Convergence, Objective function and Black-box example

Questions
1

Dear all,

  1. Can someone please explain the role of the following line:
    -(0.5/sigma**2)*np.sum((data - model)**2)
    in the example given ‘Using a “black box” likelihood function’.

    Is it objective/fitness function ?

  2. Can I modify it to improve the convergence ?

  3. How can I see the convergence of the model in pymc3 ?

Thank You!