For over 60 years, computers have been based on digital computing principles. Data is represented as bits (1s and 0s). Boolean logic gates perform operations on these bits. Lyric has invented a new kind of logic gate circuit that uses transistors as dimmer switches instead of as on/off switches. These circuits can accept inputs and calculate outputs that are between 0 and 1, directly representing probabilities – levels of certainty. A digital processor steps through these operations serially in order to perform a function. In order to improve efficiency even further, Lyric’s processors are designed to perform many probability computations in parallel. (Lyric Press Materials)

The Monty Hall Problem, (The New York Times).

Take this question out back…, (Language Log).

Finally, the great Lenny Pickett, (circa 1973). He is the one in the bow tie.

For details and deadlines, contact:
—————————
Jean Goubault-Larrecq
goubault@lsv.ens-cachan.fr
Eric Goubault
eric.goubault@cea.fr
Olivier Bouissou
olivier.bouissou@cea.fr

Project description:

This postdoctoral position is part of the ANR Blanc “Confidence, Proof and Probabilities” project. This project aims at studying the joint use of probabilistic and deterministic semantics and analysis methods in order to improve the applicability and precision of static analysis methods on numerical programs. The project includes theoretical computer scientists, specialists of abstract interpretation based static analysis, applied mathematicians and control theoreticians as well as industrial partners. The project will contain the following steps:

* construction of good approximations of the semantics for non-deterministic and probabilistic behaviours.
* abstraction of these semantics models for the tractable static analysis of complex programs and test case generation.
* characterisation of the imprecision error due to the use of floating point numbers by the program.

The role of this postdoctoral position is to explore the theoretical and practical aspects of these points. A first task will be to define a good semantics for probabilistic behaviours in numerical programms, building on work by Goubault-Larrecq at LSV [ICALP07,PRONOBIS], in particular the notion of prevision transition systems [FOSSACS08]. Then, the candidate will explore the abstraction semantics, based on deterministic abstractions using zonotopes as in [CAV09,arxiv2008,arxiv2010]. Finally, if time permits and according to the applicants preferences, he/she may look into the notion of bisimulation distances to characterize the imprecision error.

Qualifications:
———————-
The candidate must have a PhD in computer science or applied mathematics and be specialist of
one of the following subject: probabiliy theory, static analysis, numerical programs analysis,…

Practical information:
———————-
This work will be done in very close collaboration with co-advisors from two sites (Jean Goubault-Larrecq at LSV and Eric Goubault at CEA LIST), and will be deeply associated with the project. The candidate will be located both at ENS Cachan, and at CEA Saclay nearby Paris. Salaries range from 27840 to 30000 euros per year.
———————-
[CAV2009] Khalil Ghorbal, Eric Goubault, Sylvie Putot, “The Zonotope Abstract Domain Taylor1+”, CAV 2009.

[Arxiv2010] Eric Goubault, Sylvie Putot, “A zonotopic framework for functional
abstractions”, CoRR abs/0910.1763.

[Arxiv2008] Eric Goubault, Sylvie Putot, “Perturbed affine arithmetic for invariant computation in numerical program analysis”, CoRR abs/0807.2961

[ICALP07] Jean Goubault-Larrecq, “Continuous Capacities on Continuous State Spaces”, ICALP 07, LNCS 4596.

[FOSSACS08] Jean Goubault-Larrecq, “Prevision Domains and Convex Powercones”, FOSSACS 2008, LNCS 4962.

[PRONOBIS] Jean Goubault-Larrecq, “Une introduction aux capacites, aux jeux et aux previsions”, available at http://www.lsv.ens-cachan.fr/~goubault/ProBobis/pp.pdf.

Special session on Logics for Games and Strategic Reasoning
at CLIMA XI (Computational Logic in Multi-Agent Systems)
Lisbon, Portugal, August 16-17, 2010
(colocated with ECAI”10)

http://centria.di.fct.unl.pt/events/climaXI/sessions.html

1st Call for Papers

************************************************************************
Submission deadline: May 7th.
Proceedings: LNCS/LNAI volume (available at the workshop).
Postproceedings: extended versions of selected papers will be published
in a special issue of Annals of Mathematics and Artificial Intelligence.
************************************************************************

INTRODUCTION

Strategic reasoning occurs in many multi-agent systems. This is
especially evident in game-theoretical and decision-theoretical models
of MAS, but also in more informal settings using the game metaphor (like
computer games or social network services). Mathematical logic can
contribute to this view in many ways by providing specifications,
models, and/or algorithms for game-like scenarios. We invite papers that
address how logic can contribute to our understanding, modeling and
analysis of games, but also ones that investigate how the metaphor of
games and strategies can help in constructing and using logical formalisms.

The list of subjects includes (but is not limited to) the following topics:
- Logics for reasoning about games and strategies
- Modal logics of strategic ability
- Logical foundations of game and decision theory
- Logical approaches to rationality and bounded rationality
- Solving games and verification of strategies by model checking
- Logics for mechanism design
- Game semantics for logics
- Games in verification of logical specifications

SUBMISSION

We encourage submission of high quality, original papers which have not
been accepted for publication nor are currently under review for another
journal or conference. Submissions should not exceed 16 pages in the
Springer LNCS format. For templates and instructions for authors, see
http://www.springer.de/comp/lncs/authors.html . Papers should be
submitted electronically via EasyChair:
http://www.easychair.org/conferences/?conf=clima2010 .

PROCEEDINGS AND POST-PROCEEDINGS

LNCS Proceedings: All the accepted papers (including special session
papers) will be published by Springer as a volume in Lecture Notes in
Computer Science, and will be available in time for the workshop.

AMAI Special Issue: After the workshop, authors of selected papers will
be invited to extend and re-submit their work to a special issue of
Annals of Mathematics and Artificial Intelligence.

IMPORTANT DATES

Paper submission: May 7, 2010
Author notification: June 4, 2010
Camera-ready deadline: June 16, 2010
Workshop: August 16-17, 2010

ORGANIZATION AND CONTACT

The special session is organized by Wojtek Jamroga, University of
Luxembourg. In case of questions, please do not hesitate to contact us
at wojtek.jamroga@uni.lu . General questions about CLIMA XI should be
sent to clima2010@easychair.org .

What’s the most efficient way to pack a group of identical, regular tetrahedron? Here answers have been all over the board. Aristotle thought they would pack together perfectly. That’s false; it turns out they don’t pack together very easily at all. Is their packing density greater than that of the sphere problem? Around 2006, the answer appeared to be negative. Now we know it is positive, and the race is on to find the upper limit.

Gambetta and Hertog mention this and other failed attempts to find good classifiers for profiling terrorist by way of arguing that engineers are three to four times as likely as other graduates to be affiliated with Islamic terrorist groups since 1970. Perhaps, they argue, university subjects provide a key to better classifiers.

Slate has recently picked up their thesis, but the Slate article is less concerned with the data than with Gambetta and Hertog’s speculation on root causes:

We reckon that something else is going on, something at the individual level, that is, relating to cognitive traits. According to polling data, engineering professors in the US are seven times as likely to be right-wing and religious as other academics, and similar biases apply to students. In 16 other countries we investigated, engineers seem to be no more right-wing or religious than the rest of the population, but the number of engineers combining both traits is unusually high. A lot of piecemeal evidence suggests that characteristics such as greater intolerance of ambiguity, a belief that society can be made to work like clockwork, and dislike of democratic politics which involves compromise, are more common among engineers.

I found the last line about compromise very peculiar, since engineering as a profession is essentially concerned with compromising between price and performance. I am not quarreling with the basic picture here: we could surely tell a story about Hank the vegan engineer that was just as confounding as the one told about Linda the feminist bank teller.

Rather, my point is that ‘trained engineer’ and ‘practicing engineer’ might be a worthwhile distinction in exploring engineering studies as a classifier. I’d also add that, in so far as engineering students can be a bit singleminded, increasing the number and variety of optimization problems in the curriculum that include social or economic parameters might not be a bad idea.