A company wants to deploy a machine learning model for hiring. They've ensured that the model is highly accurate. However, they're facing criticism because the inner workings of their model are a "black box," and candidates want to know why they were or were not selected. This criticism mainly pertains to which aspect of machine learning?
- Explainability
- Accuracy
- Training Data
- Hyperparameter Tuning
The criticism about the model being a "black box" highlights the need for explainability in machine learning. It's essential to understand how and why the model made hiring decisions, not just the accuracy of those decisions.
In a neural network, what are the nodes that receive input data and pass it forward called?
- Neurons
- Synapses
- Layers
- Weights
In a neural network, the nodes that receive input data and pass it forward are called "Neurons." Neurons process and transmit information.
The balance between taking known rewarding actions and discovering new ones is termed as the ________ dilemma.
- Exploration-Exploitation
- Action-Selection
- Reinforcement-Reward
- Decision-Making
This is known as the Exploration-Exploitation dilemma, which is central in Reinforcement Learning, referring to the balance between exploiting known actions for rewards and exploring new actions to discover more rewarding ones.
When determining the number of clusters (K) for K-means clustering, which method involves plotting the variance as K increases and looking for an "elbow" in the plot?
- Davies-Bouldin Index
- Elbow Method
- Gap Statistics
- Silhouette Analysis
The Elbow Method involves plotting the variance (inertia) against different values of K. The "elbow" point is where the rate of decrease sharply changes, indicating an optimal K.
The ________ classifier calculates the probability of a particular event based on prior knowledge of conditions related to that event.
- Decision Tree
- K-Means
- Naive Bayes
- Support Vector Machine
The Naive Bayes classifier calculates the probability of an event based on prior knowledge of conditions. It's a probabilistic classifier.
In the Actor-Critic architecture, which part directly decides on the action to be taken?
- The Actor
- The Critic
- The Decision Maker
- The Environment
The Actor is directly responsible for deciding on the action to be taken, while the Critic provides feedback on these actions.
A research institution is working on a system to automatically summarize long medical research papers into concise abstracts. This task primarily requires expertise in which domain?
- Medical Domain
- Computer Science
- Linguistics
- Information Retrieval
Summarizing research papers into concise abstracts relies heavily on linguistic expertise to ensure the summaries are coherent and convey the key information effectively.
Using domain knowledge to transform raw data into a more suitable format or variable for modeling is known as ________.
- Data Preprocessing
- Data Transformation
- Feature Engineering
- Raw Data Analysis
The process of using domain knowledge to transform raw data into a more suitable format or create new variables is called "feature engineering." It is a critical step in data preparation to improve model performance by providing more informative features.
In a Gaussian Mixture Model, what parameter determines the shape, size, and orientation of the cluster?
- Covariance Matrix
- Number of Clusters
- Mean Vector
- Standard Deviation
In a Gaussian Mixture Model, the shape, size, and orientation of a cluster are determined by the covariance matrix, which represents the dispersion and correlation between data points in the cluster.
A telemedicine platform wants to develop a feature where patients can describe their symptoms in natural language, and the system provides potential diagnoses. This feature would heavily rely on which technology?
- Natural Language Processing (NLP)
- Artificial Intelligence (AI)
- Blockchain Technology
- Augmented Reality (AR)
The feature of interpreting natural language symptoms and providing diagnoses relies on Natural Language Processing (NLP) to understand and process the patient's input effectively.
An e-commerce company has collected data about user behavior on their website. They are now interested in segmenting their users based on similar behaviors to provide personalized recommendations. While they considered decision trees, they were concerned about stability and overfitting. Which ensemble method might they consider as an alternative?
- AdaBoost
- Bagging (Bootstrap Aggregating)
- Gradient Boosting
- XGBoost
Gradient Boosting is a strong alternative. It's an ensemble method that combines the predictions of multiple decision trees, focusing on correcting the errors of previous trees. It typically performs well, provides stability, and mitigates overfitting concerns.
When a machine learning algorithm tries to group...
- Reinforcement Learning
- Semi-Supervised Learning
- Supervised Learning
- Unsupervised Learning
Unsupervised learning involves clustering or grouping data without prior labels. Algorithms in this approach aim to identify patterns and structure in the data without any guidance from labeled examples.