CAIC 9.3 Flashcards
(124 cards)
1
Q
What are some marketing tactics employed by retailers?
A
Direct marketing emails, digital advertisements, incentives, discounts based on demographics
These tactics aim to attract potential customers.
2
Q
What is the primary goal of marketing campaigns in retail?
A
Achieve a high conversion rate while minimizing advertising costs and reducing customer disturbances.
3
Q
What role do ML models play in marketing campaigns?
A
Optimize the effectiveness of marketing campaigns by identifying potential customers and appropriate messaging.
4
Q
What is customer segmentation in marketing?
A
Understanding different customer segments to improve marketing campaign effectiveness.
5
Q
What is unsupervised clustering in customer segmentation?
A
Grouping customers based on demographic data without predefined labels.
6
Q
Fill in the blank: Highly personalized _______ can improve conversion rates.
A
marketing campaigns
7
Q
What type of data is used to create accurate individual profiles for targeted marketing?
A
Historical transaction data, response data to historical campaigns, social media data.
8
Q
What is contextual advertising?
A
A targeted marketing technique that displays ads relevant to the content on a web page.
9
Q
How does generative AI enhance targeted marketing?
A
Creates dynamically personalized content tailored to individual customer preferences.
10
Q
Why is understanding consumer perception crucial for retail businesses?
A
It significantly impacts their success and brand reputation.
11
Q
What is sentiment analysis?
A
A text classification problem that determines whether sentiment is positive, negative, or neutral.
12
Q
How do ML algorithms assist in sentiment analysis?
A
By training models to detect sentiment in text data.
13
Q
What techniques do retailers use to assess customer sentiment?
A
Soliciting feedback, monitoring social media channels.
14
Q
What is the purpose of inventory planning and demand forecasting in retail?
A
Manage inventory costs while maximizing revenue and avoiding out-of-stock situations.
15
Q
What limitations do traditional methods for demand forecasting have?
A
Accuracy and reliability issues.
16
Q
What statistical techniques are retailers using for demand forecasting?
A
Regression analysis, deep learning.
17
Q
Fill in the blank: ML-based forecasting models can generate both _______ and probabilistic forecasts.
A
point forecasts
18
Q
What is the significance of AI and ML in the automotive industry?
A
Improving efficiency, safety, and customer experience.
19
Q
What is one of the most significant applications of AI and ML in the automotive industry?
A
Autonomous driving.
20
Q
What are the three main stages of the system architecture in autonomous vehicles?
A
Perception and localization, decision and planning, control.
21
Q
What is the role of the perception stage in autonomous driving?
A
Gathering information about surroundings and determining the vehicle’s position.
22
Q
What types of sensors are used in the perception stage of autonomous vehicles?
A
RADAR, LIDAR, cameras.
23
Q
What does the decision and planning stage in autonomous driving do?
A
Controls the motion and behavior of the vehicle based on perception data.
24
Q
How do AI and ML enhance the path planning process in autonomous vehicles?
A
By analyzing real-time data, traffic patterns, and user inputs.
25
What is the function of the control module in autonomous driving?
Translates decisions into physical actions controlling the vehicle.
26
What is Adaptive Control in the context of autonomous vehicles?
Dynamically adjusts control inputs based on sensor data and real-time feedback.
27
What are Advanced Driver-Assistance Systems (ADAS)?
Technologies that enhance driving safety and experience.
28
What features do ADAS include?
Lane departure warning, automatic emergency braking.
29
What is the primary role of ML solutions architects?
Identify suitable data science solutions and design technology infrastructure.
30
What do ML algorithms learn by optimizing?
An objective function, also known as a loss function.
31
What is the purpose of optimization in ML?
Minimize or maximize an objective function.
32
Fill in the blank: The learning rate is a hyperparameter that controls the _______ of parameter updates.
magnitude
33
What is gradient descent?
An iterative approach for optimizing ML algorithms by calculating error changes.
34
What is gradient descent?
An iterative approach for optimizing neural networks and ML algorithms by calculating the rate of error change associated with each input variable.
35
What is the role of the learning rate in gradient descent?
Controls the magnitude of parameter updates at each iteration.
36
List the key steps in the gradient descent optimization process.
* Initialize the value of W randomly
* Calculate the error (loss)
* Compute the gradient of the error
* Update W to reduce the error
* Repeat until the gradient is zero
37
What is the normal equation in the context of ML algorithms?
A one-step analytical solution for calculating the coefficients of linear regression models.
38
What are the primary types of ML tasks discussed?
* Classification
* Regression
39
Fill in the blank: Classification algorithms are suitable for tasks where the goal is to ______ data into distinct classes.
[categorize]
40
What is overfitting in machine learning?
When a trained model learns the training data too well but fails to generalize to new, unseen data.
41
What factors should be considered when selecting a ML algorithm?
* Problem type
* Dataset size
* Number and nature of features
* Computational requirements
* Interpretability of results
* Assumptions about data distribution
42
What is linear regression?
A method that utilizes a linear function of a set of coefficients and input variables to predict a scalar output.
43
What does logistic regression estimate?
The probability of an event or outcome, such as transaction fraud or passing an exam.
44
True or False: Logistic regression is suitable for problems with complex non-linear relationships.
False
45
What is a decision tree in machine learning?
A model that divides data hierarchically based on rules, leading to similar data points following the same decision path.
46
List the advantages of decision trees.
* Capture non-linear relationships
* Handle both numerical and categorical features
* Minimal preprocessing required
* Highly interpretable
47
What is a limitation of decision trees?
Prone to overfitting, especially with a large number of features and noisy data.
48
What is the role of the Gini index in decision trees?
Measures the probability of misclassification.
49
What is random forest?
An ensemble method that combines the decisions of multiple decision trees to improve overall performance.
50
What is bagging in the context of random forests?
A technique that involves using the same sample multiple times in a single tree to make the model more generalized.
51
List the advantages of random forests over decision trees.
* Improved accuracy
* Reduced overfitting
* Robust to outliers
* Feature importance estimation
52
What is gradient boosting?
A sequential approach that aggregates results from different trees, where each tree corrects the errors of the previous one.
53
List the advantages of gradient boosting.
* Handles imbalanced datasets
* Potential for higher performance with tuning
* Supports custom loss functions
* Captures complex relationships in data
54
What is a limitation of gradient boosting?
Lacks parallelization capabilities, making it slower in training compared to parallelizable algorithms.
55
What is the primary advantage of gradient boosting?
It has the potential to achieve higher performance than other algorithms when properly tuned.
56
What does gradient boosting support that adds flexibility in modeling?
Custom loss functions.
57
What is a limitation of gradient boosting related to data?
It is sensitive to noisy data, including outliers.
58
What is XGBoost?
A widely-used implementation of gradient boosting that focuses on speed and performance.
59
What key improvement does XGBoost offer over traditional gradient boosting?
It enables training a single tree across multiple cores and CPUs for faster training times.
60
Name two other popular variations of gradient boosting trees.
* LightGBM
* CatBoost
61
What is K-NN primarily used for?
Both classification and regression tasks.
62
What is the underlying assumption of the K-NN algorithm?
Similar items tend to have close proximity to each other in the feature space.
63
How does K-NN determine the class label for a new data point?
Through majority voting among the K nearest neighbors.
64
What is one advantage of K-NN?
Its simplicity and lack of the need for training or tuning with hyperparameters.
65
What is a significant limitation of K-NN related to dataset size?
As the number of data points increases, the complexity grows, making predictions slower.
66
What is an artificial neuron modeled after?
The learning process of the human brain.
67
What is the role of the activation function in an artificial neuron?
It modifies the output of the linear function to capture non-linear relationships.
68
What does MLP stand for in the context of neural networks?
Multi-Layer Perceptron.
69
What is the purpose of backpropagation in neural networks?
To adjust the weights of each neuron to optimize the training objective.
70
What types of tasks can MLP handle?
* Classification
* Regression
71
What is clustering in data mining?
Grouping items together based on their shared attributes.
72
What is the K-means clustering algorithm used for?
To group similar data points together in clusters.
73
What is a significant drawback of K-means clustering?
Selecting the optimal number of clusters (K) can be subjective and challenging.
74
What are the three important characteristics of time series data?
* Trend
* Seasonality
* Stationarity
75
What does the trend of a time series indicate?
The long-term direction of the data, whether it shows an overall increase or decrease.
76
What is ARIMA used for?
Analyzing and predicting time series data.
77
What does the 'I' in ARIMA stand for?
Integrated, referring to the differencing of the time series to achieve stationarity.
78
What is a major advantage of using DeepAR for time series forecasting?
It captures complex non-linear relationships and can utilize multivariate datasets.
79
What is a significant drawback of DeepAR?
Its black-box nature makes it less interpretable compared to simpler statistical methods.
80
What does DeepAR model in marketing?
Multiple variables simultaneously
## Footnote
DeepAR provides accurate predictions and insights for marketing campaigns.
81
What is a significant drawback of DeepAR?
Black-box nature, lacks interpretability and transparency
## Footnote
This makes forecasts difficult to explain compared to simpler statistical methods.
82
What is a major disadvantage of DeepAR regarding data?
Data-hungry nature, performs poorly with small datasets
83
What is a recommender system?
An ML technology that predicts user preferences for items
## Footnote
It relies on user or item attribute similarities or user-item interactions.
84
In which industries has the recommender system gained widespread adoption?
* Retail
* Media and entertainment
* Finance
* Healthcare
85
What is collaborative filtering?
A recommendation algorithm based on similarities in user preferences
## Footnote
It analyzes collective experiences to recommend items.
86
What is one major benefit of collaborative filtering?
Provides highly personalized recommendations
87
What is the cold-start problem in collaborative filtering?
Struggles with new users or items with no ratings
88
What technique is commonly used in collaborative filtering for recommendations?
Matrix factorization
89
What is matrix factorization's role in collaborative filtering?
Scalability benefits for large catalogs
## Footnote
It sacrifices some interpretability due to latent factor modeling.
90
What is a Multi-Armed Bandit (MAB) approach?
A recommendation system inspired by trial and error
91
How do MAB algorithms learn?
Through online learning without pre-existing training data
92
What is the balance that MAB algorithms strive to achieve?
Exploration and exploitation
93
What does computer vision refer to?
Ability of computers to interpret visual representations
94
What are some tasks performed by computer vision?
* Object identification
* Image classification
* Text detection
* Face recognition
* Activity detection
95
What is a Convolutional Neural Network (CNN)?
A deep learning architecture for processing image data
96
What does a convolutional layer do in a CNN?
Extracts features from input images
97
What is max pooling in CNNs?
A technique that selects the maximum value from outputs
98
What is the vanishing gradient problem?
Signals diminish as they traverse through multiple layers
99
How does ResNet address the vanishing gradient problem?
By implementing a layer-skipping technique
100
What is Natural Language Processing (NLP)?
Focuses on the relationship between computers and human language
101
What are some tasks performed in NLP?
* Document classification
* Topic modeling
* Speech recognition
* Language translation
102
What are the two widely used methods for representing word relevance before embeddings?
* Bag of Words (BOW)
* Term Frequency-Inverse Document Frequency (TF-IDF)
103
What is the main idea behind Bag of Words (BOW)?
Counts the number of times a word appears in a text
104
What does TF-IDF measure?
Importance of a word in a document and across all documents
105
What is embedding in NLP?
Generates low-dimensional representations capturing semantic meaning
106
What is embedding?
A technique used to generate low-dimensional representations (mathematical vectors) for words or sentences, capturing the semantic meaning of the text.
107
What is the underlying idea of embedding?
Words or sentences with similar semantic meanings tend to occur in similar contexts.
108
In a multi-dimensional space, how are semantically similar entities represented?
The mathematical representations of semantically similar entities are closer to each other than those with different meanings.
109
What metric is often used to measure the similarity of embedding vectors?
Cosine similarity.
110
What does the embedding vector represent?
The intrinsic meaning of the word, with each dimension representing a specific attribute associated with the word.
111
Why have embeddings become crucial in NLP tasks?
They offer more meaningful representations of the underlying text compared to techniques like simple word counts.
112
Who created Word2Vec and in what year?
Thomas Mikolov in 2013.
113
What are the two techniques supported by Word2Vec for learning embeddings?
* CBOW (Continuous Bag of Words)
* Continuous-skip-gram.
114
What does CBOW try to predict?
A word for a given window of surrounding words.
115
What does continuous-skip-gram try to predict?
Surrounding words for a given word.
116
What is the training dataset for CBOW generated from?
A sliding window across running text.
117
How is the training problem formulated in Word2Vec?
As a multi-class classification problem where the model learns to predict classes (words in the vocabulary) for the target word.
118
What type of network is used to train Word2Vec embeddings?
A one-hidden-layer MLP (Multi-Layer Perceptron) network.
119
What does the output of the MLP network represent in Word2Vec?
A probability distribution for the target words.
120
What is the purpose of using embeddings in downstream tasks?
To serve as features for tasks such as text classification or entity extraction.
121
What is a common approach to using embeddings in NLP tasks?
Using pre-trained embeddings.
122
What limitation does Word2Vec have regarding word meanings?
It produces a fixed embedding representation for each word, disregarding contextual variations in meaning.
123
What are contextualized word embeddings?
Embeddings that take into account the surrounding words or overall context in which a word appears.
124
What advantage do contextualized embeddings provide?
They capture the diverse meanings a word can have, enabling more accurate context-specific analyses.
