Pearlmutter & Siskind joint work on Automatic Differentiation (AD)

Code

J.M. Siskind, ‘R6RS-AD,’ Purdue University, 20 May 2010.
[github] [BibTeX]

J.M. Siskind, ‘Stalin∇,’ Purdue University, 1 April 2016.
[github] [BibTeX]

Grants

J.M. Siskind, PI, ‘Algorithmic Differentiation of Functional Programs,’ NSF award CCF-0438806, May 2005–April 2010, $420,000.
[announcement]

Journal Papers

B.A. Pearlmutter and J.M. Siskind, ‘Reverse-Mode AD in a Functional Framework: Lambda the Ultimate Backpropagator,’ ACM Transactions on Programming Languages and Systems (TOPLAS), 30(2):1–36, April 2008.
[copy of record] [pdf] [BibTeX]

J.M. Siskind and B.A. Pearlmutter, ‘Nesting Forward-Mode AD in a Functional Framework,’ Higher-Order and Symbolic Computation (HOSC), 21(4):361–376, December 2008.
[copy of record] [pdf] [BibTeX]

A.G. Baydin, B.A. Pearlmutter, A.A. Radul, and J.M. Siskind, ‘Automatic Differentiation in Machine Learning: a Survey,’ Journal of Machine Learning Research (JMLR), 18(153):1–43, April 2018.
[copy of record] [pdf] [BibTeX]

J.M. Siskind and B.A. Pearlmutter, ‘Divide-and-Conquer Checkpointing for Arbitrary Programs with No User Annotation,’ Optimization Methods and Software (OMS), 33(4–6):1288–1330, September 2018.
[copy of record] [pdf] [BibTeX]

O. Manzyuk, B.A. Pearlmutter, A.A. Radul, D.R. Rush, and J.M. Siskind, ‘Perturbation Confusion in Forward Automatic Differentiation of Higher-Order Functions,’ Journal of Functional Programming (JFP), 29(e12):1–24, September 2019.
[copy of record] [pdf] [BibTeX] [announcement] [slides] [video]

Conference Papers

J.M. Siskind and B.A. Pearlmutter, ‘Perturbation Confusion and Referential Transparency: Correct Functional Implementation of Forward-Mode AD,’ Draft Proceedings of the 17th International Workshop on Implementation and Application of Functional Languages (IFL), Dublin, Ireland, 19–21 September 2005.
[pdf] [BibTeX] [slides]

B.A. Pearlmutter and J.M. Siskind, ‘Lazy Multivariate Higher-Order Forward-Mode AD,’ Proceedings of the 34th Annual Symposium on Principles of Programming Languages (POPL), pp. 155–160, Nice, France, 17–19 January 2007.
[copy of record] [pdf] [BibTeX] [slides]

J.M. Siskind and B.A. Pearlmutter, ‘First-Class Nonstandard Interpretations by Opening Closures,’ Proceedings of the 34th Annual Symposium on Principles of Programming Languages (POPL), pp. 71–76, Nice, France, 17–19 January 2007.
[copy of record] [pdf] [BibTeX] [slides]

B.A. Pearlmutter and J.M. Siskind, ‘Using Programming Language Theory to Make AD Sound and Efficient,’ Proceedings of the 5th International Conference on Automatic Differentiation (AD), pp. 79–90, Bonn, Germany, 11–15 August 2008.
[copy of record] [pdf] [BibTeX]

D. Wingate, N. Goodman, A. Stuhlmüller, and J.M. Siskind, ‘Nonstandard Interpretations of Probabilistic Programs for Efficient Inference,’ Proceedings of the Neural Information Processing Systems Conference (NIPS), Granada, Spain, 12–15 December 2011.
[copy of record] [pdf] [BibTeX]

A. Radul, B.A. Pearlmutter, and J.M. Siskind, ‘AD in Fortran: Implementation via Prepreprocessor,’ Proceedings of the 6th International Conference on Automatic Differentiation (AD), pp. 273–284, Fort Collins, CO, 23–27 July 2012.
[copy of record] [pdf] [BibTeX] [slides]

J.M. Siskind and B.A. Pearlmutter, ‘Binomial Checkpointing for Arbitrary Programs with No User Annotation,’ Extended abstract, presented at the International Conference on Algorithmic Differentiation (AD), Oxford, UK, 12–15 September 2016.
[copy of record] [pdf] [BibTeX] [slides]

J.M. Siskind and B.A. Pearlmutter, ‘Efficient Implementation of a Higher-Order Language with Built-In AD,’ Extended abstract, presented at the International Conference on Algorithmic Differentiation (AD), Oxford, UK, 12–15 September 2016.
[copy of record] [pdf] [BibTeX] [slides]

A.G. Baydin, B.A. Pearlmutter, and J.M. Siskind, ‘DiffSharp: An AD Library for .NET Languages,’ Extended abstract, presented at the International Conference on Algorithmic Differentiation (AD), Oxford, UK, 12–15 September 2016.
[copy of record] [pdf] [BibTeX] [slides]

A.G. Baydin, B.A. Pearlmutter, and J.M. Siskind, ‘Tricks from Deep Learning,’ Extended abstract, presented at the International Conference on Algorithmic Differentiation (AD), Oxford, UK, 12–15 September 2016.
[copy of record] [pdf] [BibTeX] [slides]

R. Kelly, B.A. Pearlmutter, and J.M. Siskind, ‘Evolving the Incremental λ Calculus into a Model of Forward AD,’ Extended abstract, presented at the International Conference on Algorithmic Differentiation (AD), Oxford, UK, 12–15 September 2016.
[copy of record] [pdf] [BibTeX] [poster]

Workshop Presentations

J.M. Siskind, ‘AD Nesting, Variable Capture, Programming Language Theory, and AD,’ 2nd European Workshop on Automatic Differentiation, 18 November 2005.
[schedule] [slides] [BibTeX]

J.M. Siskind, ‘AD for Probabilistic Programming,’ NIPS Workshop on Probabilistic Programming: Universal Languages and Inference; systems; and applications, 13 December 2008.
[schedule] [slides] [BibTeX]

J.M. Siskind, ‘What every machine-learning researcher should know about AD,’ DALI Workshop on Probabilistic Programming, 10 April 2015.
[schedule] [slides] [BibTeX]

B.A. Pearlmutter, ‘Automatic Differentiation: History and Headroom,’ NIPS Workshop on The Future of Gradient-Based Machine Learning Software, 10 December 2016.
[schedule] [slides] [BibTeX] [video]

J.M. Siskind, ‘The tension between convenience and performance in automatic differentiation,’ NIPS Workshop on The Future of Gradient-Based Machine Learning Software, 10 December 2016.
[schedule] [slides] [BibTeX] [video]

J.M. Siskind, ‘Binomial Checkpointing for Arbitrary Programs with No User Annotation,’ SIAM Conference on Computational Science and Engineering (CSE) minisymposium on AD, 28 February 2017.
[schedule] [slides] [BibTeX]

J.M. Siskind, ‘Divide-and-Conquer Checkpointing for Arbitrary Programs with No User Annotation,’ NIPS Workshop on The Future of Gradient-Based Machine Learning Software & Techniques, 9 December 2017.
[schedule] [slides] [BibTeX]

J.M. Siskind, ‘Scheme as a framework for Deep Learning,’ ICFP Scheme and Functional Programming Workshop, 27 August 2021.
[schedule] [slides] [BibTeX] [video]

M. Maleki, B.A. Pearlmutter, and J.M. Siskind, ‘Adding AD to Scheme by differentiation the Interpreter,’ ICFP Scheme and Functional Programming Workshop, 27 August 2021.
[schedule] [BibTeX] [video]

Public Seminars

J.M. Siskind, ‘Backpropagation Through Functional Programs: How to do Reverse-Mode AD Correctly in a Functional Framework,’ New Jersey Programming Languages and Systems (NJPLS), 16 Sepember 2005.
[schedule] [abstract] [slides] [BibTeX]

J.M. Siskind, ‘Taking Derivatives of Functional Programs: AD in a Functional Framework,’ New England Programming Languages and Systems (NEPLS), 27 October 2005.
[schedule] [abstract] [slides] [BibTeX]

J.M. Siskind, ‘Algorithmic Differentiation of Functional Programs,’ NSF Science of Design (SoD) Program PI Meeting, 28 February–2 March 2007.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs or Lambda the Ultimate Calculus,’ New England Programming Languages and Systems (NEPLS), 21 November 2008.
[schedule] [abstract] [slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs and its use for Probabilistic Programming,’ New Jersey Programming Languages and Systems (NJPLS), 3 April 2009.
[schedule] [slides] [BibTeX]

J.M. Siskind, ‘AD for CV, AC, ML, AI, and CL,’ Science Foundation Ireland (SFI) Site Visit, 23 November 2012.
[slides] [BibTeX]

J.M. Siskind, ‘Probabilistic Programming: From Principled Foundations, Through Efficient Implementation, To Innovative Applications,’ DARPA Probabilistic Programming for Advancing Machine Learning (PPAML) Program PI Meeting, 10 July 2014.
[slides] [BibTeX]

J.M. Siskind, ‘The Next 700 Probabilistic Programming Languages,’ DARPA Probabilistic Programming for Advancing Machine Learning (PPAML) Program PI Meeting, 22 January 2015.
[slides] [BibTeX]

J.M. Siskind, ‘Nested Queries in Nondeterministic and Probabilistic Programming Languages,’ New England Programming Languages and Systems (NEPLS), 10 November 2015.
[schedule] [abstract] [slides] [BibTeX]

Seminars at Institutions

J.M. Siskind, ‘Algorithmic Differentiation of Functional Programs,’ Purdue University, 16 June 2004.
[slides] [BibTeX]

J.M. Siskind, ‘Algorithmic Differentiation of Functional Programs,’ University of Maryland, 17 June 2004.
[slides] [BibTeX]

J.M. Siskind, ‘Algorithmic Differentiation of Functional Programs,’ University of Toronto, 25 June 2004.
[slides] [BibTeX]

J.M. Siskind, ‘Algorithmic Differentiation of Functional Programs,’ Northeastern University, 8 August 2004.
[slides] [BibTeX]

J.M. Siskind, ‘Algorithmic Differentiation of Functional Programs,’ University of Illinois at Urbana-Champaign, 1 September 2004.
[slides] [BibTeX]

J.M. Siskind, ‘Taking Derivatives of Functional Programs: AD in a Functional Framework,’ University of Hertfordshire, 22 November 2005.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs or Lambda the Ultimate Calculus or How I Learned to Stop Worrying and Love Map-Closure,’ Indiana University, 20 February 2007.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs or Lambda the Ultimate Calculus,’ University of Chicago, 8 April 2008.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation for Probabilistic Programming,’ Indiana University, 29 January 2009.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs and its use for Probabilistic Programming,’ Jane Street Capital, 12 March 2009.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs and its use for Probabilistic Programming,’ DE Shaw, 13 March 2009.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs or Lambda the Ultimate Calculus,’ MIT, 17 March 2009.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs or Lambda the Ultimate Calculus,’ Yale University, 19 March 2009.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs or Lambda the Ultimate Calculus,’ University of Pennsylvania, 20 March 2009.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs and its use for Probabilistic Programming,’ Purdue University, 30 March 2009.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs and its use for Probabilistic Programming,’ Northeastern University, 1 April 2009.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs and its use for Probabilistic Programming,’ TTI Chicago, 17 April 2009.
[slides] [BibTeX]

J.M. Siskind, ‘Automatic Differentiation of Functional Programs and its use for Probabilistic Programming,’ University of South Carolina, 11 September 2009.
[slides] [BibTeX]

J.M. Siskind, ‘What does CPS have to do with deep learning?,’ Northeastern University, 11 September 2017.
[slides] [BibTeX]

J.M. Siskind, ‘What does CPS have to do with deep learning?,’ Amazon, Boston, 6 June 2018.
[slides] [BibTeX]

J.M. Siskind, ‘A Functional-Programming Framework for Deep Learning,’ Meta, 16 December 2021.
[slides] [BibTeX]

Technical Reports

J.M. Siskind and B.A. Pearlmutter, ‘Using Polyvariant Union-Free Flow Analysis to Compile a Higher-Order Functional-Programming Language with a First-Class Derivative Operator to Efficient Fortran-like Code,’ Technical Report TR-ECE-08-01, School of Electrical and Computer Engineering, Purdue University, 2008.
[pdf] [BibTeX]

J.M. Siskind and B.A. Pearlmutter, ‘Putting the Automatic Back into AD: Part I, What's Wrong,’ Technical Report TR-ECE-08-02, School of Electrical and Computer Engineering, Purdue University, 2008.
[pdf] [BibTeX]

B.A. Pearlmutter and J.M. Siskind, ‘Putting the Automatic Back into AD: Part II, Dynamic, Automatic, Nestable, and Fast,’ Technical Report TR-ECE-08-03, School of Electrical and Computer Engineering, Purdue University, 2008.
[pdf] [BibTeX]

J.M. Siskind and B.A. Pearlmutter, ‘Nesting Forward-Mode AD in a Functional Framework,’ Technical Report TR-ECE-08-09, School of Electrical and Computer Engineering, Purdue University, 2008.
[pdf] [BibTeX]

arXiv

A. Radul, B.A. Pearlmutter, and J.M. Siskind, ‘AD in Fortran, Part 1: Design,’ arXiv:1203.1448. 7 March 2012.
[pdf] [BibTeX]

A. Radul, B.A. Pearlmutter, and J.M. Siskind, ‘AD in Fortran, Part 2: Implementation via Prepreprocessor,’ arXiv:1203.1450, 7 March 2012.
[pdf] [BibTeX]

O. Manzyuk, B.A. Pearlmutter, A.A. Radul, D.R. Rush, and J.M. Siskind, ‘Confusion of Tagged Perturbations in Forward Automatic Differentiation of Higher-Order Functions,’ arXiv:1211.4892, 20 November 2012.
[pdf] [BibTeX]

A.G. Baydin, B.A. Pearlmutter, A.A. Radul, and J.M. Siskind, ‘Automatic differentiation in machine learning: a survey,’ arXiv:1502.05767, 19 April 2015.
[pdf] [BibTeX]

A.G. Baydin, B.A. Pearlmutter, and J.M. Siskind, ‘DiffSharp: Automatic Differentiation Library,’ arXiv:1511.07727, 26 November 2015.
[pdf] [BibTeX]

J.M. Siskind and B.A. Pearlmutter, ‘Binomial Checkpointing for Arbitrary Programs with No User Annotation,’ arXiv:1611.03410, 10 November 2016.
[pdf] [BibTeX]

J.M. Siskind and B.A. Pearlmutter, ‘Efficient Implementation of a Higher-Order Language with Built-In AD,’ arXiv:1611.03416, 10 November 2016.
[pdf] [BibTeX]

A.G. Baydin, B.A. Pearlmutter, and J.M. Siskind, ‘DiffSharp: An AD Library for .NET Languages,’ arXiv:1611.03423, 10 November 2016.
[pdf] [BibTeX]

R. Kelly, B.A. Pearlmutter, and J.M. Siskind, ‘Evolving the Incremental λ Calculus into a Model of Forward AD,’ arXiv:1611.03429, 10 November 2016.
[pdf] [BibTeX]

A.G. Baydin, B.A. Pearlmutter, and J.M. Siskind, ‘Tricks from Deep Learning,’ arXiv:1611.03777, 10 November 2016.
[pdf] [BibTeX]

J.M. Siskind and B.A. Pearlmutter, ‘Divide-and-Conquer Checkpointing for Arbitrary Programs with No User Annotation,’ arXiv:1708.06799, 22 August 2017.
[pdf] [BibTeX]