Lesson 1B Introduction to AI: Pre Quiz * A famous 19th century proto-computer engineer was - Charles Barkley + Charles Babbage - Charles Darwin * Weak AI is a system designed to solve many tasks - True + False * Chat bots are an example of truly intelligent systems - false, they are usually designed by a series of rules. - true, they are usually considered to be 'intelligent + false, but they are increasingly able to pass Turing tests as they become more sophisticated. Lesson 1E Introduction to AI: Post-Quiz * A top-down approach to AI is a model of reasoning called - strategic reasoning + symbolic reasoning - synergistic reasoning * A bottom-up approach to AI is based on neural networks + True - False * The AI Winter occurred in this era - 1950s - 1960s + 1970s Lesson 2B Knowledge Representation and Expert Systems: Pre-Quiz * The top-down approach to creating intelligent systems was based on: - knowledge seeking and reading + knowledge representation and reasoning - knowledge reasoning and seeking * Knowledge is the same as information - True + False * Knowledge is obtained by an: + active learning process - passive learning process - both of these Lesson 2E Knowledge Representation and Expert Systems: Post-Quiz * The simplest method of knowledge representation is: + algorithmic - symbolic - synergistic * Scenarios can represent complex situations that can unfold in time + true - false * Forward inference starts with initial data and then: + executes a reasoning loop - looks for a goal - starts over Lesson 3B Introduction to Neural Networks - Perceptron: Pre-Quiz * Early neural networks required + manual weight adjusting - terabytes of data - special reasoning * A simple neuron is also called a 'threshold logic unit' + true - false * A perceptron is a ___ type of model - multi-class classification - clustering + binary classification Lesson 3E Introduction to Neural Networks - Perceptron: Post-Quiz * To train a perceptron, find a weights vector that results in the smallest ___. - size + error - nodes * To minimize the function of weights, you can use gradient descent + true - false * During gradient descent, each step updates the ___ - learning rate + weights - gradient Lesson 4B Neural Networks: Pre Quiz * The quality of prediction is measured by Loss function + True - False * One layer network is capable of classifying ____ - linearly joined classes + linearly separable classes - single layers of classes * The method of training multi-layered perceptron is called ____ + back propagation - multiple propagation - front propagation Lesson 4E Neural Networks: Post Quiz * We use ____ for regression loss functions - absolute error - mean squared error + all of the above * All but one is a type of classification loss function - 0-1 loss + binary loss - logistic loss * Cross-entropy loss is a function that can calculate similarity between two arbitrary probability distributions + True - False Lesson 5B Frameworks: Pre Quiz * Deep Neural Network training requires a lot of computations + True - False * Overfitting occurs because of ____ - Not enough testing data + Too powerful model - Too much noise in output data * Bias errors are caused by our ____ not being able to capture the relationship between training data correctly. - model + algorithm - computer Lesson 5E Frameworks: Post Quiz * After compiling our model object, we train by calling ____ function + fit - train - teach * Binary cross-entropy is also called log loss + True - False * TensorFlow is to ____ while PyTorch is to ____ - Facebook, Google + Google, Facebook - Microsoft, Google Lesson 6B Introduction to Computer Vision: Pre Quiz * Computer vision aims to allow computers gain understanding of _____ + images - text - computers * Python libraries available for image processing includes - OpenCV - Pillow + a and b * Images cannot be represented as NumPy arrays in Python - true + False Lesson 6E Introduction to Computer Vision: Post Quiz * Optical Flow helps us to understand how each pixel on video frames move. + true - false * _____ computes the vector field that shows where each pixel is moving - Sparse Optical Flow + Dense Optical Flow - none * Resizing and Blurring are steps that can be taken during? + pre-processing - training - image transformation Lesson 7B Convolutional Neural Networks: Pre Quiz * To extract patterns from images we use? + convolutional filters - extractor - filters * One of these is not a CNN Architecture - ResNet - MobileNet + TensorFlow * CNN are mostly used for computer vision tasks. + true - false Lesson 7E Convolutional Neural Networks: Post Quiz * Which pooling layer is used "scale down" the size of the image - average pooling - max pooling + a and b * Convolutional networks generalizes much better + True - False * To train our neural network, we need to convert images to tensors + true - false Lesson 8B Pre-trained Networks and Transfer Learning: Pre Quiz * Transfer learning approach uses untrained models for classification - true + false * One of these is not a normalization technique? + height normalization - weight normalization - layer normalization * We choose Stochastic Gradient Descent(SGD) in deep learning because classical gradient descent can be ____ - fast + slow Lesson 8E Pre-trained Networks and Transfer Learning: Post Quiz * Dropout layers act as a ____ technique - gradient boosting - training + regularization * freezing weights of convolutional feature extractor can be done by ____ - setting `requires_grad` property to `False` - setting `trainable` property to `False` + a and b * Batch normalization is to bring values that flow through the ____ to right interval - algorithms - batches + neural network Lesson 9B Autoencoders: Pre Quiz * Self-supervised learning uses ____ data for training - pre-trained + raw - labeled * Encoder Network coverts input images into latent spaces + true - false * VAE is short for? - Variable AutoEncoding + Variation auto-encoder - Variational automated encoders Lesson 9E Autoencoders: Post Quiz * Properties of autoencoders include - it is data Specific - works on unlabeled data + all of the above * Auto encoders can be used to effectively remove noise from images + true - false * Variational auto-encoders loss function does not consist of which of these? - reconstruction loss - KL loss + TF loss Lesson 10B Generative Adversarial Networks: Pre Quiz * Generators take vectors and produce ____ - videos + image - gif * GANs is short for? - General adversarial networks - Generative advisor networks + Generative adversarial networks * GAN uses ____ neural networks - 1 + 2 - 3 Lesson 10E Generative Adversarial Networks: Post Quiz * We can use Batch normalization and BatchNorm1D to stabilize the training + true - false * Deep Convolutional GAN uses convolutional layers for ____ and ____ + generator, discriminator - CNN, generator - training, testing * Problems of GAN training includes - Sensitivity to hyperparameters - Keeping balance between generator and discriminator + all of the above Lesson 11B Object Detection: Pre Quiz * Neural networks can only be used to classify images - true + false * With object detection, we don't just get the class of an object, but also its ____ - shape + location - type * How many objects can an object detection model detect? - one - two + any number Lesson 11E Object Detection: Post Quiz * An Object detection model gives us - object class - bounding box + both class and bounding box * Which object detection models are faster? + one-pass models - region proposal networks - Fast R-CNN * Which metric can be used to determine how well bounding boxes are aligned? - accuracy - precision + IoU Lesson 12B Segmentation: Pre Quiz * How many segmentation algorithm are there? - 1 + 2 - 3 * Segmentation is a _____ task + computer vision - natural language processing - neural networks * Segmentation networks consist of ____ and ____ parts - classifier, divider + encoder, decoder - generator, discriminator Lesson 12E Segmentation: Post Quiz * ____ extracts features from an input image - decoder - generator + encoder * ____ transforms input features into mask image + decoder - generator - encoder * SegNet relies on ____ to train multi-layered network + batch normalization - height normalization - weight normalization Lesson 13B Text Representation: Pre Quiz * Each word in a Bag of Words is linked to a vector index + true - false * Text can be represented using _____ approaches - 1 + 2 - 3 * Character level representation represents each _____ as a number + letter - word - symbol Lesson 13E Text Representation: Post Quiz * Word level representation represents _____ as a number - letter + word - symbol * N-Grams refers to _____ - combination of n number of words and symbols - combination of n number of letters + combination of n number of Words * The main drawback of N-gram is that the vocabulary size grows fast + true - false Lesson 14B Embeddings: Pre Quiz * Embedding is used to represent words with _____ dimensional dense vectors + lower - higher - average * Word2Vec pre-trained embeddings can also be used in place of embedding layer in neural networks + True - False * Using embedding layer we cannot switch from bag-of-words to embedding bag - True + false Lesson 14E Embeddings: Post Quiz * Word2Vec has _____ main architectures - 1 + 2 - 3 * Word sense disambiguation is a limitation of traditional pretrained embedding representations + True - False * An embedding layer takes _____ as input + word - symbol - number Lesson 15B Language Modeling: Pre Quiz * Which of the following can be considered a language model? + Word2Vec embeddings - Embedding layer in RNN - RNN used for text classification * A language model should be able to ____ the next word in the sentence - use + predict - guess * Language models are trained on - language vocabulary - specially labeled data + any natural text Lesson 15E Language Modeling: Post Quiz * Which of the architectures predicts a word from neighboring words? + CBoW - Skip-gram - N-Gram * When we train a CBoW model, we obtain - A model that can generate text + Word2Vec embedding vectors - Text classification model * The CBoW model is based on + Dense neural network - Convolutional neural network - Recurrent neural network Lesson 16B RNN: Pre Quiz * RNN is short for? - regression neural network + recurrent neural network - re-iterative neural network * A simple RNN cell has two weight _____ + matrices - cell - neuron * vanishing gradients is a problem of _____ + RNN - CNN - KNN Lesson 16E RNN: Post Quiz * _____ takes some information from the input and hidden vector, and inserts it into state - forget gate - output gate + input gate * Bidirectional RNNs runs recurrent computation in _____ + both directions - north-west direction - left-right direction * All RNN Cells have the same shareable weights + True - False Lesson 17B Generative networks: Pre Quiz * RNNs can be for generative tasks + yes - no * _____ is a traditional neural network with one input and one output + one-to-one - sequence-to-sequence - one-to-many * RNN generates texts by generating next output token for each input token + true - false Lesson 17E Generative networks: Post Quiz * Output encoder converts hidden state into _____ output + one-hot-encoded - sequence - number * Selecting the character with higher probabilities always gives a meaningful text. - true + false - maybe * Many-to-many can also be referred to as _____ - one-to-one + sequence-to-sequence - one-to-many Lesson 18B Transformers: Pre Quiz * Attention mechanism provides a means of _____ the impact of an inout vector on an output prediction of RNN + weighting - training - testing * BERT is an acronym for - Bidirectional Encoded Representations From Transformers + Bidirectional Encoder Representations From Transformers - Bidirectional Encoder Representatives of Transformers * In positional encoding the relative position of the token is represented by number of steps + true - false Lesson 18E Transformers: Post Quiz * Positional embedding _____ the original token and its position within the sequence - separates - compares + embeds * Multi-Head Attention is used in transformers to give network the power to capture _____ of dependencies + different types - same type - none * In transformers attention is used in _____ instances - 1 + 2 - 3 Lesson 19B Named Entity Recognition: Pre Quiz * What does NER stands for? - Nearest Estimated Region - Nearest Entity Region + Named Entity Recognition * An entity always consists of one token - true + false * To train NER model, we need + Labeled dataset - Any natural text - Translated texts in two languages Lesson 19E Named Entity Recognition: Post Quiz * NER model is essentially a ____ model - text classification + token classification - text regression * Which neural network types can be used for NER? - RNNs - Transformers + Both RNNs and Transformers * NER model is a good example of ____ network architecture - one-to-one - one-to-many + many-to-many Lesson 20B Language Models: Pre Quiz * What does GPT stand for? - Generic Pre-Trained network + Generative Pre-trained Transformers - Generic Positional Text * What can GPT be used for? - Text generation - Text classification + Both text generation and other tasks * GPT is based on transformer architecture + true - false Lesson 20E Language Models: Post Quiz * What is zero-shot learning? + Getting an answer from pre-trained network - Training the network from scratch - Training the network only for one epoch * Prompt engineering can be used with - Zero-shot learning - Few-shot learning + Both * Which metric can be used to estimate the quality of a language model? - accuracy - recall + perplexity Lesson 21B Genetic Algorithms: Pre Quiz * Genetic Algorithms are based on which of the following? - mutations - Selection + both a and b * Crossover allows us to combine two solutions together to obtain a new valid solution + true - false * Valid solutions to genetic algorithm can be represented as _____ + genes - neurons - cells Lesson 21E Genetic Algorithms: Post Quiz * Genetic Algorithms can solve which of these tasks - Schedule optimization - Optimal packing + both of a and b * In implementing a genetic algorithm the first step is to randomly select two genes - true + false * When using Crossover operation the algorithm randomly selects _____ genes - 3 - 1 + 2 Lesson 22B Reinforcement Learning: Pre Quiz * To train RL model, we need + Simulation environment - Labeled dataset - Unlabeled dataset * What is a good example of Reinforcement learning: - Zero-shot image classification - Zero-shot text classification + Learning to play chess * When creating RL-based chess engine, we need to - use all existing chess matches as a dataset + let computer play against itself many times - program exhaustive search algorithm Lesson 22E Reinforcement Learning: Post Quiz * How does RL training algorithm knows how well it did? - It achieves high accuracy - Using perplexity metric + Using reward function * Which problem(s) is RL is applicable to? - With discrete environment - With continuous environment + Both * In Actor-Critic model, critic predicts + Reward function - Best next action - Probability of next actions Lesson 23B Multi-Agent Modeling: Pre Quiz * By modeling the behavior of simple agents, we can understand more complex behaviors of a system. + true - false * The principle of metasystem transition is derived from: - Evolutionary Cybernetics - Emergentism + both of these * Multi-Agent systems emerged in the ____: - 1970s - 1980s + 1990s Lesson 23E Multi-Agent Modeling: Post Quiz * An agent is: - an entity that lives alone + an entity that lives in an environment - an entity that is intelligent * Reactive agents usually have: + simple request-response behavior - complex behavior - no behavior * Multi-agent systems are used in: - video production and systems modeling - games and automations + both the above Lesson 24B Ethical and Responsible AI: Pre Quiz * Why we need to worry about Ethical AI? - AI is a very powerful tool and can cause harm - We need to make sure AI models do not discriminate people + Both * Which is the example of interpretable AI? + Expert system - Neural network - Image classifier * It is not ethical to use AI in medicine - true + false Lesson 24E Ethical and Responsible AI: Post Quiz * Why an AI model can discriminate? - Because it may become unfriendly + Because datasets were not properly balanced - Because developers programmed it in such a way * Which of the following is not a principle of Responsible AI? - Transparency - Fairness + Cleverness * Accountability of an AI system means that + there should be a human being involved in taking decisions, who can take responsibility - AI system should be held responsible for its actions - AI system developers should be held responsible * Model fairness is related to - Interpretability + Biases - Accountability