Research Seminar Logic and AI

2022-10-04T08:28:28Z

Ali Elhalawati:

{{#leftsidenav:{{LeftsidenavHeading|Contents}}
<div class="no-toc-head">__TOC__</div>
|{{Main heading|Research Seminar Logic and AI}}

In this research seminar, we invite speakers to present their research on topics related to logic and artificial intelligence. This includes presentations of published or accepted papers, rehearsals for future talks at conferences, defenses of very good student theses, as well as presentations of ongoing research. Coffee is usually provided. All are welcome.

The seminar is organized by the [[Knowledge-Based Systems|Knowledge-Based Systems]] group. Please contact [[Piotr Ostropolski-Nalewaja/en|Piotr Ostropolski-Nalewaja]] if you would like to give a talk or invite a speaker.

===Schedule and Location===

The seminar takes place most Thursdays from 11.00 – 12.00. Because of the ongoing pandemic, this research seminar will take place online until further notice. If you want to attend a talk, please contact [[Piotr Ostropolski-Nalewaja/en|Piotr Ostropolski-Nalewaja]] to obtain the link leading to the online room.
Please see the detailed program for information on upcoming seminars.

====Recent and Upcoming Talks====

<table class="table" >
<tr><th>Date</th><th>Speaker</th><th>Topic</th></tr>
{{#ask: [[Event series::Research Seminar Logic and AI]]
| ?Titel EN=
| ?Datum#ISO=
| ?Vortragender=
| sort=Datum
| order=desc
| format=plainlist
| template=Talk short inline/en
| searchlabel=
| link=none
| limit=5
}}
<tr><td></td><td>{{#ask: [[Event series::Research Seminar Logic and AI]]|
| ?Datum=Date
| ?=Title
| ?Vortragender=Speaker
| mainlabel=-
| sort=Datum | order=desc | limit=0 | searchlabel=View older talks …}}</td><td>{{Upcoming events feed
| L=/en
| query=[[Event series::Research Seminar Logic and AI]]
| label=Subscribe to upcoming talks (iCal)
| calendartitle=ICCL Research Seminar Logic and AI
}}</td></tr>
</table>

===Calendar===

{{#ask:[[Event series::Research Seminar Logic and AI]][[Start::+]]
| ?Start
| ?Ende
| ?Titel EN=title
| ?Raum=Room
| ?Veranstaltungsart=<br/>Type
| ?Vortragender=<br/>Speaker
| ?Beschreibung EN=<br/>Details
| format=eventcalendar
| firstday=Monday
| defaultview=month
| limit=1000
| link=subject
}}

}}

{{#set:
|Title=Research Seminar Logic and AI
|Forschungsgruppe=Computational Logic
|Lecturer=Sebastian Rudolph
|Lecturer=Thomas Feller
|Modul=INF-AQUA
|Modul=INF-PM-FOR
|Term=WS
|Year=2020
|SWSLecture=0
|SWSExercise=2
|SWSPractical=0
}}
{{#subobject:
|Title=Research Seminar Logic and AI
|Forschungsgruppe=Computational Logic
|Lecturer=Sebastian Rudolph
|Lecturer=Thomas Feller
|Modul=INF-AQUA
|Modul=INF-PM-FOR
|Term=SS
|Year=2021
|SWSLecture=0
|SWSExercise=2
|SWSPractical=0
|@category=Vorlesung
}}
{{#subobject:
|Title=Research Seminar Logic and AI
|Forschungsgruppe=Wissensbasierte Systeme
|Lecturer=Markus Krötzsch
|Lecturer=Ali Elhalawati
|Modul=INF-AQUA
|Modul=INF-PM-FOR
|Term=WS
|Year=2021
|SWSLecture=0
|SWSExercise=2
|SWSPractical=0
|@category=Vorlesung
}}
{{#subobject:
|Title=Research Seminar Logic and AI
|Forschungsgruppe=Wissensbasierte Systeme
|Lecturer=Markus Krötzsch
|Lecturer=Ali Elhalawati
|Modul=INF-AQUA
|Modul=INF-PM-FOR
|Term=SS
|Year=2022
|SWSLecture=0
|SWSExercise=2
|SWSPractical=0
|@category=Vorlesung
}}

[[Kategorie:Event series]] [[Kategorie:Vorlesung]]

Discovering Fine-Grained Semantics in Knowledge Graph Relations/en

2022-09-30T08:47:22Z

Ali Elhalawati: Page created automatically by parser function on page Discovering Fine-Grained Semantics in Knowledge Graph Relations

Discovering Fine-Grained Semantics in Knowledge Graph Relations

2022-09-30T08:47:21Z

Ali Elhalawati: Die Seite wurde neu angelegt: „{{Veranstaltung |Titel EN=Discovering Fine-Grained Semantics in Knowledge Graph Relations |Beschreibung EN=Knowledge graphs (KGs) provide structured representa…“

{{Veranstaltung
|Titel EN=Discovering Fine-Grained Semantics in Knowledge Graph Relations
|Beschreibung EN=Knowledge graphs (KGs) provide structured representation of data in the
form of relations between different entities. The semantics of relations
between words and entities are often ambiguous, where it is common to
find polysemous relations that represent multiple semantics based on the
context. This ambiguity in relation semantics also proliferates KG
triples. While the guidance from custom-designed ontologies addresses
this issue to some extent, our analysis shows that the heterogeneity and
complexity of real-world data still results in substantial relation
polysemy within popular KGs. The correct semantic interpretation of KG
relations is necessary for many downstream applications such as entity
classification and question answering.

We present the problem of fine-grained relation discovery and a
data-driven method towards this task that leverages the vector
representations of the knowledge graph entities and relations available
from relational learning models.

We show that by performing clustering over these vectors, our method is
able to not only identify the polysemous relations in knowledge graphs,
but also discover the different semantics associated with them.
Extensive empirical evaluation shows that fine-grained relations
discovered by the proposed approach lead to substantial improvement in
the semantics in the Yago and NELL datasets, as compared to baselines.

The talk will take place in a hybrid fashion, physically in the seminar room, and online through the link:
https://bbb.tu-dresden.de/b/pio-zwt-smp-aus
|Veranstaltungsart=Vortrag
|Start=2022/10/06 11:00:00
|Ende=2022/10/06 12:00:00
|Raum=APB room 3027
|Vortragender=Nitisha Jain
|Eingeladen durch=Markus Krötzsch
|Event series=Research Seminar Logic and AI
|In News anzeigen=1
}}

An Existential Rule Framework for Computing Why-Provenance On-Demand for Datalog/en

2022-09-09T13:13:27Z

Ali Elhalawati: Page created automatically by parser function on page An Existential Rule Framework for Computing Why-Provenance On-Demand for Datalog

An Existential Rule Framework for Computing Why-Provenance On-Demand for Datalog

2022-09-09T13:13:27Z

Ali Elhalawati: Die Seite wurde neu angelegt: „{{Veranstaltung |Titel EN=An Existential Rule Framework for Computing Why-Provenance On-Demand for Datalog |Beschreibung EN=Why-provenance — explaining why a…“

{{Veranstaltung
|Titel EN=An Existential Rule Framework for Computing Why-Provenance On-Demand for Datalog
|Beschreibung EN=Why-provenance — explaining why a query result is obtained — is an essential asset for reaching the goal of Explainable AI. For instance, recursive (Datalog) queries may show unexpected derivations due to complex entanglement of database atoms inside recursive rule applications. Provenance, and why-provenance in particular, helps debugging rule sets to eventually obtain the desired set of rules. There are three kinds of approaches to computing why-provenance for Datalog in the literature: (1) the complete ones, (2) the approximate ones, and (3) the theoretical ones. What all these approaches have in common is that they aim at computing provenance for all IDB atoms, while only a few atoms might be requested to be explained. We contribute an on-demand approach: After deriving all entailed facts of a Datalog program, we allow for querying for the provenance of particular IDB atoms and the structures involved in deriving provenance are computed only then. Our framework is based on terminating existential rules, recording the different rule applications. We present two implementations of the framework, one based on the semiring solver FPsolve, the other one based Datalog(S), a recent extension of Datalog by set terms. We perform experiments on benchmark rule sets using both implementations and discuss feasibility of provenance on-demand.

The talk is based on the upcoming conference paper in RuleML+RR 2022 with Markus Krötzsch and Stephan Mennicke, the paper can be found through the following link: https://iccl.inf.tu-dresden.de/web/Inproceedings3342

The talk will take place in a hybrid fashion, physically in the seminar room, and online through the link:
https://bbb.tu-dresden.de/b/ali-zgz-l8d-52n
|Veranstaltungsart=Vortrag
|Start=2022/09/22 11:00:00
|Ende=2022/09/22 12:00:00
|Raum=APB room 3027
|Vortragender=Ali Elhalawati
|Event series=Research Seminar Logic and AI
|In News anzeigen=1
}}

Inproceedings3342

2022-09-09T09:21:46Z

Ali Elhalawati:

{{Publikation Erster Autor
|ErsterAutorVorname=Ali
|ErsterAutorNachname=Elhalawati
|FurtherAuthors=Markus Krötzsch; Stephan Mennicke
}}
{{Inproceedings
|Referiert=1
|Title=An Existential Rule Framework for Computing Why-Provenance On-Demand for Datalog
|To appear=1
|Year=2022
|Month=September
|Booktitle=Proceddings of the 6th International Joint Conference on Rules and Reasoning (RuleML+RR 2022)
|Publisher=Springer
}}
{{Publikation Details
|Abstract=Why-provenance — explaining why a query result is obtained — is an essential asset for reaching the goal of Explainable AI. For instance, recursive (Datalog) queries may show unexpected derivations due to complex entanglement of database atoms inside recursive rule applications. Provenance, and why-provenance in particular, helps debugging rule sets to eventually obtain the desired set of rules. There are three kinds of approaches to computing why-provenance for Datalog in the literature: (1) the complete ones, (2) the approximate ones, and (3) the theoretical ones. What all these approaches have in common is that they aim at computing provenance for all IDB atoms, while only a few atoms might be requested to be explained. We contribute an on-demand approach: After deriving all entailed facts of a Datalog program, we allow for querying for the provenance of particular IDB atoms and the structures involved in deriving provenance are computed only then. Our framework is based on terminating existential rules, recording the different rule applications. We present two implementations of the framework, one based on the semiring solver FPsolve, the other one based Datalog(S), a recent extension of Datalog by set terms. We perform experiments on benchmark rule sets using both implementations and discuss feasibility of provenance on-demand.
|Download=An Existential Rule Framework for Computing Why Provenance On Demand for Datalog (2).pdf
|Projekt=CPEC, KIMEDS, SECAI, ScaDS.AI, Cfaed
|Forschungsgruppe=Wissensbasierte Systeme
}}

Datei:An Existential Rule Framework for Computing Why Provenance On Demand for Datalog (2).pdf

2022-09-09T09:21:40Z

Ali Elhalawati:

Inproceedings3342/en

2022-09-09T09:18:01Z

Ali Elhalawati: Page created automatically by parser function on page Inproceedings3342

#REDIRECT [[Inproceedings3342]]

Inproceedings3342

2022-09-09T09:18:00Z

Ali Elhalawati: Die Seite wurde neu angelegt: „{{Publikation Erster Autor |ErsterAutorVorname=Ali |ErsterAutorNachname=Elhalawati |FurtherAuthors=Stephan Mennicke; Markus Krötzsch }} {{Inproceedings |Refer…“

{{Publikation Erster Autor
|ErsterAutorVorname=Ali
|ErsterAutorNachname=Elhalawati
|FurtherAuthors=Stephan Mennicke; Markus Krötzsch
}}
{{Inproceedings
|Referiert=1
|Title=An Existential Rule Framework for Computing Why-Provenance On-Demand for Datalog
|To appear=1
|Year=2022
|Month=September
|Booktitle=Proceddings of the 6th International Joint Conference on Rules and Reasoning (RuleML+RR 2022)
}}
{{Publikation Details
|Abstract=Why-provenance — explaining why a query result is obtained — is
an essential asset for reaching the goal of Explainable AI. For instance, recursive
(Datalog) queries may show unexpected derivations due to complex entangle-
ment of database atoms inside recursive rule applications. Provenance, and why-
provenance in particular, helps debugging rule sets to eventually obtain the desired
set of rules. There are three kinds of approaches to computing why-provenance
for Datalog in the literature: (1) the complete ones, (2) the approximate ones,
and (3) the theoretical ones. What all these approaches have in common is that
they aim at computing provenance for all IDB atoms, while only a few atoms
might be requested to be explained. We contribute an on-demand approach: After
deriving all entailed facts of a Datalog program, we allow for querying for the
provenance of particular IDB atoms and the structures involved in deriving prove-
nance are computed only then. Our framework is based on terminating existential
rules, recording the different rule applications. We present two implementations
of the framework, one based on the semiring solver FPsolve, the other one based
Datalog(S), a recent extension of Datalog by set terms. We perform experiments
on benchmark rule sets using both implementations and discuss feasibility of
provenance on-demand.
|Projekt=CPEC, KIMEDS, SECAI, ScaDS.AI, Cfaed
|Forschungsgruppe=Wissensbasierte Systeme
}}

Finite Base Contraction and Expansion via Models - The ALC case/en

2022-09-06T09:48:32Z

Ali Elhalawati: Page created automatically by parser function on page Finite Base Contraction and Expansion via Models - The ALC case

Finite Base Contraction and Expansion via Models - The ALC case

2022-09-06T09:48:32Z

Ali Elhalawati: Die Seite wurde neu angelegt: „{{Veranstaltung |Titel EN=Finite Base Contraction and Expansion via Models - The ALC case |Beschreibung EN=We propose a new paradigm for Belief Change in which…“

{{Veranstaltung
|Titel EN=Finite Base Contraction and Expansion via Models - The ALC case
|Beschreibung EN=We propose a new paradigm for Belief Change in which the new information is represented as sets of models, while the agent’s body of knowledge is represented as a finite set of formulae, that is, a finite base. The focus on finiteness is crucial
when we consider limited agents and reasoning algorithms. Moreover, having the input as arbitrary set of models is more general than the usual treatment of formulae as input. In this setting, we define new Belief Change operations akin to traditional expansion and contraction, and we identify the rationality postulates that emerge due to the finite representability requirement. We also analyse different logics concerning compatibility with our framework. In this talk, I will focus on the ALC description logic.

The talk will take place in a hybrid fashion, physically in the seminar room, and online through the link:

https://bbb.tu-dresden.de/b/ali-zgz-l8d-52n
|Veranstaltungsart=Vortrag
|Start=2022/09/08 11:00:00
|Ende=2022/09/08 12:00:00
|Raum=APB room 3027
|Vortragender=Ana Ozaki
|Eingeladen durch=Sebastian Rudolph
|Event series=Research Seminar Logic and AI
|In News anzeigen=1
}}

Tunas - Fishing for Diverse Answer Sets: A Multi-shot Trade up Strateg/en

2022-08-29T14:29:34Z

Ali Elhalawati: Page created automatically by parser function on page Tunas - Fishing for Diverse Answer Sets: A Multi-shot Trade up Strateg

Tunas - Fishing for Diverse Answer Sets: A Multi-shot Trade up Strateg

2022-08-29T14:29:34Z

Ali Elhalawati: Die Seite wurde neu angelegt: „{{Veranstaltung |Titel EN=Tunas - Fishing for Diverse Answer Sets: A Multi-shot Trade up Strateg |Beschreibung EN=Answer set programming (ASP) solvers have adv…“

{{Veranstaltung
|Titel EN=Tunas - Fishing for Diverse Answer Sets: A Multi-shot Trade up Strateg
|Beschreibung EN=Answer set programming (ASP) solvers have advanced in
recent years, with a variety of different specialisation and overall devel-
opment. Thus, even more complex and detailed programs can be solved.
A side effect of this development are growing solution spaces and the
problem of how to find those answer sets one is interested in. One gen-
eral approach is to give an overview in form of a small number of highly
diverse answer sets. By choosing a favourite and repeating the process,
the user is able to leap through the solution space. But finding highly
diverse answer sets is computationally expensive. In this paper we intro-
duce a new approach called Tunas for Trade Up Navigation for Answer
Sets to find diverse answer sets by reworking existing solution collec-
tions. The core idea is to collect diverse answer sets one after another by
iteratively solving and updating the program. Once no more answer sets
can be added to the collection, the program is allowed to trade answer
sets from the collection for different answer sets, as long as the collection
grows and stays diverse. Elaboration of the approach is possible in three
variations, which we implemented and compared to established methods
in an empirical evaluation. The evaluation shows that the Tunas app-
roach is competitive with existing methods, and that efficiency of the
approach is highly connected to the underlying logic program.

The talk will take place in a hybrid fashion, physically in the seminar room, and online through the link:

https://bbb.tu-dresden.de/b/ali-zgz-l8d-52n
|Veranstaltungsart=Vortrag
|Start=2022/09/01 11:30:00
|Ende=2022/09/01 12:00:00
|Raum=APB room 3027
|Vortragender=Elisa Böhl
|Event series=Research Seminar Logic and AI
|In News anzeigen=1
}}

2022-06-30T12:25:22Z

Ali Elhalawati: Die Seite wurde neu angelegt: „{{Veranstaltung |Titel EN=A Glimpse into Propositional Model Counting |Beschreibung EN=Model counting (#SAT) asks to compute the number of satisfying assignmen…“

{{Veranstaltung
|Titel EN=A Glimpse into Propositional Model Counting
|Beschreibung EN=Model counting (#SAT) asks to compute the number of satisfying assignments for a propositional formula. The decision version (SAT) received widespread interest in computational complexity, formed many applications in modern combinatorial problem solving, and can be solved effectively for millions of variables on structured instances. #SAT is much harder than SAT and requires more elaborate solving techniques. In this talk, we revisit the problem, its complexity, and explain its connection to quantitative AI. We briefly overview solving techniques and illustrate a parameterized algorithm and implementation to tackle the problem. While purely parameterized approaches from theory often suffer practical limitations, we elaborate that a parameterized algorithm can be successful when combining it with modern hardware that takes advantage of parallelism.
|Veranstaltungsart=Vortrag
|Start=2022/07/07 11:00:00
|Ende=2022/07/07 12:00:00
|Raum=Online
|Vortragender=Johannes K. Fichte
|Eingeladen durch=Dominik Rusovac
|Event series=Research Seminar Logic and AI
|In News anzeigen=1
}}

Efficient Emulation of Datalog(S) in the Existential Rules Engine VLog/en

2022-06-27T22:33:38Z

Ali Elhalawati: Page created automatically by parser function on page Efficient Emulation of Datalog(S) in the Existential Rules Engine VLog

Efficient Emulation of Datalog(S) in the Existential Rules Engine VLog

2022-06-27T22:33:38Z

Ali Elhalawati: Die Seite wurde neu angelegt: „{{Veranstaltung |Titel EN=Efficient Emulation of Datalog(S) in the Existential Rules Engine VLog |Beschreibung EN=We propose a translation of the expressive lo…“

{{Veranstaltung
|Titel EN=Efficient Emulation of Datalog(S) in the Existential Rules Engine VLog
|Beschreibung EN=We propose a translation of the expressive logic Datalog(S), a subset of Datalog^CV, into optimised existential rules. We look into applications of the language, in particular regarding description logic and graph theory, as well as into expressivity. We explain the translation from earlier work in detail, identify issues and propose solutions to increase performance. We explain the need for singularisation, propose partial congruence as an alternative and define criteria for when this handling is not necessary. Furthermore, we propose an alternative translation for which the standard chase universally terminates. To measure the impact of each optimisation, we conduct an empirical evaluation on real-world data.

The talk is a defence, with a hybrid attendance.

link: https://bbb.tu-dresden.de/b/ali-zgz-l8d-52n
|Veranstaltungsart=Vortrag
|Start=2022/06/30 11:00:00
|Ende=2022/06/30 12:00:00
|Raum=APB room 3027
|Vortragender=Christian Lewe
|Eingeladen durch=Markus Krötzsch
|Event series=Research Seminar Logic and AI
|In News anzeigen=1
}}

Statistics for Knowledge Graph Modelling/en

2022-06-20T11:04:46Z

Ali Elhalawati: Page created automatically by parser function on page Statistics for Knowledge Graph Modelling

Statistics for Knowledge Graph Modelling

2022-06-20T11:04:46Z

Ali Elhalawati: Die Seite wurde neu angelegt: „{{Veranstaltung |Titel EN=Statistics for Knowledge Graph Modelling |Beschreibung EN=In this talk I will provide a high level overview of how statistical approa…“

{{Veranstaltung
|Titel EN=Statistics for Knowledge Graph Modelling
|Beschreibung EN=In this talk I will provide a high level overview of how statistical approaches can be used to model knowledge graphs. The talk is intended for everyone, regardless of statistics background, and will cover the basic vocabulary and concepts before moving on to a high level view of inference and how it can be used to approximate joint distributions. Finally, I will tie these ideas with my current work.

The talk will take place in a hybrid fashion, physical attendance is in our seminar room.
Link: https://bbb.tu-dresden.de/b/ali-zgz-l8d-52n
|Veranstaltungsart=Vortrag
|Start=2022/06/23 11:00:00
|Ende=2022/06/23 12:00:00
|Raum=APB room 3027
|Vortragender=Marcin Pietrasik
|Event series=Research Seminar Logic and AI
|In News anzeigen=1
}}

Satisfiability Testing -- Recent Developments and Open Problems

2022-06-17T08:11:20Z

Ali Elhalawati:

{{Veranstaltung
|Titel EN=Satisfiability Testing -- Recent Developments and Open Problems
|Beschreibung EN=With the introduction of clause learning in 1996,
satisfiability testing (SAT) solvers have become effective to replace
other state-of-the-art algorithm in applications. Shortly after,
efficient data structures and heuristics have been introduced, boosting
the performance of these solvers further. Recent developments shift SAT
solvers towards anytime algorithms, to attempt to find a good average
solving time for a wide range of input problems. Consequently, SAT
solvers have become the core tool to solve many industrial applications.
However, there are still limitations, and open research questions that
would allow to make SAT solving even more effective. This presentation
will shed some light on various past and recent developments, as well as
point out current known limitations as well as open research questions.

The talk is Hybird, with a 20 person maximum capacity in the seminar room APB 3027

Online link: https://bbb.tu-dresden.de/b/ali-zgz-l8d-52n
|Veranstaltungsart=Vortrag
|Start=2022/06/16 11:00:00
|Ende=2022/06/16 12:00:00
|Raum=APB room 3027
|Vortragender=Norbert Manthey
|Eingeladen durch=Markus Krötzsch
|PDF=SAT
|Event series=Research Seminar Logic and AI
|In News anzeigen=1
}}

Datei:2022-06-TUD-SAT-nmanthey.pdf

2022-06-17T08:04:42Z

Ali Elhalawati: Presentation

== Beschreibung ==
Presentation