12 October 2024

Title: The transformational power of non-monotonic reasoning

Abstract:

We argue that NMR has an inherent power of transformation that has been under-exploited in LP, due to  what could be called LP's "analysis bias": a problem is solved, or knowledge is extracted, from a typically static snapshot of reality at a given time. 

We examine the reaches of overcoming this bias around the goal of making the world more equitably human-friendly. We argue that  to achieve the societal evolution needed to reach this goal, we will need to pay more attention to LP’s capabilities for NMR-aided generation, as opposed to analysis, in combination with a) the development of world patrimony knowledge repositories where properties such as truth can be certified, and b) societal literacy tools for relevant subjects not sufficiently taught at schools, such as inclusive history, civic education or media literacy. We illustrate our arguments around three projects that, in different ways, take such an approach: 

1. a novel view of grammatical inference, through abducing the grammar of an under-resourced language from the grammar of a well-studied one (with potential impact on languages on the verge of extinction);

2. LP tools for transforming social and ecological indicators of distress by generating sets of actions that can heal their causes; and 

3. teaching tools, including generative, that promote the use of logic-based computing to make logical reasoning and trustworthy coding skills more universally available. 

N.B. Project 3 will be discussed in much more detail at PEG 2.0 on October 13, see https://prolog-lang.org/Education/PrologEducationWS2024.html - do not miss it!

13 October 2024

Title: TBA

Abstract: TBA

14 October 2024

09:00-10:00

Title: Logic Programming and Logical Algorithmics

Abstract:

Logic programming, born around 1972, expresses a program as a set of Horn clauses. Computation is then performed by applying logical reasoning to that set of clauses. The approach was eventually described as "Algorithm=Logic+Control". Another approach to logic and algorithms was developed by Codd in the early 1970s. In his approach, the problem domain is expressed as a relational database, and the problem is expressed as a first-order formula, called "query". Computation is performed by a meta-algorithm, the query-evaluation algorithm.  In this talk, I will describe this approach, which I call Logical Algorithmics, in detail. I will show how this approach yielded multi-variate computational-complexity theory, which offers a more nuanced approach to complexity analysis. It also enabled the development the model-checking algorithms, which are today used in industrial semiconductor design tools.