Nov 12, 2022 09:00 AM - 11:45 AM(America/New_York)
20221112T090020221112T1145America/New_YorkPhysical Signatures of Computation
In recent years, a new generation of scholars have begun searching for physical signatures of computation. That is, they have begun investigating what it takes for a physical system to implement a computation with unprecedented attention to the scientific practices involving computation, including computer science and engineering, in hopes of identifying physical differences between systems that implement computations and systems that don't. This symposium will introduce recent progress in this area to a wider audience. It will illustrate how our understanding of physical computation has deepened and become more sophisticated and how it can be informed by scientific practices that were not on the horizon of most philosophers of science until quite recently. The participants in this symposium are also engaged in a lively debate with one another, which will stimulate both them and the audience to make further progress. Providing an adequate account of computational implementation has real implications for the foundations of the computational theory of cognition, the notion of biological computation, the construction of novel forms of computers, the foundations of physics, and more. Thus, this symposium is likely to attract a wide audience.
In recent years, a new generation of scholars have begun searching for physical signatures of computation. That is, they have begun investigating what it takes for a physical system to implement a computation with unprecedented attention to the scientific practices involving computation, including computer science and engineering, in hopes of identifying physical differences between systems that implement computations and systems that don't. This symposium will introduce recent progress in this area to a wider audience. It will illustrate how our understanding of physical computation has deepened and become more sophisticated and how it can be informed by scientific practices that were not on the horizon of most philosophers of science until quite recently. The participants in this symposium are also engaged in a lively debate with one another, which will stimulate both them and the audience to make further progress. Providing an adequate account of computational implementation has real implications for the foundations of the computational theory of cognition, the notion of biological computation, the construction of novel forms of computers, the foundations of physics, and more. Thus, this symposium is likely to attract a wide audience.
Implementation, individuation, and triviality in computational theoriesView Abstract SymposiumPhilosophy of Computer Science09:00 AM - 11:45 AM (America/New_York) 2022/11/12 14:00:00 UTC - 2022/11/12 16:45:00 UTC
Distinguishing between physical systems that compute and those which do not requires an explanation that posits the relation between the formal concept of computation and the physical implementing system. There is confusion about how an answer to the implementation question is to be articulated leading to claims about implementation which are not sufficiently distinguished from claims about individuation. I argue that confusions about computational triviality have, in part, given rise to this conflation. In this paper, I demonstrate that there are two distinct types of triviality: a trivialization of the implementation relation and a trivialization of the individuation conditions.
Presenters Danielle J. Williams Presenting Author, University Of California, Davis
Some Myths of Symbolic ComputationView Abstract SymposiumPhilosophy of Computer Science09:00 AM - 11:45 AM (America/New_York) 2022/11/12 14:00:00 UTC - 2022/11/12 16:45:00 UTC
It is shown that supposedly paradigmatic examples of classic architecture do not contain local representations. In particular, Turing Machines (TMs) carry out transformations over sub-symbols where only the initial and final states may involve interpretable strings. In contrast, examples of computing systems with local representations lack the coding efficiency that is claimed to be paradigmatic of classical architectures. Thus, distributed, sub-symbolic computation should also be considered as a hallmark of classical architectures. In light of this and other commonalities, it is proposed that the traditional divide between connectionist and classical architectures is more apparent than real.
J. Brendan Ritchie Presenting Author, National Institute Of Mental Health
Analog Computation, Continuous or Empirical: The perspective of Carnapian ExplicationView Abstract SymposiumPhilosophy of Computer Science09:00 AM - 11:45 AM (America/New_York) 2022/11/12 14:00:00 UTC - 2022/11/12 16:45:00 UTC
We discuss two different ways that the term “analog” (as opposed to “digital”) is used in the methodology of computer science and those engineering disciplines that are related to computer science. We show that formal models of computation on real numbers provide, indeed, an explication of what corresponds to the intuition that certain devices operating on continuous quantities perform computations. We call this “the analog continuous thesis” (“the AN-C thesis”), and we show how it is similar to other theses used to explicate computation, such as the Church-Turing thesis or the Cobham-Edmonds thesis.
Paweł Stacewicz Presenting Co-Author, Warsaw University Of Technology
The Robust Mapping Account of ImplementationView Abstract SymposiumPhilosophy of Computer Science09:00 AM - 11:45 AM (America/New_York) 2022/11/12 14:00:00 UTC - 2022/11/12 16:45:00 UTC
According to the robust mapping account we propose, a mapping from physical to computational states is a legitimate basis for implementation only if it includes only physical states relevant to the computation, the physical states have enough spatiotemporal structure to map onto the structure of the computational states, and the evolving physical states bear neither more nor less information about the evolving computation than do the computational states they map onto. When these conditions are in place, a physical system can be said to implement a computation in a robust sense, which does not trivialize the notion of implementation.
Neal Anderson Presenting Co-Author, University Of Massachusetts Amherst
Computation with Neural ManifoldsView Abstract SymposiumPhilosophy of Computer Science09:00 AM - 11:45 AM (America/New_York) 2022/11/12 14:00:00 UTC - 2022/11/12 16:45:00 UTC
Recent research in cognitive neuroscience has uncovered so-called neural manifolds that play a central role in explanations of behavior. Revealed through the use of a range of dimensionality reduction techniques, these manifolds are entities in low-dimensional spaces contained in high-dimensional neural spaces. In this paper, I explore a possible computational interpretation for the role of manifolds in cognition. I argue that manifolds provide evidence for what neural computations are performed. I then turn to argue that manifolds also provide evidence for how inputs are transformed into outputs during neural computation.
Danielle J. Williams