Write a Blog >>
ICFP 2020
Thu 20 - Fri 28 August 2020
Tue 25 Aug 2020 12:26 - 12:37 at ICFP NY 3 - New York 3 (JFP talks) Chair(s): Jeremy Gibbons
Tue 25 Aug 2020 23:26 - 23:37 at ICFP Asia 3 - Asia 3 (JFP talks) Chair(s): Jeremy Gibbons

Automatic differentiation (AD) is a technique for augmenting computer programs to compute derivatives. The essence of AD in its forward accumulation mode is to attach perturbations to each number, and propagate these through the computation by overloading the arithmetic operators. When derivatives are nested, the distinct derivative calculations, and their associated perturbations, must be distinguished. This is typically accomplished by creating a unique tag for each derivative calculation and tagging the perturbations. We exhibit a subtle bug, present in fielded implementations which support derivatives of higher-order functions, in which perturbations are confused despite the tagging machinery, leading to incorrect results. The essence of the bug is as follows: a unique tag is needed for each derivative calculation, but in existing implementations unique tags are created when taking the derivative of a function at a point. When taking derivatives of higher-order functions, these need not correspond! We exhibit a simple example: a higher-order function f whose derivative at a point x, namely f′(x), is itself a function which calculates a derivative. This situation arises naturally when taking derivatives of curried functions. Two potential solutions are presented, and their deficiencies discussed. One uses eta expansion to delay the creation of fresh tags in order to put them into one-to-one correspondence with derivative calculations. The other wraps outputs of derivative operators with tag substitution machinery. Both solutions seem very difficult to implement without violating the desirable complexity guarantees of forward AD.

Tue 25 Aug

Displayed time zone: Eastern Time (US & Canada) change

11:30 - 13:00
New York 3 (JFP talks)ICFP Program at ICFP NY 3
Chair(s): Jeremy Gibbons Department of Computer Science, University of Oxford

Public livestreams: YouTube, Bilibili (China)

11:30
11m
Talk
A theory of RPC calculi for client–server modelJFP
ICFP Program
DOI Media Attached
11:41
11m
Talk
The full-reducing Krivine abstract machine KN simulates pure normal-order reduction in lockstep: A proof via corresponding calculusJFP
ICFP Program
Álvaro García Perez IMDEA Software Institute, Pablo Nogueira ESNE University School of Design, Innovation and Technology
DOI Media Attached
11:52
11m
Talk
Local algebraic effect theoriesJFP
ICFP Program
Žiga Lukšič , Matija Pretnar University of Ljubljana, Slovenia
DOI Media Attached
12:03
11m
Talk
Heterogeneous binary random-access listsJFP
ICFP Program
Wouter Swierstra Utrecht University, Netherlands
DOI Media Attached
12:15
11m
Talk
POPLMark reloaded: Mechanizing proofs by logical relationsJFP
ICFP Program
Andreas Abel Gothenburg University, Guillaume Allais University of St Andrews, Aliya Hameer McGill University, Brigitte Pientka McGill University, Alberto Momigliano Università degli Studi di Milano, Steven Schäfer Google, Aarhus, Kathrin Stark Princeton University, USA
DOI Media Attached
12:26
11m
Talk
Perturbation confusion in forward automatic differentiation of higher-order functionsJFP
ICFP Program
Oleksandr Manzyuk , Barak A. Pearlmutter Maynooth University, Alexey Radul , David Rush , Jeffrey Mark Siskind School of Electrical and Computer Engineering, Purdue University
DOI Media Attached
12:37
11m
Talk
Elastic Sheet-Defined Functions: Generalising Spreadsheet Functions to Variable-Size Input ArraysJFP
ICFP Program
Matt McCutchen , Judith Borghouts , Andrew D. Gordon Microsoft Research and University of Edinburgh, Simon Peyton Jones Microsoft, UK, Advait Sarkar Microsoft Research and University of Cambridge
DOI Pre-print Media Attached
12:48
11m
Talk
Emerging languages: An alternative approach to teaching programming languagesJFP
ICFP Program
Saverio Perugini University of Dayton
DOI Media Attached
22:30 - 00:00
Asia 3 (JFP talks)ICFP Program at ICFP Asia 3
Chair(s): Jeremy Gibbons Department of Computer Science, University of Oxford

Public livestreams: YouTube, Bilibili (China)

22:30
11m
Talk
A theory of RPC calculi for client–server modelJFP
ICFP Program
DOI Media Attached
22:41
11m
Talk
The full-reducing Krivine abstract machine KN simulates pure normal-order reduction in lockstep: A proof via corresponding calculusJFP
ICFP Program
Álvaro García Perez IMDEA Software Institute, Pablo Nogueira ESNE University School of Design, Innovation and Technology
DOI Media Attached
22:52
11m
Talk
Local algebraic effect theoriesJFP
ICFP Program
Žiga Lukšič , Matija Pretnar University of Ljubljana, Slovenia
DOI Media Attached
23:03
11m
Talk
Heterogeneous binary random-access listsJFP
ICFP Program
Wouter Swierstra Utrecht University, Netherlands
DOI Media Attached
23:15
11m
Talk
POPLMark reloaded: Mechanizing proofs by logical relationsJFP
ICFP Program
Andreas Abel Gothenburg University, Guillaume Allais University of St Andrews, Aliya Hameer McGill University, Brigitte Pientka McGill University, Alberto Momigliano Università degli Studi di Milano, Steven Schäfer Google, Aarhus, Kathrin Stark Princeton University, USA
DOI Media Attached
23:26
11m
Talk
Perturbation confusion in forward automatic differentiation of higher-order functionsJFP
ICFP Program
Oleksandr Manzyuk , Barak A. Pearlmutter Maynooth University, Alexey Radul , David Rush , Jeffrey Mark Siskind School of Electrical and Computer Engineering, Purdue University
DOI Media Attached
23:37
11m
Talk
Elastic Sheet-Defined Functions: Generalising Spreadsheet Functions to Variable-Size Input ArraysJFP
ICFP Program
Matt McCutchen , Judith Borghouts , Andrew D. Gordon Microsoft Research and University of Edinburgh, Simon Peyton Jones Microsoft, UK, Advait Sarkar Microsoft Research and University of Cambridge
DOI Pre-print Media Attached
23:48
11m
Talk
Emerging languages: An alternative approach to teaching programming languagesJFP
ICFP Program
Saverio Perugini University of Dayton
DOI Media Attached