arXiv subjects

Caleb Probine

Publications and source records attributed to Caleb Probine.

6 recordsLinked to original sources

arXiv2026

Characterizing Opinion Evolution of Networked LLMs

Large language models (LLMs) increasingly interact with one another in multi-agent systems, from simulations of human discourse to influence operations and fully LLM-driven social platforms. These interactions give rise to new regimes of opinion propagation that are not yet well understood. We investigate whether classical opinion dynamics models, which have long been used to explain how interactions shape collective beliefs in human societies, can capture the behavior of LLM networks. We find that, while naive averaging-style models fail to track LLMs' opinion dynamics, simple modifications yield substantial gains in modeling fidelity. In particular, bias, an innate opinion toward which agents regress, emerges as a significant driver of LLM opinion dynamics, with its inclusion reducing cumulative estimated mean opinion error by up to 88%. We additionally find that these conclusions generalize across model families, discussion topics, and networks.

cs.MA↗

arXiv2026

When Should Agents Coordinate in Differentiable Sequential Decision Problems?

Multi-robot teams must coordinate to operate effectively. When a team operates in an uncoordinated manner, and agents choose actions that are only individually optimal, the team's outcome can suffer. However, in many domains, coordination requires costly communication. We explore the value of coordination in a broad class of differentiable motion-planning problems. In particular, we model coordinated behavior as a spectrum: at one extreme, agents jointly optimize a common team objective, and at the other, agents make unilaterally optimal decisions given their individual decision variables, i.e., they operate at Nash equilibria. We then demonstrate that reasoning about coordination in differentiable motion-planning problems reduces to reasoning about the second-order properties of agents' objectives, and we provide algorithms that use this second-order reasoning to determine at which times a team of agents should coordinate.

cs.MA↗

arXiv2025

Multi-Environment POMDPs: Discrete Model Uncertainty Under Partial Observability

Multi-environment POMDPs (ME-POMDPs) extend standard POMDPs with discrete model uncertainty. ME-POMDPs represent a finite set of POMDPs that share the same state, action, and observation spaces, but may arbitrarily vary in their transition, observation, and reward models. Such models arise, for instance, when multiple domain experts disagree on how to model a problem. The goal is to find a single policy that is robust against any choice of POMDP within the set, i.e., a policy that maximizes the worst-case reward across all POMDPs. We generalize and expand on existing work in the following way. First, we show that ME-POMDPs can be generalized to POMDPs with sets of initial beliefs, which we call adversarial-belief POMDPs (AB-POMDPs). Second, we show that any arbitrary ME-POMDP can be reduced to a ME-POMDP that only varies in its transition and reward functions or only in its observation and reward functions, while preserving (optimal) policies. We then devise exact and approximate (point-based) algorithms to compute robust policies for AB-POMDPs, and thus ME-POMDPs. We demonstrate that we can compute policies for standard POMDP benchmarks extended to the multi-environment setting.

cs.AI↗

arXiv2025

Designing Inferable Signaling Schemes for Bayesian Persuasion

In Bayesian persuasion, an informed sender, who observes a state, commits to a randomized signaling scheme that guides a self-interested receiver's actions. Classical models assume the receiver knows the commitment. We, instead, study the setting where the receiver infers the scheme from repeated interactions. We bound the sender's performance loss relative to the known-commitment case by a term that grows with the signal space size and shrinks as the receiver's optimal actions become more distinct. We then lower bound the samples required for the sender to approximately achieve their known-commitment performance in the inference setting. We show that the sender requires more samples in persuasion compared to the leader in a Stackelberg game, which includes commitment but lacks signaling. Motivated by these bounds, we propose two methods for designing inferable signaling schemes, one being stochastic gradient descent (SGD) on the sender's inference-setting utility, and the other being optimization with a boundedly-rational receiver model. SGD performs best in low-interaction regimes, but modeling the receiver as boundedly-rational and tuning the rationality constant still provides a flexible method for designing inferable schemes. Finally, we apply SGD to a safety alert example and show it to find schemes that have fewer signals and make citizens' optimal actions more distinct compared to the known-commitment case.

cs.GT↗

arXiv2025

Privacy-preserving Nash Equilibrium Synthesis with Partially Ordered Temporal Objectives

Nash equilibrium is a central solution concept for reasoning about self-interested agents. We address the problem of synthesizing Nash equilibria in two-player deterministic games on graphs, where players have private, partially-ordered preferences over temporal goals. Unlike prior work, which assumes preferences are common knowledge, we develop a communication protocol for equilibrium synthesis in settings where players' preferences are private information. In the protocol, players communicate to synthesize equilibria by exchanging information about when they can force desirable outcomes. We incorporate privacy by ensuring the protocol stops before enough information is revealed to expose a player's preferences. We prove completeness by showing that, when no player halts communication, the protocol either returns an equilibrium or certifies that none exists. We then prove privacy by showing that, with stopping, the messages a player sends are always consistent with multiple possible preferences and thus do not reveal some given secret regarding a player's true preference ordering. Experiments demonstrate that we can synthesize non-trivial equilibria while preserving privacy of preferences, highlighting the protocol's potential for applications in strategy synthesis with constrained information sharing.

cs.GT↗

arXiv2024

A Decentralized Shotgun Approach for Team Deception

Deception is helpful for agents masking their intentions from an observer. We consider a team of agents deceiving their supervisor. The supervisor defines nominal behavior for the agents via reference policies, but the agents share an alternate task that they can only achieve by deviating from these references. As such, the agents use deceptive policies to complete the task while ensuring that their behaviors remain plausible to the supervisor. We propose a setting with centralized deceptive policy synthesis and decentralized execution. We model each agent with a Markov decision process and constrain the agents' deceptive policies so that, with high probability, at least one agent achieves the task. We then provide an algorithm to synthesize deceptive policies that ensure the deviations of all agents are small by minimizing the worst Kullback-Leibler divergence between any agent's deceptive and reference policies. Thanks to decentralization, this algorithm scales linearly with the number of agents and also facilitates the efficient synthesis of reference policies. We then explore a more general version of the deceptive policy synthesis problem. In particular, we consider a supervisor who selects a subset of agents to eliminate based on the agents' behaviors. We give algorithms to synthesize deceptive policies so that, after the supervisor eliminates some agents, the remaining agents complete the task with high probability. We demonstrate the developed methods in a package delivery example.

math.OC↗