Artificial Intelligence in Dermatology: the Essentials

Quiz Summary

0 of 11 Questions completed

Questions:

You have already completed the quiz before. Hence you can not start it again.

Quiz is loading…

You must sign in or sign up to start the quiz.

You must first complete the following:

Results

Quiz complete. Results are being recorded.

Results

0 of 11 Questions answered correctly

Your time:

You have reached 0 of 0 point(s), (0)

Current
Review
Answered
Correct
Incorrect

Question 1 of 11

1. Question
Artificial Intelligence in Dermatology is currently involved in developing recognition patterns. Which of the following statements is false.
- Artificial Neural Networks (ANNs) are a subset of machine learning called Natural Language Processing (NLP)
- In Dermatology, most of the ANNs called Convolutional Neural Networks (CNNs).
- CNNs are named after the operation they perform: convolutions
- Parallelism is the structure of functioning, and is also known as perceptrons.
Correct

Incorrect
Question 2 of 11

2. Question
Deep learning refers to a learning methodology mimicking the processes of the brain. Which of the following statement (s) is/are true ?
- Deep Learning can only done on unstructured data.
- Deep Learning can be used for supervised learning.
- Deep Learning can be used to find patterns.
- Solutions derived using Deep Learning can be difficult to explain.
Correct

Incorrect
Question 3 of 11

3. Question
The term algorithm is a common term in Artificial Intelligence. The following are correct statements except:
- In artificial neural networks, algorithms connect what is inputed and what comes out of the output layer.
- Algorithms are the equation developed from what happens in the hidden layer.
- Explainability is often possible by analyzing what is happening in the hidden layer.
- The term "Algorithm" is of recent origin.
- The training of an algorithm is facilitated to the access of large quality datasets.
Correct

Incorrect
Question 4 of 11

4. Question
Machine Learning uses Artificial Neural Networks (ANN) to learn to discover patterns. Which of the following statements is false ?
- Artificial Networks are structures like neurons and called nodes.
- Convolutional Neural Networks is the subset of ANNs which is adapted to imaging.
- When data is labelled, the results (values) are compared with the correct diagnosis. Backpropagation is then used to adapt the signal strengths between the nodes.
- Like neurons nodes are polarized and the signal transmission can proceed in 2 directions.
- In convolutions reduction of pixel size of the images can be minimzed using zero padding.
- In Gasnets the signal is not a hard wired signal between nodes, but similar to diffusion.
Correct

Incorrect
Question 5 of 11

5. Question
Explainability in health solutions using AI and data science in health will be needed to make it acceptable in practice. Which of the following statement(s) is/are true?
- Explainbility applies to Data.
- Explainability applies to Health.
- Unexplainable solutions using unsupervised learning are acceptable, provided that they work.
- Healthcare providers need to implement solutions in the real clinical setting. It is their responsibility to implement safe solutions.
- Physicians are tied to a code of medical ethics to their patients only. The art of medecine is only between a patient and the doctor. In primum non nocere, their knowledge doesn't extend to advising beyond this relationship.
Correct

Incorrect
Question 6 of 11

6. Question
Training of a model in health is easier in supervised learning as explainbility is easier. Which of the following statements is false ?
- Supervised learning consists of labelled datasets.
- Training sets, validation sets and test sets are types of datasets used.
- Test sets are labelled.
- Training sets are labelled in supervised learning.
Correct

Incorrect
Question 7 of 11

7. Question
Often models can look pretty good, but they fail to predict on data they haven’t seen before. This is detected by passing a test dataset to the algorithm. Passing the test set is a prerequisite to any potential implementation of a model. Please mark as correct if approporiate the following statements:
- Underfitting can be corrected by reducing the complexity of the model, such as by applying the "dropout" method.
- Overfitting occurs, when a test set is accurate according to the validation set results.
- Overfitting is easier to correct in general than underfitting. The model can be improved by adding more data.
- Overfitting can also be correct by increasing the complexity of the model.
Correct

Incorrect
Question 8 of 11

8. Question
There are measurement parametrics which influence the speed of a model to learn. Which of the following statments is false !
- En epoch is the passing of the data to train the algorithm through an artificial neural network.
- Learning rates can influence adversively processing power.
- An increase in the number of epochs proportionally increases the learning rate.
- An decreas in time decreases the learning rate.
Correct

Incorrect
Question 9 of 11

9. Question
Concerning personalized medicine, which statement is true ?
- Personalized medicine consists of dealing with a patient as a person, not taking into account clinical guidelines.
- Datasets studied include patient information.
- Datasets studied can also include environmental factors.
- Personomics are datasets that define and individual: nique biological characteristics are defined by the tools of precision medicine: genomics, proteomics, metabolomics, epigenomics, pharmacogenomics, and other ″- omics. ″ They are the only datasets used in personalized medicine.
Correct

Incorrect
Question 10 of 11

10. Question
As of 2021, a lot of solutions developed in AI for health fail to make it to the real world. Which of the following statement(s) is(are) true ?
- The model is the algorithmic solution to a problem using datasets to train them.
- Data collected into the deployment of a solution is not useful for further training of the model.
- The deployment of a model is it's integration into the real world.
- Accurate models can always be integrated provided they are safe to use.
Correct

Incorrect
Question 11 of 11

11. Question
Model prediction in AI uses prediction and recall, instead of sensitivity and specificy as is usually done in Medicine. There are relationships between, them. Regarding model prediction, please state whether the folling is true (correct) or false:
- Recall measure the Positive predictive value (PPV) of the model prediction.
- Recall measure the Negative predictive value (NPV).
- Precision is equivalent to specificity.
- Precision is equivalent to sensitivity.
Correct

Incorrect

Artificial Intelligence in Dermatology: the Essentials

Artificial Intelligence in Dermatology: the Essentials

Quiz Summary

Information

Results

Results

Categories

1. Question

2. Question

3. Question

4. Question

5. Question

6. Question

7. Question

8. Question

9. Question

10. Question

11. Question