Module 7 Assignment: Discriminative vs generative comparison

Module 7 Assignment: Discriminative vs generative comparison#

Theme#

Generative models and applications

Scenario#

A team is considering generated text, images, or synthetic records and needs to distinguish useful generation from unsupported claims.

Exercises#

  1. Choose one task and describe discriminative and generative approaches.

  2. Use the starter code to sample from a simple latent-variable generator.

  3. Explain how evaluation differs across classifiers, autoencoders, diffusion models, and language models.

  4. Identify one misuse risk and one documentation safeguard.

Evidence Requirements#

  • A short explanation of the data, tensors, objective, and evaluation signal used in the starter experiment.

  • At least one meaningful modification to the starter code, with the changed variable named explicitly.

  • A comparison against the unmodified starter result or another defensible baseline.

  • A limitation statement that separates what the toy experiment demonstrates from what a production model would require.

Submission#

Submit a 600-900 word technical memo plus code, plots, tables, or shape traces needed to support your claims. The memo should read like a review artifact for another AI practitioner: concise, reproducible, and honest about uncertainty.

Rubric Focus#

  • Technical correctness and appropriate neural-network vocabulary.

  • Evidence from the starter experiment or a documented extension.

  • Connection between design choices and data/problem structure.

  • Clear treatment of limitations, failure modes, or next experimental gates.

import torch
from torch import nn

torch.manual_seed(7)
generator = nn.Sequential(nn.Linear(2, 8), nn.Tanh(), nn.Linear(8, 2))
z = torch.randn(6, 2)
samples = generator(z).detach()
print(samples.round(decimals=3))
print("These are untrained synthetic samples; describe what training objective would make them useful.")

tensor([[ 0.2320, -0.5090],
        [ 0.5230, -0.0620],
        [ 0.5890, -0.3290],
        [ 0.4260, -0.4030],
        [ 0.1220, -0.5300],
        [ 0.0290, -0.5160]])
These are untrained synthetic samples; describe what training objective would make them useful.

Reflection Prompts#

  • What changed when you modified the starter experiment, and why should that change matter?

  • Which result surprised you, and what diagnostic would you run next?

  • What assumption would you document before handing this model to another practitioner?

  • Which failure mode from the module reading is most relevant to your result?