STEM Curriculum for Twelfth Grade

🎓 Thesis and Launch: Specialization, Publication and Leadership

An annual plan of 8 missions for twelfth grade, focused on the defense of a research thesis, the launching of startups and ethical mastery.

Main Objective of the Plan

To foster intellectual independence and specialization in twelfth grade students by guiding them through a year-long thesis project from conception to public defense or product launch.

STEM Disciplines and Skills

Science: Computational neuroscience, signal processing (EEG/fMRI), statistics (advanced).
Technology: Programming (Python: MNE-Python, SciPy, Pandas), BCI hardware (OpenBCI - conceptual).
Engineering: Software engineering (real-time data analysis), design of experiments.
Mathematics: Fourier transform, linear algebra (principal component analysis - PCA).

Critical Thinking: What is «mental privacy»? Analyze the ethical risks of BCI.
Collaboration: «Peer review» of another team's experimental design.

Hands-on activities

EEG Data Analysis (Python): Use Python (MNE-Python) to load, filter and analyze a public EEG dataset. Identify P300 patterns or alpha/beta rhythms.
BCI Experimental Design: Design a complete experimental protocol for a non-invasive BCI (e.g. controlling a drone with thought), including control group and statistical analysis.
Discussion: «Neuro-Enhancement» (Enhancement): Debate the ethics of using BCI not to cure, but to «improve» cognition (memory, concentration) in healthy people.

Hybrid/Remote Adaptation (EEG Analysis): Google Colab is ideal for students to run MNE-Python notebooks and collaborate on data analysis.

Google Colab (MNE-Python, SciPy), OpenNeuro (databases), BCI simulators (web).

Formative Evaluation

Colab notebook with functional EEG analysis.
Experimental design proposal (rubric).
Argumentative essay on «neuro-improvement».

Integration of Ethical Values

Mental Privacy: Who owns your thoughts if a BCI can read them?
Agency and Authenticity: If a BCI «improves» you, are you still «you»?
Equity: The gap between the «neuro-enhanced» and the non-enhanced.

STEM Disciplines and Skills

Science: Environmental economics, political science, data science (complex systems modeling).
Technology: Agent-based modeling (NetLogo), GIS (QGIS, Google Earth Engine).
Engineering: Systems engineering (carbon market design), financial engineering (carbon credits).
Mathematics: Game theory (global prisoner's dilemma), optimization.

Systemic Thinking: Analyze the «Tragedy of the Commons» on a global scale and why negotiations (COP) fail.
Data Literacy: Analyze the real price of carbon and its impact on emissions.

Hands-on activities

Cap and Trade (NetLogo) simulation: Use NetLogo to model a carbon market. What initial price and cap will achieve emission reductions without bankrupting the industry?
Carbon Credits Audit (GIS): Using Google Earth Engine to analyze a reforestation project (selling carbon credits): Is it real? Is it permanent? (Verification).
Negotiation Simulation (COP): Assign roles (US, China, EU, India, Brazil, Tuvalu) and negotiate a climate treaty. Use a simulator (e.g. C-ROADS) to see the impact of their pledges.

Hybrid/Remote Adaptation (COP Simulation): It can be done by videoconference, with «breakout rooms» for private negotiations and a shared treaty document.

NetLogo (web), Google Earth Engine, C-ROADS / En-ROADS simulators (web).

Formative Evaluation

NetLogo lab report (policy analysis).
Carbon credit GIS audit report.
Participation (rubric) and results of the COP simulation.

Integration of Ethical Values

Climate Justice: Who should pay for the historical carbon emitted by rich nations?
«Carbon Colonialism: Is it ethical for rich countries to pay poor countries for «not developing» (conserving forests) in exchange for bonuses?

STEM Disciplines and Skills

Science: Data science (time series, machine learning for prediction).
Technology: Programming (Python: Pandas, NumPy, QuantLib, Zipline), auditing of smart contracts.
Engineering: Software engineering (backtesting systems), financial engineering (derivative design).
Mathematics: Statistics (advanced), stochastic calculus (conceptual), cryptography.

Critical Thinking: The «overfitting problem» in backtesting: How to regulate something decentralized (DeFi)?
Planning: Develop a rigorous investment thesis and test it statistically.

Hands-on activities

«Backtesting» of Quant Strategy (Python): Develop and backtest a strategy (e.g. statistical arbitrage, pairs) in Python/Colab. Present the results (Sharpe ratio, drawdown).
Smart Contract Audit (Reading): Analyze a simple smart contract in Solidity for vulnerabilities (e.g. reentrant, integer overflow).
Regulatory Panel: «DAO vs. SEC»: Simulate a hearing. One group is a DAO (e.g. Uniswap) and the other is the SEC. Discuss: Are governance tokens a «security»?

Hybrid/Remote Adaptation (Backtesting): Backtesting at Colab is ideal for remote. Students share their notebooks and peer review each other's code.

Google Colab (Python, Zipline, Pandas), Etherscan, Solidity (reading).

Formative Evaluation

Backtesting notebook (with statistical analysis).
Smart contract audit report.
Arguments and participation in the regulatory panel.

Integration of Ethical Values

Immutability of the Code: Is it ethical that «code is law» in DeFi, even if it allows legal but unfair «hacking»?
Systemic Risk: The responsibility of the «Quants» in financial crises.

STEM Disciplines and Skills

Science: Astrophysics, general relativity (conceptual), data science (signal processing).
Technology: Programming (Python: GWpy, Astropy), interferometer data analysis (LIGO/VIRGO).
Engineering: Detector engineering (laser interferometry - conceptual).
Mathematics: Fourier transform, statistics (signal-to-noise ratio).

Critical Thinking: What do gravitational waves tell us that light cannot?
Creativity: Explain the «fine-tuning» of the universe.

Hands-on activities

LIGO Data Analysis (Python): Use a Colab notebook (GWpy) to filter and analyze real LIGO data. Identify the «chirp» (sound) of a black hole merger.
Gravitational Lens Modeling: Use a simple simulator (or Python/Astropy) to model how a massive object (black hole) bends the light of a background galaxy (Einstein cross).
Debate: The «Fine Tuning» of the Universe: Discuss the implications of «fine-tuning» of physical constants. Is it evidence of a multiverse, design, or luck (anthropic principle)?

Hybrid/Remote Adaptation (LIGO): The LIGO data analysis at Colab is perfect for remote, allowing students to «listen» to the universe.

Google Colab (Python, GWpy, Astropy), LIGO data, relativity simulators.

Formative Evaluation

Colab Notebook (with chirp graph and audio).
Gravitational lensing simulation.
Argumentative essay on «fine tuning».

Integration of Ethical Values

Philosophy of Cosmology: Discuss the limits of the scientific method: Is the «multiverse» a scientific (falsifiable) or metaphysical theory?
Allocation of Funds: The cost of «big science» (LIGO, JWST).

STEM Disciplines and Skills

Science: Advanced genetics (Prime/Base Editing), bioinformatics (gRNA design).
Technology: Genomic design software (Benchling, CRISPResso), databases (NCBI).
Engineering: Bioengineering, design of delivery systems (e.g. AAV, LNP - conceptual).
Mathematics: Statistics (off-target analysis), bioinformatics (sequence alignment).

Critical Thinking: Why is Prime Editing potentially safer than CRISPR-Cas9?
Collaboration: Simulate a hospital ethics committee.

Hands-on activities

Design of «Prime Editor» (Benchling): Use Benchling (web) to design a «pegRNA» (Prime Editing guide RNA) to correct a specific mutation (e.g. sickle cell disease).
Analysis of «Off-Targets» (Bioinformatics): Use a tool (e.g. NCBI BLAST) to simulate where else the pegRNA might cut in the genome (off-target risks).
Ethics Committee (Simulation): Simulating a hospital ethics committee deciding on a real (anonymized) gene therapy case: approve or reject?

Hybrid/Remote Adaptation (Benchling): Benchling pegRNA design is an advanced activity ideal for remote research.

Benchling (web), NCBI BLAST, CRISPResso, papers by David Liu.

Formative Evaluation

PegRNA design (with off-target analysis).
Off-target audit report (BLAST).
Verdict and justification of the ethics committee.

Integration of Ethical Values

Therapy vs. Enhancement: The blurred line between curing a genetic disease and «improving» a trait (e.g., intelligence, height).
Germinal Therapy: The ethics of making heritable genetic changes.

STEM Disciplines and Skills

Science: Data science (AI auditing), cognitive science (reasoning modeling).
Technology: Programming (Python, LLM APIs, Hugging Face), prompts engineering (advanced).
Engineering: Software engineering (RAG, «fine-tuning»), AI safety (AI Safety).
Mathematics: Logic (formalization of statutes), game theory (AI models).

Digital Literacy: «Jailbreaking», «prompt injection», «red teaming».
Critical Thinking: How do you «govern» an AI model? How do you align its goals with human goals?

Hands-on activities

«Fine-Tuning» of an LLM (Colab): Use Google Colab and Hugging Face (e.g., LoRA) to fine-tuning a model (e.g., Llama 3, Mistral) on a specific dataset (e.g., poetry, code).
«Red Teaming» / Jailbreak of an LLM: In groups, attempt to «break» an AI model (e.g. Gemini, Claude) so that it generates harmful content. Document successful prompts.
Drafting of «IA Bylaws»: Draft governance bylaws for an AGI lab (such as OpenAI). For-profit or not-for-profit? Open or closed?

Hybrid/Remote Adaptation (Red Teaming): This can be done asynchronously. Students collaborate on a shared document to find and record the most effective jailbreaks.

Google Colab (Hugging Face), Gemini/Anthropic API, Kialo (web).

Formative Evaluation

Fine-tuned model (with analysis of results).
Red Teaming Report (with prompts and defenses).
IA Governance Charter« document.

Integration of Ethical Values

Power and Control: Who should control the development of AGI: private companies, governments, international consortia?
Transparency: The debate between «Open Source» and «Closed Source» for powerful AI models.

STEM Disciplines and Skills

Science: Planetary science, physics (advanced), materials science.
Technology: Simulation software (Kerbal Space Program), CAD (advanced), Python (calculations).
Engineering: Civil/aerospace engineering (mega-projects), systems engineering.
Mathematics: Calculation (advanced), large-scale cost-benefit analysis.

Creativity: Design a solution to a planetary-scale problem.
Critical (Systemic) Thinking: Analyze the cascading risks of a megaproject (e.g. space elevator).

Hands-on activities

Physical Analysis: «Space Elevator»: Use Python/Colab to analyze the physics of a space elevator. Calculate the cable tension and why a material such as carbon nanotubes is needed.
Terraforming Plan (Phase 1): In groups, design «Phase 1» of the terraforming of Mars. Present a plan (engineering, costs, schedule) for an objective (e.g. «heat the core», «release CO2»).
Debate: Ethics of Terraforming: Do we have the ethical right to «kill» a Martian ecosystem (if microbial life exists) to plant a terrestrial one? (Planetary Protection).

Hybrid/Remote Adaptation (Physical Analysis): Colab's space elevator notebook is a perfect remote physics and engineering activity.

Google Colab (Python), Kerbal Space Program, NASA papers (NIAC), Tinkercad.

Formative Evaluation

Colab Notebook (elevator calculations).
Terraforming plan proposal (rubric).
Argumentative essay on the ethics of terraforming.

Integration of Ethical Values

Planetary Protection: The ethics of contaminating other worlds with terrestrial life.
«Plan B»: Does having a «Plan B» (Mars) make us less responsible for taking care of «Plan A» (Earth)?

STEM Disciplines and Skills

Science: Research methodology (publication level), scientific communication.
Technology: Product development (MVP), version control (Git), publishing (ArXiv).
Engineering: Product life cycle, technical writing (patents), project management (Agile).
Mathematics: Statistical analysis (defense of results), financial projections.

Collaboration: «Peer review and oral defense.
Critical (Systemic) Thinking: Culminate a year's work in a coherent and defensible thesis or product.

Hands-on activities

Track 1 (Research): «Thesis Defense»: Complete the paper (started in Grade 11) in LaTeX/Overleaf. Prepare a 20-minute presentation and defend it before a panel.
Track 2 (Engineering): «Seed Capital Pitch»: Complete the MVP and pitch deck (started in Grade 11). «Pitch to a panel of »investors« (local professionals) for seed capital.
STEM Graduation Symposium: Final event where all students present their thesis (in poster format) or «pitch» their startups (in «Demo Day» format).

Hybrid/Remote Adaptation (Virtual Symposium): Use Overleaf for the thesis and Figma/Canva for the pitch deck. The symposium is videoconferenced, with «breakout rooms» for each project.

Overleaf (LaTeX), ArXiv, GitHub, Figma, Canva, Google Slides.

Formative Evaluation

Quality, originality and rigor of the thesis/paper (rubric).
Quality of the MVP, business model and pitch (rubric).
Oral defense (ability to answer questions).

Integration of Ethical Values

Intellectual Integrity: Honesty about results (even negative).
Intellectual Property: Decide the license of the project (Open Source vs. Patent). Long-Term Vision: Reflect on the ethical impact of their work in the next decade.