Lecture 11 - Tiny Machine Learning Flashcards
What is trustworthy AI?
Trustworthy AI should be:
* Lawful - respecting all applicable laws and regulations
* Ethical - respecting ethical principles and values
* Robust - both from a technical perspective while taking into account its social environment ( e.g fairness, inclusivity, alignment with social norms and values etc)
What are the EU Ethics guidelines for trustworthy AI?
- Human agency & oversight
- Technical robustness & safety
- Privacy & data governance
- Transparency
- Diversity, fairness & non-discrimination
- Societal & environmental wellbeing
- Accountability
What is human agency and oversight?
- AI systems should empower human beings, allowing
to make informed decisions and fostering their fundamental rights. - The allocation of functions between humans and AI systems should follow human-centric design principles and leave meaningful opportunity for human choice.
- At the same time, proper oversight mechanisms need to be ensured, which can be achieved through human-in-the-loop, human-on-the-loop, and humanin-command approaches.
What is technical robustness and safety?
- AI systems need to be resilient and secure.
- They need to be safe, ensuring a fall back plan in case something goes wrong.
- They need to be accurate, reliable and reproducible. That is the only way to ensure that also unintentional harm can be minimized and prevented
What is privacy and data governance?
Besides ensuring full respect for privacy and data protection, adequate data governance mechanisms must also be ensured, taking into account the quality and integrity of the data, and ensuring legitimised access to data.
To allow individuals to trust the data gathering process, it must be ensured that data collected about them will not be used to unlawfully or unfairly discriminate against them.
What is privacy and data governance - Quality and integrity of data?
Quality and integrity of the data: When data is gathered, it may contain socially constructed biases, inaccuracies, errors and mistakes. This needs to be addressed prior to training with any given data set.
Access to data:
* Data protocols governing data access should be put in place.
* These protocols should outline who can access data and under which circumstances.
* Only duly qualified personnel with the competence and need to access individual’s data should be allowed to do so.
What is transparency?
The data, system and AI business models should be transparent. Traceability mechanisms can help achieving this.
Moreover, AI systems and their decisions should be explained in a manner adapted to the stakeholder concerned. Humans need to be aware that they are interacting with an AI system and must be informed of the system’s capabilities and limitations.
What is diversity, non-discrimination and fairness?
- Unfair bias must be avoided, as it could have multiple negative implications, from the marginalization of vulnerable groups, to the exacerbation of prejudice and discrimination.
- Fostering diversity, AI systems should be accessible to all, regardless of any disability, and involve relevant stakeholders throughout their entire life circle.
What is societal and environmental well-being?
AI systems should benefit all human beings, including future generations.
It must hence be ensured that they are sustainable and environmentally friendly.
Moreover, they should take into account the environment, including other living beings, and their social and societal impact should be carefully considered.
What is accountability - in reference to AI?
- Mechanisms should be put in place to ensure responsibility and accountability for AI systems and their outcomes.
- Auditability, which enables the assessment of algorithms, data and design processes plays a key role therein, especially in critical applications.
- Further, adequate and accessible redress should be ensured.
What are the limitations
REFER TO SLIDES FOR LINKS TO STUDY FROM
What are some performance metrics for model evaluation?
- Confusion Matrix
- Precision, Recall, and F1 Score
- Balanced Accuracy
- Receiver Operator Characteristics Curve (ROC)
What is some of the terminology used in pattern classification?
Often used in Pattern Classification Problems:
True positive
The object is there and our classifier says it is there
True negative
The object is not there and our classifier says it is not there
False negative (false misses)
The object is there and our classifier says it is not there
False positive (false hits)
The object is not there and our classifier says it is there
What is a confusion matrix?
REFER TO SLIDES FOR EXAMPLE AND FORMULA
What is limitation of accuracy?
REFER TO SLIDES FOR EXAMPLE
What is true postive rate and false positive rate?
REFER TO SLDIES FOR FORMULA
What is sensitivity and specificity? - GO OVER LECTURE RECORDING FOR THIS
- Sensitivity (True Positive Rate) is the probability of a positive test result, conditioned on the individual truly being positive.
- Sensitivity Formula: REFER TO SLIDES
- Specificity (True Negative Rate) is the probability of a negative test result, conditioned on the individual truly being negative.
- # Specificity Formula: REFER TO SLIDES
- Sensitivity
Probability of a true-positive = TP/(TP+FN) - Specificity
Probability of a true-negative = TN/(TN+FP) - The probability of a correct decision = (TP+TN)/S,
where S is the total number of samples
What is precision and Recall?
Precision is the ratio between true positives versus allpositives
Recall is the measure of how accurate the model is in identifying true positives
REFER TO SLIDES FOR FORMULA FOR BOTH
Formula for precision, recall and accuracy
Precision = TP/(TP+FP)
Recall = TP/(TP+FN)
Accuracy = (TP+TN)/(TP+TN+FP+FN)
What is an F1 score?
F1 score takes into account both precision and recall and is based on a balance of the two.
The F1 score is a useful performance metric in machine learning, especially for imbalanced classification problems, where the number of samples in each class is not equal. The F1 score combines two important metrics: precision and recall, into a single value, helping to balance their trade-off.
REFER TO SLIDES FOR FORMULA
What is balanced accuracy?
Balanced accuracy provides a more insightful measure by accounting for both your model’s sensitivity (true positive rate) and specificity (true negative rate). This makes it particularly valuable in real-world scenarios where imbalanced data is common, and the minority class is usually more important.
* Balanced Accuracy accounts for the performance on both the positive and negative classes, making it particularly useful in imbalanced datasets.
* It is the average of Sensitivity(Recall) and Specificity (True Negative Rate).
Balanced Accuracy = (sensitivity + specificity) / 2
* Standard Accuracy may be misleading in imbalanced datasets because a model could achieve high accuracy simply by predicting the majority class, while balanced accuracy ensures both classes are considered equally.
What is parameters vs performance?
- Once we have designed our classifier, we invariably have some parameters we’d like to adjust. e.g. – Prior probability, Threshold
- The optimal classifier is one with sensitivity (Probability of True Positive) as close to 100% as possible, and at the same time with specificity (Probability of True Negative) as close to 100% as possible
What is ROC
Developed in 1950s for signal detection theory to analyze noisy signals
– Characterize the trade-off between positive hits and false alarms
ROC curve plots TP (on the y-axis) against FP (on the x-axis)
Performance of each classifier represented as a point on the ROC curve
– changing the threshold of algorithm, sample distribution or cost matrix changes the location of the point.
* A Receiver Operating Characteristic Curve (ROC) is a standard technique for summarizing classifier performance over a range of trade-offs between true positive (TP) and false positive (FP) error rates (Sweets, 1988).
* ROC curve is a plot of sensitivity (the ability of the model to predict an event correctly) versus 1-specificity for the possible cut-off classification probability values .
ROC GRAPH
REFER TO SLIDES FOR EXAMPLE (SLIDE 101 and 102)
