Exam AWS Certified Machine Learning - Specialty topic 1 question 124 discussion

I agree with an answer of C Attention mechanism. The disadvantage of an encoder-decoder framework is that model performance decreases as and when the length of the source sequence increases because of the limit of how much information the fixed-length encoded feature vector can contain. To tackle this problem, in 2015, Bahdanau et al. proposed the attention mechanism. In an attention mechanism, the decoder tries to find the location in the encoder sequence where the most important information could be located and uses that information and previously decoded words to predict the next token in the sequence.

C. By tuning attention-related hyperparameters (such as attention type, attention layer size, and dropout), the model can focus on relevant parts of the input sequence during translation.

A. NO - n-grams are more the opposite, it is to capture local information B. NO - it could help, but not best C. YES - best practice (https://docs.aws.amazon.com/sagemaker/latest/dg/seq-2-seq-hyperparameters.html) D. NO - weight are for vectorized words, they do not relate to sequences

Action C: Adjusting hyperparameters related to the attention mechanism can help improve the translation quality for long sentences, because the attention mechanism allows the decoder to focus on the most relevant parts of the source sentence at each time step. According to the Amazon SageMaker documentation, the seq2seq algorithm supports several types of attention mechanisms, such as dot, general, concat, and location. The data scientist can experiment with different values of the hyperparameters attention_type, attention_coverage_type, and attention_num_hidden to find the optimal configuration for the translation task.

C. Adjust hyperparameters related to the attention mechanism. The seq2seq algorithm uses an attention mechanism to dynamically focus on relevant parts of the input sequence for each output sequence element. Increasing the attention mechanism's ability to learn dependencies between long input and output sequences might help improve the translation quality for long sentences. The data scientist could try adjusting relevant hyperparameters such as attention depth or attention scale, or try a different attention mechanism such as scaled dot-product attention, to see if that improves the translation quality for long sentences.

i go with C

Ans: C Explanation: https://docs.aws.amazon.com/sagemaker/latest/dg/seq-2-seq- howitworks.html

This is such a niche question for a niche market. Geared towards someone who specializes in NLP.

c is correct https://docs.aws.amazon.com/sagemaker/latest/dg/seq-2-seq-howitworks.html

I believe the answer is C https://docs.aws.amazon.com/sagemaker/latest/dg/seq-2-seq-howitworks.html