CS代考 VE445: Introduction to machine learning

Ve492: Introduction to Artificial Intelligence
Conclusion
Paul M-SJTU Joint Institute

Copyright By PowCoder代写 加微信 powcoder

What Have We Learned?
❖ Single agent, deterministic known model, fully-observable ❖ A* with admissible/consistent heuristics
❖ Multi-agent, known model, fully-observable ❖ Minimax, expectimax, expectiminimax
❖ Constraint satisfaction problems ❖ Backtracking, constraint graphs
❖ Single agent, stochastic known model, fully-observable ❖ Value iteration and policy iteration
❖ Decision theory and game theory

What Have We Learned?
❖ Single agent, deterministic known model, structured ❖ Propositional logic and first-order logic
❖ Single agent, stochastic unknown model, fully-observable ❖ (Approximate) Q-learning
❖ Probabilistic known model, partially observable ❖ Bayesian networks, Markov chain, HMM
❖ Classification, unknown model
❖ Generative models: Naïve Bayes, Bag of words
❖ Discriminative models: perceptron, logistic regression, neural networks

❖ Everyone is doing AI
❖ AI is more than deep learning
❖ Moving goalposts:
❖ AI: what a computer cannot yet do
What is AI?
Understanding
Decision- making
Knowledge representation
Perception

AI Research
❖ Deep learning
❖ What does deep learning work so well?
❖ => More robust, efficient, interpretable models
Loss landscape Adversarial example

AI Research
❖ Neuro-symbolic approaches
❖ How to combine deep learning with other AI techniques?
❖ => Understanding, reasoning
VQA Dataset CLEVR Dataset

AI Research
❖ (Deep) reinforcement learning ❖ How to make it more efficient?
Autonomous driving Robotics

Artificial General Intelligence?
❖ Narrow AI vs AGI
❖ Is AGI needed for any complex task?
❖ A recipe from :
❖ A Path Towards Autonomous Machine Intelligence

Should we be afraid of AI?

DARPA Robotics Challenge

What to Learn Next?
❖ VE445: Introduction to machine learning
❖ ECE4710J: Introduction to data science
❖ ECE4880J: Introduction to computer vision
❖ VE593: Solving problems with AI techniques
❖ VE693: Deep reinforcement learning

程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com