Knowing others' intentions and taking others' perspectives are two core components of human intelligence typically considered as instantiations of theory of mind. Infiltrating machines with these abilities is an important step towards building human-level artificial intelligence. We here investigate intentionality understanding and perspective-taking in Vision Language Models and, for the purpose, we have created IntentBench and PerspectBench datasets, which contain over 400 cognitive experiments grounded in real-world scenarios and classic cognitive tasks. Surprisingly, we find that VLMs achieve high performance in intentionality understanding but lower performance in perspective-taking using our two datasets. This challenges the common belief in the cognitive science literature that perspective-taking at the corresponding modality is necessary for intentionality understanding.

Game theory provides a mathematical way to study the interaction between multiple decision makers. However, classical game-theoretic analysis is limited in scalability due to the large number of strategies, precluding direct application to more complex scenarios. This survey provides a comprehensive overview of a framework for large games, known as Policy Space Response Oracles (PSRO), which holds promise to improve scalability by focusing attention on sufficient subsets of strategies. We first motivate PSRO and provide historical context. We then focus on the strategy exploration problem for PSRO: the challenge of assembling effective subsets of strategies that still represent the original game well with minimum computational cost. We survey current research directions for enhancing the efficiency of PSRO, and explore the applications of PSRO across various domains. We conclude by discussing open questions and future research.

Policy-Space Response Oracles (PSRO) as a general algorithmic framework has achieved state-of-the-art performance in learning equilibrium policies of two-player zero-sum games. However, the hand-crafted hyperparameter value selection in most of the existing works requires extensive domain knowledge, forming the main barrier to applying PSRO to different games. In this work, we make the first attempt to investigate the possibility of self-adaptively determining the optimal hyperparameter values in the PSRO framework. Our contributions are three-fold: (1) Using several hyperparameters, we propose a parametric PSRO that unifies the gradient descent ascent (GDA) and different PSRO variants. (2) We propose the self-adaptive PSRO (SPSRO) by casting the hyperparameter value selection of the parametric PSRO as a hyperparameter optimization (HPO) problem where our objective is to learn an HPO policy that can self-adaptively determine the optimal hyperparameter values during the running of the parametric PSRO. (3) To overcome the poor performance of online HPO methods, we propose a novel offline HPO approach to optimize the HPO policy based on the Transformer architecture. Experiments on various two-player zero-sum games demonstrate the superiority of SPSRO over different baselines.

In iterative approaches to empirical game-theoretic analysis (EGTA), the strategy space is expanded incrementally based on analysis of intermediate game models. A common approach to strategy exploration, represented by the double oracle algorithm, is to add strategies that best-respond to a current equilibrium. This approach may suffer from overfitting and other limitations, leading the developers of the policy-space response oracle (PSRO) framework for iterative EGTA to generalize the target of best response, employing what they term meta-strategy solvers (MSSs). Noting that many MSSs can be viewed as perturbed or approximated versions of Nash equilibrium, we adopt an explicit regularization perspective to the specification and analysis of MSSs. We propose a novel MSS called regularized replicator dynamics (RRD), which simply truncates the process based on a regret criterion. We show that RRD is more adaptive than existing MSSs and outperforms them in various games. We extend our study to three-player games, for which the payoff matrix is cubic in the number of strategies and so exhaustively evaluating profiles may not be feasible. We propose a profile search method that can identify solutions from incomplete models, and combine this with iterative model construction using a regularized MSS. Finally, and most importantly, we reveal that the regret of best response targets has a tremendous influence on the performance of strategy exploration through experiments, which provides an explanation for the effectiveness of regularization in PSRO.

Qualitative probabilistic networks (QPNs) combine the conditional independence assumptions of Bayesian networks with the qualitative properties of positive and negative dependence. They formalise various intuitive properties of positive dependence to allow inferences over a large network of variables. However, we will demonstrate in this paper that, due to an incorrect symmetry property, many inferences obtained in non-binary QPNs are not mathematically true. We will provide examples of such incorrect inferences and briefly discuss possible resolutions.

On a recent Sunday at the Queen Anne Lutheran Church basement, parishioners sat transfixed as the Rev. Dr. Ted Peters discussed an unusual topic for an afternoon assembly: "Can technology enhance the image of God?" Peters' discussion focused on a relatively new philosophical movement. Its followers believe humans will transcend their physical and mental limitations with wearable and implantable devices. The movement, called transhumanism, claims that in the future, humans will be smarter and stronger and may even overcome aging and death through developments in fields such as biotechnology and artificial intelligence (AI). "What does it mean to be truly human?" Peters asked in a voice that boomed throughout the church basement, in a city that boasts one of the world's largest tech hubs.

We exploit player symmetry to formulate the representation of large normal-form games as a regression task. This formulation allows arbitrary regression methods to be employed in in estimating utility functions from a small subset of the game's outcomes. We demonstrate the applicability both neural networks and Gaussian process regression, but focus on the latter. Once utility functions are learned, computing Nash equilibria requires estimating expected payoffs of pure-strategy deviations from mixed-strategy profiles. Computing these expectations exactly requires an infeasible sum over the full payoff matrix, so we propose and test several approximation methods. Three of these are simple and generic, applicable to any regression method and games with any number of player roles. However, the best performance is achieved by a continuous integral that approximates the summation, which we formulate for the specific case of fully-symmetric games learned by Gaussian process regression with a radial basis function kernel. We demonstrate experimentally that the combination of learned utility functions and expected payoff estimation allows us to efficiently identify approximate equilibria of large games using sparse payoff data.

This article provides an overview of the field of qualitative decision theory: its motivating tasks and issues, its antecedents, and its prospects. Qualitative decision theory studies qualitative approaches to problems of decision making and their sound and effective reconciliation and integration with quantitative approaches. Although it inherits from a long tradition, the field offers a new focus on a number of important unanswered questions of common concern to AI, economics, law, psychology, and management. As developed by philosophers, economists, and mathematicians over some 300 years, these disciplines have developed many powerful ideas and techniques, which exert major influences over virtually all the biological, cognitive, and social sciences. Their uses range from providing mathematical foundations for microeconomics to daily application in a range of fields of practice, including finance, public policy, medicine, and now even automated device diagnosis.

We investigate the effects of market making on market performance, focusing on allocative efficiency as well as gains from trade accrued by background traders. We employ empirical simulation-based methods to evaluate heuristic strategies for market makers as well as background investors in a variety of complex trading environments. Our market model incorporates private and common valuation elements, with dynamic fundamental value and asymmetric information. In this context, we compare the surplus achieved by background traders in strategic equilibrium, with and without a market maker. Our findings indicate that the presence of the market maker strongly tends to increase total welfare across various environments. Market-maker profit may or may not exceed the welfare gain, thus the effect on background-investor surplus is ambiguous. We find that market making tends to benefit investors in relatively thin markets, and situations where background traders are impatient, due to limited trading opportunities. The presence of additional market makers increases these benefits, as competition drives the market makers to provide liquidity at lower price spreads. A thorough sensitivity analysis indicates that these results are robust to reasonable changes in model parameters.

Artificial intelligence boosters predict a brave new world of flying cars and cancer cures. Detractors worry about a future where humans are enslaved to an evil race of robot overlords. Veteran AI scientist Eric Horvitz and Doomsday Clock guru Lawrence Krauss, seeking a middle ground, gathered a group of experts in the Arizona desert to discuss the worst that could possibly happen - and how to stop it. Their workshop took place last weekend at Arizona State University (ASU) with funding from Tesla co-founder Elon Musk and Skype co-founder Jaan Tallinn. Officially dubbed "Envisioning and Addressing Adverse AI Outcomes", it was a kind of AI doomsday games that organised some 40 scientists, cyber-security experts and policy wonks into groups of attackers - the red team - and defenders - blue team - playing out AI-gone-very-wrong scenarios, ranging from stock-market manipulation to global warfare.