Online Access Free NCA-GENM Exam Questions

You are building a multimodal AI system that generates 3D models of furniture from text descriptions and a few 2D images of similar furniture pieces. The system uses separate encoders for text and images. You want to fuse the information from both modalities effectively. Which TWO of the following fusion techniques would be the most appropriate for this task, considering the different nature of the text and image data?

• A.Using a cross-attention mechanism where the text embedding attends to the image features, and vice-versa, allowing the model to dynamically weight the importance of different parts of each modality.
• B.Applying element-wise addition of the text and image embeddings.
• C.Using a gating mechanism (e.g., a learned weight) to control the contribution of the text and image embeddings based on the input.
• D.Simple concatenation of the text and image embeddings before feeding them into the decoder.
• E.Performing a simple average of the text and image embeddings.

You're developing a system that analyzes video footage and generates textual summaries of the events occurring in the video. Which of the following architectures would be the MOST appropriate starting point for this task?

• A.A generative adversarial network (GAN) for image generation.
• B.A long short-term memory (LSTM) network for time series forecasting.
• C.A purely convolutional neural network (CNN) for image classification.
• D.A support vector machine (SVM) for classification.
• E.A transformer-based encoder-decoder architecture, with the encoder processing video frames (e.g., using a 3D CNN) and the decoder generating text.

You are evaluating a generative A1 model for image captioning. The model produces captions that are grammatically correct but often miss key objects in the image. Which of the following evaluation metrics would be MOST suitable to identify this deficiency?

• A.CIDEr (Consensus-based Image Description Evaluation)
• B.Inception Score
• C.ROUGE (Recall-Oriented Understudy for Gisting Evaluation)
• D.Perplexity
• E.BLEU (Bilingual Evaluation Understudy)

You are developing a generative A1 model to create music based on textual descriptions of mood and genre. You have a dataset of paired text descriptions and music tracks. When evaluating the generated music, you realize it's difficult to objectively quantify the quality of the music. Which of the following evaluation methods would provide the MOST comprehensive assessment of the generated music's quality and alignment with the text descriptions?

• A.Conduct a human evaluation study where participants rate the generated music on various subjective criteria such as pleasantness, mood alignment, and overall quality based on the text description provided.
• B.Use a pre-trained music genre classification model to predict the genre of the generated music and compare it to the genre in the text description.
• C.Measure the file size of the generated music tracks.
• D.Calculate the Bit Rate of the generated music tracks.
• E.Calculate the Root Mean Square (RMS) energy of the generated music tracks.

You are building a multimodal RAG application that integrates text documents and images. You've noticed that when a user query relates strongly to the visual content, the retrieved documents are less relevant than desired. Which of the following strategies would MOST effectively improve the retrieval of relevant information in this scenario?

• A.Use a larger language model for the generative component of the RAG pipeline.
• B.Implement cross-modal embedding, training a model to create a joint embedding space for text and images.
• C.Fine-tune the existing text embedding model with more text data.
• D.Increase the k value (number of retrieved documents) in the vector search.
• E.Prepend the image captions to all the source documents to enhance text-based retrieval.