TY - GEN
T1 - Lagrangian Duality for Constrained Deep Learning
AU - Fioretto, Ferdinando
AU - Van Hentenryck, Pascal
AU - Mak, Terrence W.K.
AU - Tran, Cuong
AU - Baldo, Federico
AU - Lombardi, Michele
N1 - Publisher Copyright:
© 2021, Springer Nature Switzerland AG.
PY - 2021
Y1 - 2021
N2 - This paper explores the potential of Lagrangian duality for learning applications that feature complex constraints. Such constraints arise in many science and engineering domains, where the task amounts to learning to predict solutions for constraint optimization problems which must be solved repeatedly and include hard physical and operational constraints. The paper also considers applications where the learning task must enforce constraints on the predictor itself, either because they are natural properties of the function to learn or because it is desirable from a societal standpoint to impose them. This paper demonstrates experimentally that Lagrangian duality brings significant benefits for these applications. In energy domains, the combination of Lagrangian duality and deep learning can be used to obtain state of the art results to predict optimal power flows, in energy systems, and optimal compressor settings, in gas networks. In transprecision computing, Lagrangian duality can complement deep learning to impose monotonicity constraints on the predictor without sacrificing accuracy. Finally, Lagrangian duality can be used to enforce fairness constraints on a predictor and obtain state-of-the-art results when minimizing disparate treatments.
AB - This paper explores the potential of Lagrangian duality for learning applications that feature complex constraints. Such constraints arise in many science and engineering domains, where the task amounts to learning to predict solutions for constraint optimization problems which must be solved repeatedly and include hard physical and operational constraints. The paper also considers applications where the learning task must enforce constraints on the predictor itself, either because they are natural properties of the function to learn or because it is desirable from a societal standpoint to impose them. This paper demonstrates experimentally that Lagrangian duality brings significant benefits for these applications. In energy domains, the combination of Lagrangian duality and deep learning can be used to obtain state of the art results to predict optimal power flows, in energy systems, and optimal compressor settings, in gas networks. In transprecision computing, Lagrangian duality can complement deep learning to impose monotonicity constraints on the predictor without sacrificing accuracy. Finally, Lagrangian duality can be used to enforce fairness constraints on a predictor and obtain state-of-the-art results when minimizing disparate treatments.
UR - http://www.scopus.com/inward/record.url?scp=85103240731&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85103240731&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-67670-4_8
DO - 10.1007/978-3-030-67670-4_8
M3 - Conference contribution
AN - SCOPUS:85103240731
SN - 9783030676698
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 118
EP - 135
BT - Machine Learning and Knowledge Discovery in Databases. Applied Data Science and Demo Track - European Conference, ECML PKDD 2020, Proceedings
A2 - Dong, Yuxiao
A2 - Mladenic, Dunja
A2 - Saunders, Craig
PB - Springer Science and Business Media Deutschland GmbH
T2 - European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2020
Y2 - 14 September 2020 through 18 September 2020
ER -