Azure Data Science Certification Flashcards

1
Q

What do Inbound Security Rules do?

A

Ensure only required communications from the internet are allowed.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Batch Node Management

what is it?
what does it allow you to do?
what is the destination port #?

A

Type of service tag which is used to identify groups IP addresses. Service tags are groups IP addresses.

Allows you to define security rules without having to manually identify IP addresses.

The service tag requires a destination port of 29876 - 29877

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Secure Shell (SSH) Protocol

(what is the destination range?)

A

destination port range: 22

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Service Principal (SP)

definition & what is created during the creation process

A

Directory object (an alias for a physical directory path) that can be used for token-based authentication. As part of the creation process a client secret is generated which is comparable to a password.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

token-based authentication

A

Requires users to obtain a computer-generated code (or token) before they’re granted network entry.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Apache Ambari

(used for)

A

Management of Hadoop Clusters

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Azure Log Analytics

(used for)

A

Querying Diagnostic Logs

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Azure Sentinel

(used for)

A

cloud security information event management (SIEM)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Explainers

(definition & 2 examples)

A

Interpretable techniques used to explain how a machine learning model works.

TabularExplainer - used tabular datasets
LinearExplainer - explains linear models

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Image classification

(what is the primary metric?)
(does imply maximizing or minimizing this metric?)

A
  1. accuracy metric (the proportion of instances that have been correctly classified)
  2. it implies that you want to maximize the metric
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Spearman Correlation

A

calculates the monotonic relationship between two values

monotonic relationship
(1) as the value of one variable increases, so does the value of the other variable; or 

(2) as the value of one variable increases, the other variable value decreases.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Azure Kubernetes Services (AKS)

(what type of compute instances does it support?)
(which deployment is it recommended for?)

A
  1. GPU-based compute instances for deploying real-time inference models.
  2. It is recommended for production deployment
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Azure Container Instances (ACI)

(what is it suited for? what are its limitations?)

A
  • suited for testing /debugging workloads but does not support GPU-based instance configuration, only low-scale CPU-based workloads
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Which computer cluster should you use?

  1. Batch Inference
  2. Real Time
  3. Testing & Debugging
A
  1. Compute Cluster (cannot handle real-time data, because they utilize low priority VMs that may not scale to the load required)
  2. Azure Kubernetes Service (AKS)
  3. Computer Instance Web Service (testing and debugging are best done on low-cost cloud or local resources, per-hour cost associated with compute clusters and AKS make them poor candidates for testing and debugging)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

run.log_list()

A

method that provides a log list of values, arrays, dictionaries, vectors, or any non-scalar value

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

run.log_image()

A

logs transformed PNG images to the run record, log matplotlib plots to the run record

17
Q

truncation selection policy

A

cancels a percentage of runs with low performance on the primary metric