Build fluency in the vocabulary of reusing pretrained weights and fine-tuning them on a small, task-specific dataset.
0 / 5 completed
1 / 5
A teammate explains that a team starts model training from weights already learned on a large, general dataset, then fine-tunes only a portion of those weights on a much smaller, task-specific dataset, instead of training a new model entirely from scratch. What technique is being described?
Transfer learning is exactly this: training starts from weights already learned on a large, general dataset, and only a portion of those weights is fine-tuned on a much smaller, task-specific dataset, instead of training an entirely new model from scratch on the small dataset alone. A hash collision is an unrelated hash-table concept about two keys sharing a bucket. This reuse-pretrained-weights-and-fine-tune approach is exactly why transfer learning lets small, task-specific datasets still produce strong models.
2 / 5
During a design review, the team fine-tunes a large pretrained image classifier on a few thousand labeled medical scans, specifically because training an equally accurate model from scratch would require far more labeled data than is available. Which capability does this provide?
Transfer learning here provides strong task-specific performance from a small dataset, since the pretrained weights already encode general visual features that only need light fine-tuning for the new task. Training a new model from scratch on the medical scans alone would require far more labeled examples than a few thousand to reach comparable accuracy. This reuse-general-features-then-fine-tune behavior is exactly why transfer learning is favored when labeled data for the specific task is scarce.
3 / 5
In a code review, a dev notices a medical-imaging model is being trained entirely from randomly initialized weights on only a few thousand labeled scans, instead of starting from a large pretrained image classifier's weights and fine-tuning. What does this represent?
This is a missed transfer-learning opportunity, since starting from pretrained weights and fine-tuning would reach far better accuracy on the small labeled dataset than training from random initialization. A cache eviction policy is an unrelated concept about discarded cache entries. This train-from-random-initialization-on-a-small-dataset pattern is exactly the kind of avoidable underperformance a reviewer flags once labeled data is scarce.
4 / 5
An incident report shows a medical-imaging model trained from randomly initialized weights on a few thousand labeled scans performed far worse than expected, because so little labeled data was insufficient to learn general visual features from scratch. What practice would prevent this?
Starting training from a large pretrained image classifier's weights and fine-tuning on the labeled medical scans lets the model reuse already-learned general visual features instead of learning everything from scratch. Continuing to train from randomly initialized weights regardless of how few labeled scans are available is exactly what caused the poor performance described in this incident. This pretrained-weights-plus-fine-tuning approach is the standard fix once a small labeled dataset is confirmed to be insufficient for training from scratch.
5 / 5
During a PR review, a teammate asks why the team reaches for transfer learning instead of training a smaller, simpler model from scratch that's tailored exactly to the medical-imaging task. What is the reasoning?
Transfer learning trades a larger base model and some fine-tuning setup for reusing general features learned from vastly more data than the task-specific dataset alone provides, while training a smaller model from scratch is tailored to the task but is limited by how little task-specific labeled data exists. This is exactly why transfer learning is favored when task-specific labeled data is scarce, while training a smaller model from scratch remains viable when a large, task-specific labeled dataset is already available.
What does the "Transfer Learning Vocabulary" vocabulary exercise cover?
This exercise tests real IT vocabulary related to transfer learning vocabulary through 5 multiple-choice questions, each built from realistic workplace sentences rather than abstract definitions.
Is this vocabulary exercise free to use?
Yes. Every exercise on CoderSlingo, including this one, is completely free — no account, sign-up, or payment required.
How many questions does this exercise have?
This exercise has 5 questions. Each one shows a real-world sentence or scenario with multiple-choice options and an explanation once you answer.
What happens after I answer a question?
You'll see immediate feedback showing whether your answer was correct, along with a short explanation of why — then a button to move to the next question, and a full results screen at the end.
Can I retry the exercise if I get questions wrong?
Yes. Once you reach the results screen, click "Try again" to reset your answers and go through the exercise from the start as many times as you like.
Do I need to create an account to take this exercise?
No account is needed. Your answers are scored in your browser during the session — nothing is saved to a server, so you can jump straight in.
Is my progress saved if I leave the page?
No — progress within an exercise resets if you navigate away or reload. Each exercise is short enough to complete in a few minutes in one sitting.
Are these vocabulary exercises connected to other topics?
Yes — browse the full vocabulary exercises hub to find related modules covering adjacent IT topics and roles.
How is this different from reading a glossary or blog article?
Exercises like this one are active recall drills — you have to choose the correct term or phrasing yourself, which builds retention faster than passively reading a definition.
Where can I find more vocabulary exercises?
Browse the full Vocabulary exercises hub for hundreds of modules covering Agile, DevOps, security, databases, architecture, and more — organised by IT role and skill.