Search Results for author:
Found 11 papers, 6 papers with code
Social Choice for AI Alignment: Dealing with Diverse Human Feedback
Foundation models such as GPT-4 are fine-tuned to avoid unsafe or otherwise problematic behavior, so that, for example, they refuse to comply with requests for help with committing crimes or with producing racist text.
Conditional and Modal Reasoning in Large Language Models
Assessing LLMs on these inference patterns is thus highly relevant to the question of how much the reasoning abilities of LLMs match those of humans.
Learning to Manipulate under Limited Information
By classic results in social choice theory, any reasonable preferential voting method sometimes gives individuals an incentive to report an insincere preference.
An impossibility theorem concerning positive involvement in voting
In social choice theory with ordinal preferences, a voting method satisfies the axiom of positive involvement if adding to a preference profile a voter who ranks an alternative uniquely first cannot cause that alternative to go from winning to losing.
An extension of May's Theorem to three alternatives: axiomatizing Minimax voting
In particular, we add two axioms stating that the voting method should mitigate spoiler effects and avoid the so-called strong no show paradox.
An Axiomatic Characterization of Split Cycle
In this paper, we go further and show that Split Cycle is the only rule satisfying the axioms of Holliday and Pacuit together with two additional axioms: Coherent Defeat and Positive Involvement in Defeat.
Impossibility theorems involving weakenings of expansion consistency and resoluteness in voting
A fundamental principle of individual rational choice is Sen's $\gamma$ axiom, also known as expansion consistency, stating that any alternative chosen from each of two menus must be chosen from the union of the menus.
Escaping Arrow's Theorem: The Advantage-Standard Model
While previous literature shows that only weakening IIA to weak IIA or only weakening negative transitivity of $P$ to acyclicity still leads to impossibility theorems, we show that jointly weakening IIA to AS rationalizability and weakening negative transitivity of $P$ leads to no such impossibility theorems.
Preferential Structures for Comparative Probabilistic Reasoning
Qualitative and quantitative approaches to reasoning about uncertainty can lead to different logical systems for formalizing such reasoning, even when the language for expressing uncertainty is the same.
