Implement MAML for adaptive regression with documentation and dependencies

[leonvanbokhorst]

Added documentation for the model_agnostic_meta_learning.py proof of concept,
highlighting its key features including:

• Meta-learning architecture with skip connections
• Task generation and adaptation capabilities
• Visualization and analysis tools
• Rapid context adaptation functionality

Summary by Sourcery

Introduce a Model-Agnostic Meta-Learning (MAML) implementation for rapid adaptation in narrative modeling, complete with documentation and necessary dependencies.

New Features:

• Implement a Model-Agnostic Meta-Learning (MAML) approach for rapid adaptation of narrative models to new contexts, featuring a meta-learning architecture with skip connections and adaptive learning rate scheduling.

Build:

• Update requirements.txt to include wandb and matplotlib for experiment tracking and visualization.

Documentation:

• Add detailed documentation for the model_agnostic_meta_learning.py proof of concept, highlighting its meta-learning architecture, task generation, and visualization tools.

[sourcery-ai]

Reviewer's Guide by Sourcery

This PR implements a Model-Agnostic Meta-Learning (MAML) system with a focus on rapid adaptation of narrative models. The implementation features a neural network architecture with skip connections, comprehensive visualization tools, and robust training components. The code is well-structured with detailed documentation and includes both the core MAML implementation and supporting utilities for task generation and analysis.

Sequence diagram for meta-training step

sequenceDiagram
    participant MetaModelGenerator
    participant TaskBatch
    participant Device
    participant FastWeights
    participant MetaOptimizer
    participant Scheduler

    MetaModelGenerator->>TaskBatch: Get task batch
    loop for each task in task_batch
        MetaModelGenerator->>FastWeights: Clone parameters
        loop Inner loop steps
            FastWeights->>MetaModelGenerator: forward_with_fast_weights(support_x)
            MetaModelGenerator->>FastWeights: Compute gradients
            FastWeights->>FastWeights: Update fast weights
        end
        FastWeights->>MetaModelGenerator: forward_with_fast_weights(query_x)
        MetaModelGenerator->>MetaOptimizer: Accumulate meta-gradients
    end
    MetaOptimizer->>MetaModelGenerator: Apply meta-update
    MetaModelGenerator->>Scheduler: Step with avg_loss

Loading

Class diagram for MetaModelGenerator

classDiagram
    class MetaModelGenerator {
        +int input_size
        +List~int~ hidden_sizes
        +int output_size
        +float inner_lr
        +float meta_lr
        +Linear input_layer
        +ModuleList hidden_layers
        +Linear output_layer
        +SGD meta_optimizer
        +ReduceLROnPlateau scheduler
        +forward(torch.Tensor) torch.Tensor
        +meta_train_step(List~Tuple~torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor~~, torch.device) Tuple~float, float~
        +forward_with_fast_weights(torch.Tensor, Dict~str, torch.Tensor~) torch.Tensor
        +compute_metrics(torch.Tensor, torch.Tensor) Dict~str, float~
        +visualize_adaptation(torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor, Dict~str, torch.Tensor~, str) void
    }

Loading

File-Level Changes

Change	Details	Files
Implemented core MAML architecture with skip connections and adaptive learning	Created MetaModelGenerator class with skip connections and multi-layer architecture Implemented bi-level optimization with inner/outer training loops Added adaptive learning rate scheduling and gradient clipping Implemented forward propagation with fast weights for quick adaptation	src/maml_model_agnostic_meta_learning.py
Added comprehensive task generation and evaluation system	Created synthetic task generation with controlled complexity Implemented support/query set splitting for evaluation Added task difficulty analysis functionality Created task-specific DataLoader implementation	src/maml_model_agnostic_meta_learning.py
Implemented visualization and analysis tools	Added feature importance visualization Created adaptation progress tracking plots Implemented error distribution analysis Added comprehensive metrics computation	src/maml_model_agnostic_meta_learning.py src/maml.md
Added project documentation and dependencies	Added detailed MAML documentation with implementation details Updated project dependencies to include visualization libraries Added README section describing MAML implementation	README.md requirements.txt src/maml.md

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time. You can also use
  this command to specify where the summary should be inserted.

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[sourcery-ai]

Hey @leonvanbokhorst - I've reviewed your changes - here's some feedback:

Overall Comments:

• The extracted visualization helper methods need cleanup - there are duplicate methods with TODO comments that should be properly refactored into a single well-named utility function.

Here's what I looked at during the review

• 🟡 General issues: 2 issues found
• 🟢 Security: all looks good
• 🟢 Testing: all looks good
• 🟡 Complexity: 1 issue found
• 🟢 Documentation: all looks good

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}