### *Education on the Horizon: Teaching and Learning with Technology*
## Week 8: Personalized and Adaptive Learning with AI Pt. 1
===
#### Dr. Yizhen (Eejain) Huang
###### Department Erziehungswissenschaft, University of Potsdam
---
### Have you experienced these in your everyday life?
- YouTube suggests new videos based on what you have watched
- Your smart phone can automatically create an album with photos with your face in it
- You have to prove "I'm not a robot" during login
- These are possible because of “artificial intelligence” (AI)
---
### What is artificial intelligence (AI)?
- AI are intelligent computer programs that mimic the problem-solving and decision-making capabilities of the human mind
- All the current AI applications are narrow and weak:
- solves a specific kind of problem: e.g., autonomous car, ad optimization, chatbot
- only acting intelligently in a few areas (pattern recognition and generation), not "thinking" intelligently (self-conscious mind)
<!-- - Traditional computer program: Input → Output is predetermined -->
<!-- - AI: Input → Output is conditional -->
<!-- - training data → algorithm → real data -->
---
### Key features of AI
- Autonomy: The ability to perform tasks in complex environments without constant guidance by a user.
- Adaptivity: The ability to improve performance by learning from experience.
---
### Why is it "intelligent"?
- Learning is the acquisition of associations: what input leads to what output
- AI can make inferences about new inputs based on learned associations
- Therefore AI is capable of "learning" based on "past experience", similar to human decision making
- [Computer Scientist Explains Machine Learning](https://www.youtube.com/watch?v=5q87K1WaoFI)
---
### What is ML?
- Machine learning (ML) is a subfield of AI, which deals with building systems that improve their performance in a given task with more and more experience or data.
---
### Two types of ML
- supervised learning: use labeled data as examples to achieve classification and prediction, e.g., identify fake Twitter bots, recognize handwritten numbers
- unsupervised learning: use unlabeled data to reveal hidden structures and generate output similar to input, e.g., write in the tone of Shakespeare, generate image based on text prompt
---
---
---
### Some interesting AI demos
---
## Generative Text
<!-- - [AI writes news](https://talktotransformer.com/) -->
- [AI answers any questions](https://chat.openai.com/chat)
<!-- - [AI writes your email](https://app.flowrite.com/write) -->
---
## Generative Graphic
<!-- - [AI draws landscape](http://gaugan.org/gaugan2/) -->
<!-- - [AI draws anything based on your text](https://openai.com/blog/dall-e/) -->
- [AI draws anything based on your prompt](https://www.midjourney.com/showcase/recent/)
<!-- - https://stability.ai/blog/stable-diffusion-v2-release -->
- [AI video](https://imagen.research.google/video/)
<!-- - [Another competitor](https://imagen.research.google/) -->
<!-- - [AI creates art](https://app.wombo.art/) -->
<!-- - [Create anime styled selfies](https://waifu.lofiu.com/) -->
<!-- - [How about oil-painting styled selfies](https://ai-art.tokyo/en/) -->
<!-- - [Generative story](https://ai-adventure.appspot.com/) -->
<!-- - [AI guessing memes](https://twitter.com/ResNeXtGuesser/status/1477391701807300608) -->
<!-- - [more demos](https://www.nvidia.com/en-us/research/ai-demos/) -->
---
## Generative Voice and Sound
<!-- - [Deep fake voice](https://fakeyou.com/) -->
- [AI deep fake voice](https://beta.elevenlabs.io/speech-synthesis)
<!-- - [Another highly realistic voice generator](https://play.ht/blog/introducing-truly-realistic-text-to-speech-with-emotion-and-laughter/) -->
<!-- - [Generative music](https://www.nvidia.com/en-us/research/ai-art-gallery/artists/aiva/) -->
- [Generative music](https://www.riffusion.com/)
---
## Moving AI (virtual human)
- [AI talk to human](https://youtu.be/BpnnD_0IlbE?t=380)
- [AI talk to AI](https://www.youtube.com/watch?v=jz78fSnBG0s)
---
# Human or AI?
- https://app.humanornot.ai/