Scientific Abstract Critique

Critique a small-sample teaching-method abstract for bias and flawed conclusions, comparing gpt-4o built-in chain-of-thought with a deepseek-reasoner reasoning model.

Demonstrates built-in and DeepSeek-backed reasoning for methodology critique.

scientific_research.py

"""
Scientific Abstract Critique
============================

Demonstrates built-in and DeepSeek-backed reasoning for methodology critique.
"""

from agno.agent import Agent
from agno.models.deepseek import DeepSeek
from agno.models.openai import OpenAIChat

# ---------------------------------------------------------------------------
# Create Agents
# ---------------------------------------------------------------------------
task = (
    "Read the following abstract of a scientific paper and provide a critical evaluation of its methodology,"
    "results, conclusions, and any potential biases or flaws:\n\n"
    "Abstract: This study examines the effect of a new teaching method on student performance in mathematics. "
    "A sample of 30 students was selected from a single school and taught using the new method over one semester. "
    "The results showed a 15% increase in test scores compared to the previous semester. "
    "The study concludes that the new teaching method is effective in improving mathematical performance among high school students."
)

cot_agent = Agent(
    model=OpenAIChat(id="gpt-4o"),
    reasoning=True,
    markdown=True,
)

deepseek_agent = Agent(
    model=OpenAIChat(id="gpt-4o"),
    reasoning_model=DeepSeek(id="deepseek-reasoner"),
    markdown=True,
)

# ---------------------------------------------------------------------------
# Run Agents
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    print("=== Built-in Chain Of Thought ===")
    cot_agent.print_response(task, stream=True, show_full_reasoning=True)

    print("\n=== DeepSeek Reasoning Model ===")
    deepseek_agent.print_response(task, stream=True, show_full_reasoning=True)

Run the Example

1. Set up your virtual environment

   Mac

   uv venv --python 3.12
   source .venv/bin/activate
   

   uv venv --python 3.12
   .venv\Scripts\activate
   
2. Install dependencies

   uv pip install -U agno openai
   
3. Export your API keys

   Mac/Linux

   export DEEPSEEK_API_KEY="your_deepseek_api_key_here"
   export OPENAI_API_KEY="your_openai_api_key_here"
   

   $Env:DEEPSEEK_API_KEY="your_deepseek_api_key_here"
   $Env:OPENAI_API_KEY="your_openai_api_key_here"
   
4. Run the example

   Save the code above as scientific_research.py, then run:

   python scientific_research.py
   

Full source: cookbook/10_reasoning/agents/scientific_research.py