Nikos Karalias

Nikos Karalias

I’m a postdoctoral researcher at MIT working with Stefanie Jegelka. I work on solving hard combinatorial optimization problems using machine learning methods. I’m currently on the academic job market. Current CV (PDF)

Brush-written formula F = c₁ ∧ c₂ ∧ c₃ ∧ c₄ beside its K4 graph, with clause nodes c₁–c₄ and edges x₁–x₆. Clauses: c₁ = (x₁ ⊕ x₂ ⊕ x₃); c₂ = (¬x₁ ⊕ x₄ ⊕ x₅); c₃ = (¬x₂ ⊕ x₄ ⊕ x₆); c₄ = (¬x₃ ⊕ x₅ ⊕ x₆).

#Learning with constraints    #Neural Networks    #Combinatorial optimization    #Boolean satisfiability   

Geometric Algorithms for Neural Combinatorial Optimization

#Learning with constraints    #Neural Networks    #Combinatorial optimization   

Are Neural Networks Optimal Approximation Algorithms?

#Learning with approximation guarantees    #Neural Networks    #Combinatorial optimization   

Learning with Neural Extensions

#Learning with discrete functions    #Neural Networks    #Combinatorial optimization   

Erdos Goes Neural

#Erdős goes neural    #Combinatorial Optimization    #Neural Networks