SEPTEMBER 19th, 2016, LogiCIC-LIRa Seminar

Speaker: Branden Fitelson
Date and Time: Monday, September 19th 2016, 11:00 – 12:30.
Venue: ILLC Seminar Room F1.15, Science Park 107.
Title: Two New(ish) Triviality Results for Indicative Conditionals.
Abstract. I will do two things in this talk: (1) present an axiomatic generalization of Gibbard’s (logical) triviality result for indicative conditionals, and (2) present an algebraic strengthening of Lewis’s (probabilistic) triviality result for indicative conditionals. Both results start from a very weak background theory (either logical or probabilistic) of the indicative conditional, and (relative to these weak backgrounds) both results will rely only on the so-called Import-Export Law. So, these results can be viewed as (general, and strong) “odd consequences” of Import-Export.

---

LIRa/LogiCIC Lecture Series with Eric Pacuit
June 1st - June 15th 2016

Eric Pacuit (University of Maryland) will give a LIRa/LogiCIC Lecture Series between June 1st and June 15th 2016 on Neighborhood Semantics for Modal Logic. Everyone is welcome to attend the lectures. Students in the Master of Logic can subscribe to the lecture series and take part in the exercise sessions for a June Project. For more details an the full schedule, see the MSc Logic website.

 

---

 

MAY 13th, 2016, LogiCIC-LIRa Seminar

Speaker: Guido Bacciagaluppi
Date and Time: Friday, May 13th 2016, 13:00 – 14:30.
Venue: ILLC Seminar Room F1.15, Science Park 107.
Title: Von Neumann’s no-Hidden-Variables Theorem (and Hermann’s Critique).

Abstract. There has been some recent debate about what von Neumann intended with his famous no-hidden-variables theorem (from his 1932 book), and on whether Bell’s now equally famous criticism is really fair to von Neumann. I revisit von Neumann’s result, which was originally published in 1927 in a paper trying to give what one would now call a “reconstruction” of quantum mechanics, and suggest a more nuanced reading. I pay special attention also to the critique of von Neumann provided by Grete Hermann in her 1935 essay on the foundations of quantum mechanics (and in an unpublished manuscript from 1933).

 

---

APRIL 29th, 2016, LogiCIC-LIRa Seminar

Speaker: Rohit Parikh
Date and Time: Friday, April 22nd 2016, 13:00 – 14:30.
Venue: ILLC Seminar Room F1.15, Science Park 107.
Title: Influencing behavior by influencing beliefs.

Abstract. Agents act on the basis of their preferences and their beliefs. These include beliefs both about the world and about the actions of other agents.

By influencing the beliefs of agents we can also influence what choices they make. The influence need not involve deception but may simply involve withholding or offering certain truths.
We show that for a finite set of agents who know nothing about some proposition, an arbitrary state of information can be achieved by sending an n-tuple of signals, one to each agent. We consider how this will influence the actions of cautious (risk averse) or aggressive (risk loving) agents.
We next consider the case of a candidate campaigning for election who seeks the approval of voters. What sorts of things can she say to increase their approval? And is it possible to increase approval while, at the same time losing votes? We show that it is always possible to increase net approval but that this can result in losing votes.

 

---

APRIL 22nd, 2016, LogiCIC-LIRa Seminar

Speaker: Rohit Parikh
Date and Time: Friday, April 22nd 2016, 16:00 – 17:30.
Venue: ILLC Seminar Room F1.15, Science Park 107.
Title: An Epistemic Generalization of Rationalizability.

Abstract. Savage showed us how to infer an agent’s subjective probabilities and utilities from the bets which the agent accepts or rejects. But in a game theoretic situation an agent’s beliefs are not just about the world but also about the probable actions of other agents which will depend on their, beliefs and utilities. Moreover, it is unlikely that agents know the precise subjective probabilities or cardinal utilities of other agents. An agent is more likely to know something about the preferences of other agents and something about their beliefs. In view of this, the agent is unlikely to to have a precise best action which we can predict, but is more likely to have a set of “not so good” actions which the agent will not perform.

Ann may know that Bob prefers chocolate to vanilla to strawberry. She is unlikely to know whether Bob will prefer vanilla ice cream or a 50-50 chance of chocolate and strawberry. So Ann’s actions and her beliefs need to be understood in the presence of such partial ignorance. We propose a theory which will let us decide when Ann is being irrational, based on our partial knowledge of her beliefs and preferences, and assuming that Ann is, rational, how to infer her beliefs and preferences from her actions.
Our principal tool is a generalization of rational behavior in the context of ordinal utilities and partial knowledge of the game which the agents are playing.

 

---

APRIL 22nd, 2016, LogiCIC-LIRa Seminar

Speaker: Jon Williamson
Date and Time: Friday, April 22nd 2016, 13:00 – 14:30.
Venue: ILLC Seminar Room F1.15, Science Park 107.
Title: Inductive Logic for Automated Decision Making.

Abstract. According to Bayesian decision theory, one’s acts should maximise expected utility. To calculate expected utility one needs not only the utility of each act in each possible scenario but also the probabilities of the various scenarios. It is the job of an inductive logic to determine these probabilities, given the evidence to hand. The most natural inductive logic, classical inductive logic, attributable to Wittgenstein, was dismissed by Carnap due to its apparent inability to capture the phenomenon of learning from experience. I argue that Carnap was too hasty to dismiss this logic: classical inductive logic can be rehabilitated, and the problem of learning from experience overcome, by appealing to the principles of objective Bayesianism. I then discuss the practical question of how to calculate the required probabilities and show that the machinery of probabilistic networks can be fruitfully applied here. This culminates in an objective Bayesian decision theory that has a realistic prospect of automation.

 

---

 

APRIL 15th, 2016, LogiCIC-LIRa Seminar

Speaker: Jeffrey M. Keisler
Date and Time: Friday, April 15th 2016, 13:00 – 14:30.
Venue: ILLC Seminar Room F1.15, Science Park 107.
Title: Observing, reporting and deciding in networks of agents.

Abstract. We define an observation network as a network of agents who each have a vocabulary, a knowledge base in first-order logic, and a set of potential observations. Agents can make inferences and report them to other agents to whom they are linked, but communication is restricted to the common vocabulary of the reporter and recipient. Agents add to their knowledge bases by making observations, receiving reports from other agents, and making inferences, and share useful statements with other agents who have different vocabularies.

A well-known result from mathematical logic, the Craig Interpolation Theorem, can be interpreted in terms of knowledge sharing between pairs of agents. It is a starting point for our formal results. Certain agents called deciders are faced with alternative statements, and must decide which one is true. We consider various questions about when the network is rich enough to ensure that decisions can be made successfully if the knowledge to do so exists somewhere in the network. We obtain some easily checked completeness conditions that ensure that this is so.
The focus of the presentation will be on explanation, interpretation and application of the framework and results, rather than on their derivation. Observation networks are general enough to model many situations. We conclude by considering the example of junction tree algorithms for solving Bayes nets.

Joint work with H. Jerome Keisler, University of Wisconsin.

Biography. Jeffrey Keisler is currently Visiting Professor and Fulbright Distinguished Chair in the Mathematics and Systems Analysis Department at Aalto University in Helsinki. A Professor of Management Information Systems at the University of Massachusetts Boston, he has two books and over fifty journal articles and book chapters in decision analysis and related fields. Professor Keisler is past-President of the INFORMS Decision Analysis Society and has received its Publication Award. His PhD in Decision Sciences is from Harvard University. He will present joint work with H. Jerome Keisler who has been a leader of mathematical logic for over fifty years.

 

---

 

APRIL 8th, 2016, LogiCIC-LIRa Seminar

Speaker: Elias Tsakas
Date and Time: Friday, April 8th 2016, 13:00 – 14:30.
Venue: ILLC Seminar Room F1.15, Science Park 107.
Title: Reasonable Doubt Revisited.

Abstract. According to the common definition, the standard of reasonable doubt is a threshold such that, the defendant is convicted if and only if the probability that the juror attaches to the defendant being guilty is above this threshold. In this paper we prove that generically such a threshold exists if and only if the juror reasons only about the defendant’s guilt and nothing else. We discuss the implications of this result, and subsequently we propose a weakening of the aforementioned definition, by substituting the standard of reasonable doubt with a pair of standards, an upper and a lower one, thus obtaining a sufficient condition for conviction and a sufficient condition for acquittal. Finally, we prove that the lower standard always exists, whereas the upper standard exists if and only if the juror prefers to always convict a guilty defendant, irrespective of the circumstances.

See also http://www.elias-tsakas.com/Research/Papers/ReasonableDoubt.pdf.