DEVELOPMENT AND APPLICATION OF AI

What is AI (Artificial intelligent)

• AI refers to the simulation of human intelligence processes by machines, especially computer systems.
• The form of existence for AI can be
• Software
• Machine learning
• Deep learning
• ChatGPT
• Hardware
• Robot

Quotes of AI

• Elon Musk on AI's Existential Threat: "With Artificial Intelligence, We Are Summoning the Demon“
• "Once humans develop artificial intelligence, it would take off on its own and redesign itself at an ever-increasing rate. Humans, who are limited by slow biological evolution, couldn't compete, and would be superseded.“ from Stephen Hawking

Quotes of AI

• “AI will affect every product and every service that we have.” – Tim Cook, the CEO of Apple
• “Artificial intelligence will have a more profound impact on humanity than fire, electricity and the internet.” – Sundar Pichai, the CEO of Alphabet
• “AI is like electricity. Just as electricity transformed every major industry a century ago, AI is now poised to do the same.” – Andrew Ng, the founder of Google Brain and former Vice President of Baidu
• Mark Zuckerberg Envisions an AI Assistant: "Jarvis, My Virtual Assistant, Manages My Life“
• Satya Nadella Sees AI's Potential: “AI Will Make Us More Human, Not Less”
• Bill Gates Predicts AI's Societal Impact: “AI Can Be Our Friend”

​

Companies

Machine learning

• Machine learning (ML) is the study of computer algorithms that can improve automatically through experience and by the use of data
• Machine learning algorithms build a model based on sample data, known as training data, in order to make predictions or decisions without being explicitly programmed to do so
• Data mining, optimization, generalization, statistics

Framework of machine learning

Training data

Training

Model

Testing data

New query

Output

Black Box

Training and testing

• Training set is for feeding the information (features) to the machine for building the model
• Testing set is for evaluating the accuracy of the built model, it should be totally separated from the training set
• The model is specific based on the selected features
• If the other features may influence the new query, the training needs to be re-initiated with new such new feature

Machine learning types

• Supervised learning
• Need known labels for the input data
• Majority for classification
• Unsupervised learning
• Do not need the labels
• Majority for data grouping
• Reinforcement learning
• Learn from mistakes
• Reward-based questions

Machine learning types

• Supervised learning

Machine learning types

• Unsupervised learning

Machine learning types

• Reinforcement learning

Patient classification

A predictor assigns a new case to an existing group

​

​

​

​

​

​

​

Protein expression

Train

Predictor

Use predictor to classify new cases

Predictor

Proteins that best distinguish the groups of patients

​

​

​

​

​

​

​

Use predictor to find a prognostic signature

Protein expression

Train

Predictor

Protein A

Protein B

Predictor

Data balance V.S imbalance

• Balanced Dataset — In our dataset, positive values which are approximately same as negative values
• Imbalanced Dataset — High different between the positive values and negative values

Over-sampling (Up sampling)

• Balance by increasing the size of rare samples
• Advantages
• No loss of information
• Mitigate overfitting caused by oversampling
• Disadvantages
• Overfitting

Under-sampling (Down Sampling)

• Balance by reducing the size of majority samples
• Advantages
• Run-time can be improved by decreasing the amount of training dataset.
• Helps in solving the memory problems
• Disadvantages
• Losing critical information

Overfitting

• Definition – the production of an analysis that corresponds too closely or exactly to a particular set of data, and may therefore fail to fit additional data or predict future observations reliably
• The situations may cause over fitting
• Data imbalanced
• Over training

Feature selection

• Need to select the features which is highly related with the target question
• Learning the wrong lesson
• an image classifier trained only on pictures of brown horses and black cats might conclude that all brown patches are likely to be horses
• Different features may contribute the predictions differently

Feature weight

• Some methods do produce the information of feature weighting, but some do not.

The input features should not have bias

Validation

• In order to measure the real accuracy of the built model, the validation needs to be done.
• Methods
• K-fold cross validation
• Leave-one-out cross validation (Jackknife)

K-fold cross validation

Hyperparameter optimization

Hyperparameter optimization

Hyperparameter optimization

Hyperparameter optimization

Hyperparameter optimization

Bayesian optimization

Bayesian optimization

3/10 ?

Bayesian optimization

3/10 ?

Deep learning

• Optimization of feature selection
• Order of sequence can be included
• Required large dataset

Recurrent neural network

• The most important feature of RNN is Hidden state and they have memory which remembers each and every information through time
• Exploding gradient problem

Applications

Global AI market

Global AI market

Combination of AI and wearable technology

• These devices will provide personalized usage for many applications, such as health monitoring, real-time translation.
• With the integration of AI technology, wearable devices will no longer be merely passive display tools for information, but rather become intelligent devices capable of dynamic interaction with users.
• With the support of AI technology, wearable devices will also demonstrate tremendous potential in augmented reality (AR) and virtual reality (VR). Users can experience highly immersive interactions and entertainment activities through these devices, further driving the application of AR and VR technologies in daily life.

Applying to professional fields

• AI can be trained to specialize according to the data, regulations, and practices of specific industries, thereby providing more professional and accurate recommendations.
• In the healthcare sector, specialized generative AI can learn from vast amounts of medical data and clinical research to assist doctors in disease diagnosis, treatment planning, and prognosis.
• In the legal field, specialized generative AI can analyze extensive case files and legal documents to provide legal research, case analysis, and judgment references for lawyers and judges.
• AI is its ability for continuous learning. Over time and with the accumulation of more data, this AI can evolve and refine itself, thereby providing more accurate and real-time recommendations and analyses.

The future of AI

• Decentralized AI (分散式)
• Create a system that can share data and computing resources among multiple nodes, including personal computers, mobile devices, sensors, and cloud servers.
• This approach enables artificial intelligence systems to operate more flexibly and efficiently, while reducing the risk of single points of failure and helping to protect user data privacy.
• By leveraging distributed computing technology, artificial intelligence systems can better adapt to changing environments and more effectively address the demands of large-scale data processing and analysis.

The future of AI

• Generative AI (生成式)
• Aimed at generating new data, images, text, or other types of content, rather than just analyzing or classifying existing data.
• Using deep learning models to learn the distribution of data and generate new data similar to it.
• Generative AI can be used in various applications including image synthesis, text generation, music composition, video generation, and more.
• The development of generative AI is significant for many fields as it opens up new possibilities for artistic creation, content generation, virtual reality, simulation, and more. It has wide-ranging applications in areas such as entertainment, design, healthcare, security, and beyond.

The future of AI

• Trusted AI (可信式)
• Refers to emphasizing the characteristics of credibility, transparency, and accountability when designing, developing, and implementing artificial intelligence systems.
• The black box is one of the weak points in AI development. With the proliferation of AI applications, enterprises are also beginning to evaluate the problems associated with AI adoption, such as network security, personal safety, legality, explainability, privacy protection, data bias, fairness, diversity, and negative environmental impacts.

The future of AI

• Sustainable AI (永續式)
• Refers to the emphasis on ensuring the sustainability of artificial intelligence technology in social, environmental, and economic aspects when designing, developing, and implementing it.
• This includes considering the impact of AI systems on resource usage, energy efficiency, environmental footprint, social equity, economic development, etc., and taking measures to minimize negative impacts while maximizing the long-term sustainability of AI technology.

Challenges and concerns

• Copyright Issues
• AI technologies, especially generative AI, often utilize large amounts of data from the internet during training, including copyrighted content. This raises the question: does the content created by AI constitute infringement of the original works?
• Privacy Protection
• When AI applications process personal data, particularly in areas such as facial recognition and behavior prediction, they may unintentionally infringe on individual privacy.

Challenges and concerns

• Ethical Standards
• Whether AI decision-making processes are biased and whether they adhere to societal ethical standards, issues that need to be taken seriously, especially for medical diagnosis and financial services.
• Legal Adjustments
• Existing legal frameworks may not fully adapt to the new challenges of the AI era.

Challenges and concerns

• Deepfake
• In the political arena, deepfake technology may be utilized to create false information, influencing elections and discussions on public policies.
• The emergence of false images and audio could erode trust among people, undermining social structures and ethical foundations.
• In the realm of military operations, deepfake technology could potentially alter the course of warfare entirely.

​

Job replaced by AI ?

如何不被AI取代？張忠謀給年輕人的4個建議

Prepare and invest yourself

• Be professional
• Keep learning
• Be creative
• Humanity
• Develop special skills

Use it wisely

Neural Network (NN)

• A collection of connected units or nodes (artificial neurons) which loosely model the neurons in a biological brain
• Neurons and edges typically have a weight that adjusts as learning proceeds
• Different layers may perform different transformations on their inputs
• Regression model

Random forest

• Make many sub-decision trees
• Each tree works independently to form their own output and give a prediction
• Take the prediction from each tree and selects the majority of the class each tree predicted as the true predicted class of the dataset

Support vector machine (SVM)

• Separates two categories of data in the dataset with the help of a hyperplane

​

​

Principal component analysis (PCA)

K-means