Skip to content

Feedback Module

Overview

The Feedback Module is an integral component designed to collect and process feedback for actions performed by the agent. It enables the system to evaluate outcomes, track performance, and adapt behavior based on historical data. This module is vital for improving the agent's decision-making and maintaining accountability.

The Feedback Module is implemented in src/feedback/feedback_module.py and provides structured methods for collecting, storing, and analyzing feedback related to agent actions.

How It Works

The Feedback Module operates as follows:

  1. Feedback Collection: Captures feedback based on the action performed and its outcome, assigning a feedback score to evaluate success or failure.

  2. Feedback Storage: Maintains an internal history of feedback entries, including details about the action, its outcome, and the assigned feedback score.

  3. Feedback Retrieval: Allows querying of recent feedback entries for analysis and monitoring.

  4. Feedback Reset: Provides functionality to clear feedback history, ensuring the module can be reset when needed.

Technical Features

1. Feedback Collection

The collect_feedback method records feedback for a specific action and its outcome. It assigns a feedback score based on predefined criteria:

  • Failure: Assigned a score of -1.0 if the outcome is None.
  • Success: Assigned a score of 1.0 for successful outcomes.

Implementation:

feedback_score = -1.0 if outcome is None else 1.0

  • Inputs:

  • action (str): The name of the action performed.

  • outcome (Any): The outcome of the action; None for failure or a value indicating success.

  • Output:

  • Returns a feedback score (float) for the action.

  • Example Usage:

feedback_score = feedback_module.collect_feedback("fetch_data", outcome=data)

2. Feedback History Retrieval

The get_feedback_history method retrieves the most recent feedback entries, limited by the specified count.

Features:

  • Enables monitoring of agent performance.
  • Default limit is set to 10 entries.

Example:

recent_feedback = feedback_module.get_feedback_history(limit=5)

3. Feedback Reset

The reset_feedback_history method clears the internal feedback history, resetting the module's state. This is useful for testing or reinitializing feedback tracking.

Example:

feedback_module.reset_feedback_history()

Key Methods

collect_feedback

Captures feedback for a given action and outcome.

Arguments:

  • action (str): The action name.
  • outcome (Optional[Any]): The outcome of the action.

Returns:

  • float: Feedback score.

get_feedback_history

Retrieves recent feedback entries.

Arguments:

  • limit (int): Number of entries to retrieve (default: 10).

Returns:

  • List[Dict[str, Any]]: Recent feedback entries.

reset_feedback_history

Clears the feedback history.

Returns:

  • None

Example Workflow

  1. An action is performed by the agent (e.g., fetch_data).
  2. The outcome of the action is evaluated, and feedback is collected: 
    feedback_module.collect_feedback("fetch_data", outcome=data)
  3. The collected feedback is stored in the feedback history for analysis.
  4. The feedback history can be retrieved for review: 
    recent_feedback = feedback_module.get_feedback_history(limit=5)
  5. Feedback history is reset when needed: 
    feedback_module.reset_feedback_history()

Benefits

  • Performance Monitoring: Tracks agent actions and their outcomes, enabling better performance evaluation.
  • Adaptability: Facilitates improvements by learning from historical data.
  • Simplicity: Provides clear and structured methods for feedback collection and retrieval.

Best Practices

  1. Consistent Feedback Collection: Ensure feedback is collected for all critical actions to build a comprehensive history.
  2. Periodic Reset: Use reset_feedback_history to clear stale feedback during testing or major updates.
  3. Logging: Leverage loguru for detailed insights into feedback processing.

Known Limitations

  • Scoring Granularity: Feedback scores are currently binary (-1.0 or 1.0). Future versions can introduce more nuanced scoring.
  • Debugging Notes: Debugging comments in collect_feedback indicate areas for improvement.

If you have any questions or need further assistance, please refer to the GitHub Discussions.