點判定之轉播視角桌球影片動作區間偵測 | 余明璠 | 1. (1) The report has a lot of room for improvement. (2) You did a lot of work. For each module you used an existing method. Why not focus on the most essential module and work out your own method? Would it be more significant? 2. 你們好像是把影片切成一張張照片去做判讀的。有考慮利用影片連續的特性參考前後frame做判斷嗎? |
|||||
| 分組報告組別：C組(報告地點：台達614會議室) | ||||||||
| C-01 13:30-13:40 |
《Project Fusion》- 特殊回合制戰鬥遊戲之開發製作 | 黃冠諺 | 簡家聖 | 問卷調查收回了多少問卷? | ||||
| C-02 13:40-13:50 |
Transforming Product Reviews through Summarization, Key Aspect and Sentiment Analysis | 陳鮑比 | 貝仁義 | 施平治 | 劉其生 | 鄭聰明 | ||
| C-03 13:50-14:00 |
實作優化3D重建流程 | 徐聖旂 | 什麼是"相機姿態"? | |||||
| C-04 14:00-14:10 |
Contextual Multi-armed Bandit for Personalized Machine Teaching via Flashcards | 劉祥暉 | 吳嘉桐 | 施雪梅 | 楊慧英 | |||
| C-05 14:10-14:20 |
機器學習於靜態壓降評估之應用 | 梁遠 | ||||||
| C-06 14:20-14:30 |
基於車聯網的異常偵測平台開發與應用 | 張紘齊 | ||||||
| C-07 14:30-14:40 |
刑事判決書法條適用性AI判別系統 | 徐偉晉 | ||||||
| C-08 14:40-14:50 |
利用自架Pulpino平台破解Coremark跑分 | 王郁欣 | ||||||
| C-09 15:10-15:20 |
基於Content Planning與Mask Filling的中文旋律歌詞生成 | 王政淯 | 陳錫宏 | |||||
| C-10 15:20-15:30 |
Automated Localized Boxing Event Recognition and Classifier | 蕭澤然 | ||||||
| C-11 15:30-15:40 |
Recommendation Systems with Graph Collaborative Filtering and Automated Supervised Learning | 邱慧莉 | ||||||
| C-12 15:40-15:50 |
Multi-bit Large-scale Boolean Matching | 曹瀚文 | 陳祈瑋 | 許閎喆 | 林妤謙 | 鍾鎮嶸 | 黃盛揚 | |
| C-13 15:50-16:00 |
SCC 學生叢集競賽 | 吳邦寧 | 郭品毅 | |||||
| C-14 16:15-16:25 |
Prediction of Prognosis from EEG Brainwave Signals using an XGBoost Classifier Model | 紀維鑫 | ||||||
| C-15 16:25-16:35 |
可變形模組在固定輪廓的平面規劃 | 吳紀臻 | 曾怡茹 | |||||
| C-16 16:35-16:45 |
引入考量使用者視覺感受的虛擬實境應用程式隱私防護策略 | 鄭幸怡 | ||||||
| 分組報告組別：D組(報告地點：台達615會議室) | ||||||||
| D-01 13:30-13:40 |
Fixed-Outline Floorplanning with Rectilinear Soft Blocks | 羅心 | 1. 研究方法就是在實作UFO? 還是有獨到的設計來改良UFO? 2. 是否有跟其他方法的效能比較？ |
|||||
| D-02 13:40-13:50 |
Input Pruning for Fine-Grained Image Recognition | 游金勝 | 梁文勝 | 1. Interesting survey study | ||||
| D-03 13:50-14:00 |
3D Placement with Macros | 林晨 | 余侞璇 | 陳美晴 | 1. 表格呈現上可以強調舊的算法跟改良後算法的效能差異 2. 分析每個步驟對於效能上的影響 |
|||
| D-04 14:00-14:10 |
Prediction of Recovery from Coma Post-Cardiac Arrest using spontaneous EEG signals | 鍾宛里 | 林之耀 | 1. Why choose LSTM as a backbone model? Why not choose a newer architecture like Transformer as the backbone? 2. Are there any competing methods that tackle the same problem as yours?? |
||||
| D-05 14:10-14:20 |
Multi-bit Large-scale Boolean Matching | 胡昭泓 | 羅文佑 | 1. 是否有一些實際數據分析結果？ | ||||
| D-06 14:20-14:30 |
Malicious javascript and Phishing dectection | 謝長錡 | 莊凱威 | 廖品睿 | 1. 是否有跟現有產品進行數據上的比較？ 2. 蒐集16K良好跟70K惡意樣本, 訓練只取3K, 為何測試不拿剩下的樣本？只拿4200? |
|||
| D-07 14:30-14:40 |
The extended problems of wythoff's game | 林育頡 | 1. Interesting study | |||||
| D-08 14:40-14:50 |
Exploring Association of Odors Discrimination by Human Brain Olfactory Pathway and by Deep Learning Models | 蕭皓隆 | 彭馨屏 | 1. The topic seems interesting, and the experiments conducted are thorough. 2. However, as I am not an expert in this field, I would suggest using 'plain English' to describe what you are doing, why it is an important problem to solve, and what the main contributions of this study are. |
||||
| D-09 15:10-15:20 |
Lightweight Motion Capture System | 陳寬宸 | 1. 相機數量對於準確度跟辨識效能的影像為何？ 2. 報告只有提到速度，建議補上骨架重建準確度 |
|||||
| D-10 15:20-15:30 |
Traffic Digital Twin | 徐嘉徽 | 李佳栩 | 1. 有實作一個完整的系統 2. 因為方法分好幾個步驟，可以分析每個步驟誤差對於最後成果的影響 3. 目前只測試在1個路口？ |
||||
| D-11 15:30-15:40 |
對抗式攻擊與防禦應用於仇恨言論偵測 | 許育棠 | 劉昀翰 | 吳岱蔚 | 1. 需要強調在攻擊跟防禦兩端的技術貢獻在哪？目前看起來比較像是採用前人的方法來兜出兩個網業介面 | |||
| D-12 15:40-15:50 |
Animatable Musician with InstantAvatar | 洪聖祥 | 張雅涵 | 蔡侑廷 | 1. Why not directly capture a human playing an instrument? It seems unreasonable to me to separate the human performer from the instrument unless there is a specific purpose for the application. 2. I am curious about the timing performance of the proposed system. How long does it take to create a new scene from scratch? |
|||
| D-13 15:50-16:00 |
研讀及實作具效能保證深度學習方法於基地台換手問題 | 張哲語 | 呂尚豪 | 1. 整個研究方法跟試驗設置的描述有點混亂，無從得知該方法的具體貢獻在哪，報告的結構上有很大的改進空間 | ||||
| D-14 16:15-16:25 |
Sleep Amplitude Threshold: A Low-Cost Approach for N1 Sleep Stage Classification without Machine Learning | 姚林飛 | 秦文峰 | 1. It is unclear to me why the proposed method is more effective than previous approaches. What kind of insights in your design make it happen? 2. Also, I don't understand how exactly the baseline methods approach the problem and what the major differences are between your architecture and the previous ones. |
||||
| D-15 16:25-16:35 |
雷射光金屬上色技術之色域映射及應用 Color Mapping and its Applications in Laser Coloring | 連軒 | 林沛佳 | 1. 整體實驗很完整，不過我想了解如何定義最佳的彩色空間模型？ 2. 是否有一個量化的指標來評估模擬結果跟真實成像之間的差異？ |
||||
| 分組報告組別：E組(報告地點：台達617會議室) | ||||||||
| E-01 13:30-13:40 |
Recommendation with GNN | 陳天意 | 蔡沃林 | Why are you focusing on LightGCL? Have you really dedicated several months to just read one paper? | ||||
| E-02 13:40-13:50 |
Unreal Engine 5 Streaming Virtual Texture Benchmark | 謝明浩 | Could you share what was the most significant obstacle you encountered in your project? How did you tackle this challenge and how long did it take to overcome it? | |||||
| E-03 13:50-14:00 |
Performance Analysis Before and After the Upgrade of Virtualization-Based Fault Tolerance | 葉昱揚 | 可以描述一下你再升級改版的時候有遇到什麼問題嗎?你怎麼解決的?因為從現在的報告中看不到這些細節 | |||||
| E-04 14:00-14:10 |
The Evolution of Recommenders: Improving SimGCL with a Comparative Insight into Graph Neural Networks | 黎忠望 | 周志偉 | Could you explain the driving factors behind the improvements you proposed for each component? Also, which component experienced the most significant impact, and why? | ||||
| E-05 14:10-14:20 |
運用加密技術的分散式公民投票系統 | 簡弘哲 | 楊瑞凱 | 祝語辰 | 吳振群 | 在你們的報告中, 看不到過去針對投票系統設計的背景文獻, 我稍微查了一下網路, 很多強調安全的投票系統(such as "Decentralized online voting system using blockchain", "DVTChain: A blockchain-based decentralized mechanism to ensure the security of digital voting system voting system")會使用blockchain的概念, 你們可以講述一下為什麼採行不一樣的策略嗎? | ||
| E-06 14:20-14:30 |
Active Learning for 3D Human Pose Estimation: Combining the Strategies of Uncertainty and Diversity | 洪偉銓 | Human Pose Estimation的確是個很受重視的題目, 但也是因為如此有很多論文探討, 舉例來說:"Deep learning-based human pose estimation: A survey", 裡面有提到很多相關的策略, 因此你報告中提的策略, 其實只是一部分, 能說說為什麼不用其他策略?或是為什麼覺得這幾個你用的策略比較好嗎? | |||||
| E-07 14:30-14:40 |
Human-Machine Co-Creation Generating Imaginative Paintings from EEG Brain Activity | 蘇芊卉 | Firstly, the report is excellently crafted – great work! I'd like to clarify, did you simply implement the model architecture as suggested by Singh, without any modifications or enhancements? If that's not the case, could you specify what aspects were newly introduced? Additionally, how did these proposed improvements influence the final outcomes? | |||||
| E-08 14:40-14:50 |
Using brain identity and image feature to improve object detection and prediction | 黃筠蘋 | Why did you choose to adopt a linearizing encoding model? Are there specific reasons for this choice? Were there alternative methods available, and if so, why were they not used? Also, could you share what was the most challenging aspect of the project? | |||||
| E-09 15:10-15:20 |
網路危機文章偵測系統 | 陳致寧 | 這個資料集在兩年前已經有同學做過了, 做出來的正確率都不太好, 你覺得這樣正確率很低的方法還適合推廣嗎?另外, BERT是2019年的產品, 為什麼沒有使用現在比較新的方式例如GPT3.5 or 4? | |||||
| E-10 15:20-15:30 |
隨機多臂老虎機問題中以不同參數及環境對改良的LinUCB等演算法之實驗分析。 | 孫柏翰 | 如果用LinUCB Multi-armed Bandit problem當關鍵字去Google Scholar查, 可以查到很多LinUCB的改版, 想請問你有比較過你的改善方法和其他改版的差別嗎?為什麼會這樣設計呢? | |||||
| E-11 15:30-15:40 |
基於無參考影像之渲染影像鋸齒瑕疵偵測 | 吳宜宸 | 邱向辰 | 你們在專題上的最大貢獻是提出一個不一樣的標註方式嗎?如果是, 為什麼它能減低主觀判斷帶來的不一致性呢? | ||||
| E-12 15:40-15:50 |
Recommendation System with Graph Neural Network | 張惇媛 | 陳宏致 | Why do you still believe that temporal information is valuable despite noting that it doesn't markedly enhance FIRE's overall performance? What elements of your design strategy make it effective? Do you have any underlying intuition or reasoning for this approach? | ||||
| E-13 15:50-16:00 |
Urban Digital Twins: A Comparative Study | 賴政嘉 | You identified two primary challenges in urban development and management: infrastructure configuration in the planning stage and resource allocation in the operational phase. Subsequently, you listed four additional challenges: Object Mirroring, Data Availability, Object Relation, and User Interaction. In the following section, you introduce further challenges specific to implementation. However, the connection between these sets of challenges is not clear, leading to confusion. Could you explain how they are interrelated? Additionally, your evaluation results lack numerical data. Could you provide these figures to demonstrate the performance differences? | |||||
| E-14 16:15-16:25 |
LayoutStudio | 戴誌宏 | It appears that you collaborated with another graduate student on this project. Could you specify which specific components you worked on or designed? What was your specific role in the team? | |||||
| E-15 16:25-16:35 |
Optimizing Serverless Workflows with Local Caching | 林劭軒 | 葛奕宣 | 吳展維 | 蔡芝泰 | It seems your significant contribution is the development of the "two-speedup path architecture." How does this differ from the current state-of-the-art (SOTA) in the field? Could you also clarify what the SOTA methodologies are at present? Additionally, what specific challenges did you encounter in designing this architecture, and how complex was the process? In the evaluation section, you mention comparisons with an "original," but there's a lack of details about this benchmark. Could you explain what this "original" refers to? | ||
| E-16 16:35-16:45 |
AI橋牌_合約橋牌手牌預測模型及應用 | 宋育峻 | 你們的方法和其他人的方法差別在哪裡? 根據目前的文件, 無法看出來。報告中說『本次主要是針對上次沒完成的功能與許多錯誤部分做修改與改進。』上次是指哪一次？所以這些錯誤部分怎麼影響到最後結果呢？從實驗結果上也沒看到這些內容 | |||||
| 分組報告組別：F組(報告地點：台達629會議室) | ||||||||
| F-01 13:30-13:40 |
Customize your art souvenir on edge devices in 1.5 sec | 劉姿伶 | ||||||
| F-02 13:40-13:50 |
無人機手勢辨識及追蹤系統 | 蔡竣亦 | 廖奕愷 | 陳璽正 | ||||
| F-03 13:50-14:00 |
衛星換手方法模擬 | 陳柏翔 | 郭柏均 | |||||
| F-04 14:00-14:10 |
以啟發式演算法及強化學習求解機台排程問題 | 林奕廷 | 簡楷恒 | 林芷儀 | ||||
| F-05 14:10-14:20 |
HOW TO CHOOSE YOUR BACKBONE FOR FINE-GRAINED IMAGE RECOGNITION | 徐美妮 | 王文強 | 吳家錫 | 張箕淳 | |||
| F-06 14:20-14:30 |
演算法與程式設計訓練 | 黃昱嘉 | 李昕威 | |||||
| F-07 14:30-14:40 |
基於Transformer Encoder之中文社群負面文句辨識及應用 | 楊子翰 | 王采葳 | |||||
| F-08 14:40-14:50 |
高階合成之神經網路加速器於影像辨識模型的實作與分析 | 陳家輝 | 黃莉婷 | 潘鏡評 | 潘勝元 | 蔡政穎 | ||
| F-09 15:10-15:20 |
Interactive Coding Tutor with GPT3.5 | 李緯倫 | 江咏宸 | 許淳 | 林宗翰 | 黃志偉 | ||
| F-10 15:20-15:30 |
SIMD Everywhere Optimization from ARM NEON to RISC-V Vector Extensions | 蘇勇誠 | 朱季葳 | 郭禮德 | 文逸雲 | |||
| F-11 15:30-15:40 |
智慧插座大數據分析 | 丁旭寬 | ||||||
| F-12 15:40-15:50 |
Sorting Unsigned Sequence by Symmetric Reversals | 蘇裕恆 | ||||||
| F-13 15:50-16:00 |
背景音樂自動化生成 | 陳禹勳 | ||||||
| F-14 16:15-16:25 |
Application of Seccomp user-space notification in Security | 楊東翰 | ||||||
| F-15 16:25-16:35 |
高頻交易之加密貨幣幣種間低風險插針捕捉策略與系統整合 | 陳佑祥 | ||||||
| F-16 16:35-16:45 |
Physical Zero-Knowledge Proof for Killer Sudoku | 張庭嫣 | ||||||
重要說明：
- 每分組評審所提供的Q&A問題內容，請見該欄位敘述。若Q&A欄位顯示為空白的隊伍，於報告時間的10分鐘內，預留最後3分鐘，由師長現場提問；評審已先預提供Q&A內容之隊伍，可在簡報報告時間10分鐘內，先製作Q&A回覆於簡報內容中，並預留最後3分鐘答覆問題。
- 專題成果報告當天，請各隊在所指定的報告時間，提前10分鐘到達報告之場地。
- 每組專題學生之成果報告解說評分佔專題總成績50%，指導教授給予之分數佔總成績50%。
本學期修讀系統整合與實作二”之學生皆需參與專題成果報告，請12/06(三)當天務必準時出席。
清華大學資工系辦公室 2023.11.27.
瀏覽數: