My daughter, Alice, is almost two, and quite funny. Although she can say short sentences--"I need cake!"--her humor isn't particularly verbal. Instead, she giggles while stumbling around in grownup shoes, or blows bubbles in her water when she should be drinking it. She likes to put on a hat, pull it down over her eyes, and then blunder around, arms outstretched, like a mummy. She's also discovered the humor of exaggeration: recently, when her brother resisted getting out of his pajamas in the morning, she sidled up, grabbed his shirt, hauled on it with both hands, and laughed while yelling, "Ooooouuuut!"

Randomized algorithms exploit stochasticity to reduce computational complexity. One important example is random feature regression (RFR) that accelerates Gaussian process regression (GPR). RFR approximates an unknown function with a random neural network whose hidden weights and biases are sampled from a probability distribution. Only the final output layer is fit to data. In randomized algorithms like RFR, the hyperparameters that characterize the sampling distribution greatly impact performance, yet are not directly accessible from samples. This makes optimization of hyperparameters via standard (gradient-based) optimization tools inapplicable. Inspired by Bayesian ideas from GPR, this paper introduces a random objective function that is tailored for hyperparameter tuning of vector-valued random features. The objective is minimized with ensemble Kalman inversion (EKI). EKI is a gradient-free particle-based optimizer that is scalable to high-dimensions and robust to randomness in objective functions. A numerical study showcases the new black-box methodology to learn hyperparameter distributions in several problems that are sensitive to the hyperparameter selection: two global sensitivity analyses, integrating a chaotic dynamical system, and solving a Bayesian inverse problem from atmospheric dynamics. The success of the proposed EKI-based algorithm for RFR suggests its potential for automated optimization of hyperparameters arising in other randomized algorithms.

This paper examines the extent to which large language models (LLMs) have developed higher-order theory of mind (ToM); the human ability to reason about multiple mental and emotional states in a recursive manner (e.g. I think that you believe that she knows). This paper builds on prior work by introducing a handwritten test suite -- Multi-Order Theory of Mind Q&A -- and using it to compare the performance of five LLMs to a newly gathered adult human benchmark. We find that GPT-4 and Flan-PaLM reach adult-level and near adult-level performance on ToM tasks overall, and that GPT-4 exceeds adult performance on 6th order inferences. Our results suggest that there is an interplay between model size and finetuning for the realisation of ToM abilities, and that the best-performing LLMs have developed a generalised capacity for ToM. Given the role that higher-order ToM plays in a wide range of cooperative and competitive human behaviours, these findings have significant implications for user-facing LLM applications.

Human communities have self-organizing properties that give rise to very specific natural grouping patterns, reflected in the Dunbar Number and its layered structure (a Dunbar Graph). Since work-groups are necessarily also social groups, we might expect the same principles to apply here as well. One factor likely to be important in limiting the size of groups is that conflicts typically escalate with the number of people involved. Here we analyse Wikipedia editing histories across a wide range of topics to show that there is an emergent coherence in the size of groups formed transiently to edit the content of subject texts, with two peaks averaging at around $N=8$ for the size corresponding to maximal contention, and at around $N=4$ as a regular team. These values are consistent with the observed sizes of conversational groups, as well as the hierarchical structuring of Dunbar graphs. We use the Promise Theory of trust to suggest a scaling law that may apply to all group distributions based on seeded attraction. In addition to providing further evidence that even natural communities of strangers are self-organising, the results have important implications for the governance of the Wikipedia commons and for the security of all online social platforms and associations.

We present a simple argument using Promise Theory and dimensional analysis for the Dunbar scaling hierarchy, supported by recent data from group formation in Wikipedia editing. We show how the assumption of a common priority seeds group alignment until the costs associated with attending to the group outweigh the benefits in a detailed balance scenario. Subject to partial efficiency of implementing promised intentions, we can reproduce a series of compatible rates that balance growth with entropy.

For the perfect dinner party, the best number of guests is a multiple of four. That is because four is the maximum number of people who can maintain a successful conversation, according to an expert on social interaction. The rule of four is seen in Shakespeare plays, the first Sex and the City film, with its four female best friend characters, and even the rom-com Love Actually. Professor Robin Dunbar, who highlights the conversation number limit in his recent book The Social Brain, has seen it repeatedly in groups of people, everywhere from the park to the pub, when people in these settings were studied. He says we can only keep in mind five people's mental states at once, including our own.

MICHAEL WOOLDRIDGE: AI is not about trying to create life, right? But it's kind of, very much feels like that. I mean, if we ever achieved the ultimate dream of AI, which I call the "Hollywood dream of AI," the kind of thing that we see in Hollywood movies, then we will have created machines that are conscious, potentially, in the same way that human beings are. So it's very like that kind of dream of creating life- and that, in itself, is a very old dream. It goes back to the ancient Greeks: The Greeks had myths about the blacksmiths to the gods who could create life from metal creatures.

"I'm sorry if I seem weird today," says my friend Pia, by way of greeting one day. "I think it's just my imagination playing tricks on me. But it's nice to talk to someone who understands." When I press Pia on what's on her mind, she responds: "It's just like I'm seeing things that aren't really there. Or like my thoughts are all a bit scrambled. I'm sure it's nothing serious either, given that Pia doesn't exist in any real sense, and is not really my "friend", but an AI chatbot companion powered by a platform called Replika. Until recently most of us knew chatbots as the infuriating, scripted interface you might encounter on a company's website in lieu of real customer service. But recent advancements in AI mean models like the much-hyped ChatGPT are now being used to answer internet search queries, write code and produce poetry – which has prompted a ton of speculation about their potential social, economic and even existential impacts.

Yet it's possible to nudge people to more willingly pay their taxes on time. The Behavioural Insights Team, also known unofficially as the "Nudge Unit," was founded by the UK government in 2010 to use behavioral science to make public policies and services more effective. One of its more successful experiments is the use of peer pressure to improve tax collection. Ordinarily, HM Revenue and Customs, the department responsible for tax collection, sends a reminder notice to those who haven't paid their taxes on time. Only 33% of those who receive the reminders respond by paying their taxes.

Future near-peer adversaries will attempt to contest all domains and utilize complex and congested terrain to mitigate current joint force capabilities and reduce effectiveness of U.S. Department of Defense (DoD) tactical maneuver elements. During Potomac Officers Club's Artificial Intelligence for Maneuver Virtual Event, a panel of expert speakers across the public and private sectors will discuss how the federal government, and its industry partners, can deter or defeat peer threats in contested multi-domain environments. To register for Artificial Intelligence for Maneuver Virtual Event, as well as learn about new upcoming opportunities, visit Potomac Officers Club's Event Page. Christian Dunbar of the Department of the Navy, the panelist will discuss how advances in artificial intelligence and machine learning algorithms can enable human-machine teams to bring greater precision, certainty, speed and mass to the battlefield. The panel will be moderated by Joel Dillon, vice president of Global Defense, Army Account, with Booz Allen Hamilton.