001038109 001__ 1038109
001038109 005__ 20250203133240.0
001038109 0247_ $$2datacite_doi$$a10.34734/FZJ-2025-01156
001038109 037__ $$aFZJ-2025-01156
001038109 082__ $$a500
001038109 1001_ $$0P:(DE-Juel1)192147$$aLohoff, Jamie$$b0
001038109 1112_ $$a38th Conference on Neural Information Processing Systems$$cVancouver$$d2024-12-09 - 2024-12-16$$gNeurIPS$$wCanada
001038109 245__ $$aOptimizing Automatic Differentiation with Deep Reinforcement Learning
001038109 260__ $$c2024
001038109 300__ $$an/a
001038109 3367_ $$2ORCID$$aCONFERENCE_PAPER
001038109 3367_ $$033$$2EndNote$$aConference Paper
001038109 3367_ $$2BibTeX$$aINPROCEEDINGS
001038109 3367_ $$2DRIVER$$aconferenceObject
001038109 3367_ $$2DataCite$$aOutput Types/Conference Paper
001038109 3367_ $$0PUB:(DE-HGF)8$$2PUB:(DE-HGF)$$aContribution to a conference proceedings$$bcontrib$$mcontrib$$s1738249743_31342
001038109 3367_ $$0PUB:(DE-HGF)7$$2PUB:(DE-HGF)$$aContribution to a book$$mcontb
001038109 4900_ $$aAdvances in neural information processing systems
001038109 500__ $$aAccepted as a spotlight paper.
001038109 520__ $$aComputing Jacobians with automatic differentiation is ubiquitous in many scientific domains such as machine learning, computational fluid dynamics, robotics, and finance. Even small savings in the number of computations or memory usage in Jacobian computations can already incur massive savings in energy consumption and runtime. While there exist many methods that allow for such savings, they generally trade computational efficiency for approximations of the exact Jacobian. In this paper, we present a novel method to optimize the number of necessary multiplications for Jacobian computation by leveraging deep reinforcement learning (RL) and a concept called cross-country elimination while still computing the exact Jacobian. Cross-country elimination is a framework for automatic differentiation that phrases Jacobian accumulation as ordered elimination of all vertices on the computational graph where every elimination incurs a certain computational cost. We formulate the search for the optimal elimination order that minimizes the number of necessary multiplications as a single player game which is played by an RL agent. We demonstrate that this method achieves up to 33% improvements over state-of-the-art methods on several relevant tasks taken from diverse domains. Furthermore, we show that these theoretical gains translate into actual runtime improvements by providing a cross-country elimination interpreter in JAX that can efficiently execute the obtained elimination orders.
001038109 536__ $$0G:(DE-HGF)POF4-5234$$a5234 - Emerging NC Architectures (POF4-523)$$cPOF4-523$$fPOF IV$$x0
001038109 536__ $$0G:(BMBF)16ME0400$$aBMBF 16ME0400 - Verbundprojekt: Neuro-inspirierte Technologien der künstlichen Intelligenz für die Elektronik der Zukunft - NEUROTEC II - (16ME0400)$$c16ME0400$$x1
001038109 536__ $$0G:(EU-Grant)953775$$aGREENEDGE - Taming the environmental impact of mobile networks through GREEN EDGE computing platforms (953775)$$c953775$$fH2020-MSCA-ITN-2020$$x2
001038109 7001_ $$0P:(DE-Juel1)188273$$aNeftci, Emre$$b1
001038109 773__ $$0PERI:(DE-600)1012320-9$$v38$$x1049-5258$$y2024
001038109 8564_ $$uhttps://neurips.cc/virtual/2024/poster/94064
001038109 8564_ $$uhttps://juser.fz-juelich.de/record/1038109/files/AlphaGrad_camera_rdy.pdf$$yOpenAccess
001038109 909CO $$ooai:juser.fz-juelich.de:1038109$$pdnbdelivery$$pec_fundedresources$$pVDB$$pdriver$$popen_access$$popenaire
001038109 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)192147$$aForschungszentrum Jülich$$b0$$kFZJ
001038109 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)188273$$aForschungszentrum Jülich$$b1$$kFZJ
001038109 9131_ $$0G:(DE-HGF)POF4-523$$1G:(DE-HGF)POF4-520$$2G:(DE-HGF)POF4-500$$3G:(DE-HGF)POF4$$4G:(DE-HGF)POF$$9G:(DE-HGF)POF4-5234$$aDE-HGF$$bKey Technologies$$lNatural, Artificial and Cognitive Information Processing$$vNeuromorphic Computing and Network Dynamics$$x0
001038109 9141_ $$y2024
001038109 915__ $$0StatID:(DE-HGF)0510$$2StatID$$aOpenAccess
001038109 915__ $$0StatID:(DE-HGF)0300$$2StatID$$aDBCoverage$$bMedline
001038109 920__ $$lyes
001038109 9201_ $$0I:(DE-Juel1)PGI-15-20210701$$kPGI-15$$lNeuromorphic Software Eco System$$x0
001038109 980__ $$acontrib
001038109 980__ $$aVDB
001038109 980__ $$aUNRESTRICTED
001038109 980__ $$acontb
001038109 980__ $$aI:(DE-Juel1)PGI-15-20210701
001038109 9801_ $$aFullTexts