IBM A1000-103 Intermediate Practice Exam: Medium Difficulty 2025

A data science team is building a sentiment analysis model and notices that their training data contains 85% positive reviews and only 15% negative reviews. The model performs well on training data but poorly identifies negative sentiment in production. What is the primary issue and recommended solution?

An organization wants to implement IBM Watson Assistant to handle customer service inquiries. They need the assistant to escalate complex queries to human agents while handling routine questions autonomously. Which Watson Assistant feature combination best supports this requirement?

A machine learning engineer needs to evaluate a binary classification model for fraud detection where false negatives (missing actual fraud) are far more costly than false positives. Which evaluation metric should be prioritized when tuning the model?

A retail company is deploying a product recommendation AI model to production. They want to ensure the model continues to perform well as customer preferences evolve over time. What deployment strategy should they implement?

During the training of a deep neural network, a data scientist observes that the training loss continues to decrease while the validation loss starts increasing after epoch 15. What phenomenon is occurring and what is the appropriate action?

An enterprise is implementing IBM Watson Discovery to analyze thousands of legal documents. They need to extract specific contract clauses and understand relationships between entities. Which Watson Discovery capabilities should they leverage for this use case?

A healthcare organization is developing a diagnostic AI system. They must ensure the model's decisions can be explained to medical professionals and patients. Which AI concept and approach should be prioritized in the model design?

A manufacturing company wants to deploy a computer vision model for quality control inspection on their production line. The model must process images in real-time with latency under 100ms. What deployment consideration is most critical for meeting this requirement?

A data scientist is preparing a dataset for training a supervised learning model. The dataset contains numerical features with vastly different ranges: age (0-100), income (20000-500000), and credit score (300-850). What preprocessing step is most important before training?

A company is using IBM Watson Natural Language Understanding to analyze customer feedback across multiple languages. They notice the sentiment analysis performs well for English but poorly for other languages. What approach should they take to improve multi-language performance?