TQ Generative AI: Test Your Knowledge! Which technique is used to combine predictions from multiple models in MLOps?Transfer learningContinuous trainingModel compressionEnsemble learningCan AI help in preserving historical artifacts?Yes, by digitizing and analyzing artifactsNo, it's only for the futureOnly during historical reenactmentsOnly during holidaysWhat is the primary purpose of the "value" vector in transformer attention mechanisms?To calculate attention scoresTo represent the content at each positionTo determine the order of tokensTo normalize the inputWhat is the main advantage of using XLNet over BERT in NLP?It's always fasterIt can learn bidirectional contexts using permutation language modelingIt requires less memoryIt works only for EnglishWhich US space agency has been applying AI in various missions and research?NOAAUSGSNASAFAAIn the context of LLMs, what does "tokenization" refer to?Encrypting the input textCompressing the input textSplitting the input text into smaller unitsTranslating the input text into another languageWhich of these is NOT a common challenge in prompt engineering?Ensuring consistent outputsManaging token limitsHandling multi-turn conversationsCompiling the AI modelWhat is the main cause of mode collapse in GANs?The generator is too simpleThe discriminator is too complexThe GAN objective encourages the generator to fool the discriminatorThe training data is not diverse enoughCan generative models be used to create personalized and adaptive AR/VR experiences?Yes, by generating content based on user preferences and behaviorNo, AR/VR experiences are pre-programmed and staticOnly for certain types of AR/VR applications like gamingOnly if the user explicitly provides feedback and inputIn prompt engineering, why is it important to specify the desired format?To confuse the AITo ensure the AI output matches the intended structureTo increase processing speedTo reduce model sizeWhat contributes to trustworthy AI development?Unpredictable outcomesAdherence to ethical guidelines and standardsMaximized autonomyMinimal human oversightWhat is a 'latent space' in the context of generative AI?A hidden physical spaceA concept in quantum mechanicsA mathematical representation of dataA storage area for dataIn computer vision, what is the main purpose of federated learning in US healthcare AI applications?To reduce computational costsTo preserve patient privacy while training modelsTo enhance image resolutionTo generate synthetic medical imagesWhat is virtual try-on in the context of AI and fashion?Using AI to identify clothing sizesUsing AI to recommend outfitsUsing AI to let customers see clothes on their own body digitallyUsing AI to design custom clothingWhat is a common challenge in generating long-form coherent text with Generative AI?training data availabilitycomputational resource needsmaintaining consistency and logical flowgenerating grammatically correct textWhich Python library is specifically designed for processing code and programming languages?NLTKspaCytree-sitterTextBlobWhich US-based company is known for its development of the DALL-E image generation model?GoogleMicrosoftOpenAINvidiaWhat is the main application of AI in the optometry industry?To replace optometristsFor prescription accuracy and disease detectionTo manufacture glassesTo perform eye surgeryWhat is the primary use of AI in energy grid management?To create infinite energyTo optimize energy distribution and usageTo replace power plantsTo make grids AI-controlledWhat is transfer learning in the context of PyTorch?Using a pre-trained model as a starting point for a new taskTransferring data between different storage devicesMoving models between CPU and GPUTranslating models from one framework to anotherHow should you complete the URI to build the REST endpoint for creating multilingual product descriptions in the Wide World product creation project?api-nam.cognitive.microsofttranslator.com, /translatetranslate.api.cognitive.microsoft.com, /v3.0/translateapi.microsofttranslator.com, /v2/translatetranslator.azure.ai, /api/v1/translateWhich layer in a CNN is responsible for introducing non-linearity?Convolutional layerPooling layerBatch Normalization layerActivation layerWhat is the importance of interdisciplinary research in Responsible AI?It's not importantTo address complex ethical and societal implications comprehensivelyTo complicate AI researchTo slow down AI progressWhat is the role of 'pooling layers' in convolutional neural networks?To create pools of dataTo make data swim togetherTo reduce the spatial dimensions of the inputTo pool resources for AIWhat is the purpose of a Conditional GAN?To generate images from labels or attributesTo generate images without any conditionsTo classify images into categoriesTo detect objects in images Score: 0/25 Retake Quiz Next Set of Questions