Describe Compute Architectures Used with AI Workloads (CPU, GPU, TPU, FPGA) CERTIFIED-AI-EXPERT Practice Questions (Page 3)

Questions 11–15

1. 11application · medium

   A data science team is building a preprocessing pipeline that reads large CSV files, performs data cleaning, and applies feature scaling before feeding data to a GPU-based training job. The preprocessing steps are mostly sequential and involve complex branching logic. Which architecture is most appropriate for this preprocessing stage?

   • AA CPU
   • BAn FPGA
   • CA TPU
   • DA GPU
   Reveal Solution Discussion Select an answer first
2. 12expert · hard

   A financial services firm needs to deploy a custom AI model for real-time fraud detection. The model's architecture is still evolving, and the team expects to make frequent changes to the inference logic. They have strict power and space constraints in their data center, and the workload is latency-sensitive but not throughput-bound. Which compute architecture best balances flexibility and performance for this evolving, latency-critical workload?

   • AA field-programmable gate array (FPGA)
   • BA fixed-function ASIC
   • CA high-end GPU
   • DA high-core-count CPU
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   What does 'field-programmable' mean in the context of an FPGA?

   • AThe hardware logic can be reconfigured after manufacturing
   • BThe chip can be updated with new firmware after deployment
   • CThe chip can be overclocked to increase performance
   • DThe chip can be programmed using high-level languages like Python
   Reveal Solution Discussion Select an answer first
5. 14expert · hard

   A defense contractor needs to deploy an AI model for real-time object detection on an unmanned aerial vehicle (UAV). The model must be updated in the field to adapt to new environments, and the hardware must be power-efficient and able to handle low-latency inference. The team has expertise in hardware description languages. Which architecture is most appropriate?

   • AA high-core-count CPU
   • BA high-end GPU
   • CAn FPGA
   • DA fixed ASIC
   Reveal Solution Discussion Select an answer first
6. 15foundation · easy

   How does a TPU differ from a GPU in terms of flexibility?

   • ATPUs are more flexible because they can be reprogrammed for any task
   • BTPUs are less flexible because they are specialized for tensor operations
   • CTPUs are equally flexible as GPUs because both are general-purpose processors
   • DTPUs are more flexible because they support dynamic reconfiguration
   Reveal Solution Discussion Select an answer first