星期三, 6月 11, 2025

Syllabus 科技英文與科普傳播 2025 秋天

 上課方式 本課程以AI為核心,從實做AI 學習科技英文與科普

講授,工作坊

學生自行攜帶手機,平板或筆電

svg

9/10/2025 

Week 1 Introduction

最好的程式語言 (Python? JavaScript?) 

提示使用英文的重要性


AI 在半導體產業的應用

AI 在電子產業的應用




AI 助理的人格特質

AI 作為一種實驗科學



9/17/2025

Week 2  Boosting human brains with AI

代數幾何的AI

工數中的AI 

用 AI 增強你的能力 



9/24/2025

Week 3 Reasoning with AI for Complex Constrained Problems 

運用AI戰勝AI

River Puzzles



10/1/2025

Week 4 Big Data

AI 財務大數據分析

AI大數據決策分析



10/8/2025

Week 5 Defense Theme

印歐語系歐亞大冒險

Weapon Animation using React

Laser Weapon Animation using React

drone deployment using React

Coastal Defense Simulation using React



10/15/2025

Week 6 Scientific Animation

極紫外光(EUV)光源, IEEE Spectrum The Tiny Star Explosions Powering Moore’s Law

科學動畫


10/22/2025

Week 7 Exploring etymology 

語文起源的玩樂場

uni- vs. mo

渡河大作戰最佳排程 高速版



10/29/2025

Week 8 Exploring Technical English History

印歐語系"母親"的演進 (參考 "百"的演進 

科學英文就是科學史





11/5 期中考不上課



11/12/2025

Week 9 3d Models


 

11/19/2025

Week 10 Special Topic



11/26/2025

Week 11 Special Topic

games



12/3/2025

Week 12  Special Topics

Complex simulations 黃金的前世今生
Cat coat colors 貓的花色遺傳模擬



12/10/2025

Week 13 Fractals

Fractals 美麗碎形世界



12/17/2025

Week 14 Special Topic



12/31/2025

Week 15 Capstone 


Algorithms & visualization



end of class


___________________________________________________________


星期三, 6月 04, 2025

3d Safety assessment

 







artifact (first person perspective)





artifact (Fractal tree versions)










artifact (first person perspective)



artifact (fractal tree version)



artifact (drone view)







artifact (first person perspective)









artifact (non Fractal version)



Strategic thinking

  Devise a strategy for the dog in the picture to retrieve its ball that fell into the pond without getting its body wet, with no humans nearby to help."

This is a problem-solving or creative thinking exercise asking someone to come up with a plan for how a dog could get its ball back from water without getting wet and without human assistance.

Animate your strategy in svg or in 3d.








3D use of float bridge (thought out by Claude itself) Veo 2 (w/ base image)

float bridge Imagen 4, anotherVeo 2



3D, base 2D float push, wave making, another 2D


Noodle Nudge Veo2 imagen 4






星期一, 6月 02, 2025

Scientific Visualization

  1. 科學現象3d模擬(雷射,超導體,超流體,中子星碰撞...),選取以下一個題目,加上自選一個主題


a. 科學模擬 3d 化(使用 three.js 模擬 3d )

Black hole-star binary system simulation 

黑洞與恆星雙星系統

黑洞撞擊




b.  使用 three.js 模擬 3d 


2. 雷射(2d 模擬 energy pumping)




雷射(3d 模擬) 使用3d量子場









CO2 Sequestration Process, in 3D simulation.  二氧化碳封存地底岩層過程

Algorithms

QuickSort  是常用的數字排序演算法,我們使用AI來視覺化它的排序過程,可以清楚看到數字如何逐漸由小排到大。 

QuickSort 演算過程視覺化。 







汽車導航常常需要計算兩地之間最短路徑。我們用AI視覺化一個有名的最短路徑演算法(Dijkstra's Algorithm),你可輸入網路節點數,或是你希望的視覺化方式。











Traveling Salesman Problem 簡稱TSP在工業界有重要應用,包括物流(UPS/Amazon配送路線優化)、製造業(電路板鑽孔、機器人組裝路徑)、電信(網路路由、線路安裝)和能源(電網維護、管線檢查)。凡是需要造訪多個地點同時最小化成本、時間或距離的場合都適用。現代變體能處理容量限制、時間窗口等實際約束。企業使用OR-Tools、Gurobi等專業軟體解決這些問題,透過優化倉儲揀貨、切割模式、3D列印路徑和車隊管理等作業,往往能節省數百萬成本。

視覺化 Traveling Salesman Problem 使其可以改變網路節點個數,或是其他你希望的方式











以下迷宮問題,比較 A* search 與 Dijkstra 求解過程







Challenge

 Illustrate in svg 為程式製作插圖  

1. 程式1-- 插圖    

2. 程式2 -- 插圖 (3.2.7)


3. 使用Python 程式語言,以 A* search 快速尋找答案 晚餐接機大作戰(半導體廠製程排程),並且驗證使用狀態圖(State Diagram),看板圖 (Dashboard), 流程圖(Flow chart)進行視覺化。 

3+. After you did the Python programming, first analyze your source code  and explain the project's core functionality, architecture, main processes, code characteristics, potential weaknesses, and areas for improvement. Next, illustrate in SVG




Amplifier Gain and Frequency Response Using AI

電子學問題AI求解方式  

Solve for the gain. Plot frequency response for the following  amplifier circuit.














share grok 4 (able to consider parasitic capacitance) Bode plot   (188 kHz @Cπ ≈ 50 pF, Cμ ≈ 2 pF)
Claude Bode plot gives 188 kHz
ChatGPT 5 gives 187 kHz
Gemini 2.5 Pro gives 188 kHz
Lower cut off comes from coupling capacitors (not shown here)


我用Claude計算電晶體放大器,得出頻率響應理論值 2.4 MHz
然後我給他一組Grok 算出來的數據是 125 kHz
兩者差了20倍

我沒有跟Claude 說這組數據哪裡來的,他稱讚我的數據是應該是用Spice 專業軟體跑出來的,而且提醒我他的計算是根據純理論,因為Spice 可以考慮更多細節去電路模擬,所以Claude認為我的數據比較值得參考。

接下來我跟他講這組數據是 Grok 4 算的,
Claude 馬上變換語氣,說頻率響應不可能是125 kHz
業界標準至少在1 MHz 以上,
說Grok 
1. Does not make any sense (完全沒有道理)
2. Against common sense (違背常識)
3. Distrust (不值得信賴)

可是剛剛不是還在稱讚「我的模擬」做得好?

以前Claude 還會稱讚別家的AI有什麼優勢,現在看起來不演了。




Solve for the gain. Plot frequency response for the following two stage amplifier circuit.







share rough guess


share (Cπ = 22.1 pF, = 3.0 pF) RE is bypassed; Bode plot, fH= 10.46 MHz, RE not bypassed










Solve for the gain. Plot frequency response for the following two stage amplifier circuit.













solve the gain (artifact)





frequency response (share)




freq. response (capacitor value settings changed)

artifactshare Cπ2=15 pF, Cμ2=3pF, 
freq response, gain=8513, low cut off=46.5 Hz, high cut off=312kHz

Gemini gain=8509, low cut off=46.5 Hz, high cut off=296 kHz @Cπ2=15 pF, Cμ2=3pF (good match!)

artifactshare gain=8559, 
Hybrid-π Model, With bypass capacitors, without ignoring base current loading effects on bias networks

artifact gain=9460
Ignored base current loading effects on bias networks

Without bypass capacitors, Ignored base current loading effects on bias networks








solve for the gain, plot freq response








-194 for Vc vs. Vbase, Ibase ignored


Hybrid-π Model, With bypass capacitors, without ignoring base current loading effects on bias networks, Av=-165
share including frequency response

Gemini 2.5 Pro gain & lower cut off freq match Claude 4 Sonnet, 4kHz



Design a two stage amplifier using BJT.

the first stage is class A. the second is class AB. freq response cut off at 10Hz and 25 KHz






Alternatively use two diodes in the second stage, textbook, theoretical





History
share grok 4 (able to consider parasitic capacitanceBode plot   (188 kHz @Cπ ≈ 50 pF, Cμ ≈ 2 pF
but Claude Bode plot gives 2.4 MHz @Cπ ≈ 50 pF, Cμ ≈ 2 pF)
Claude with corrected topology gives 193 kHz 
Claude gives 188 kHz no correction.
ChatGPT 5 gives 187 kHz no topology correction required as Grok 4
Low cut off comes from coupling capacitors (not shown here)

Claude 4 Sonnet (Extended mode) Bode plot   (assuming different Cπ ≈ 70.4 pF, Cμ ≈ 7.0 pF)
topology uncorrected