Utilizing advanced geometric deep learning methodologies, invariant point atention is employed to predict the binding affinity (ΔG) of protein complexes through the detailed three-dimensional characterization of their structures.