Autonomous systems that operate around humans will likely always rely on kill switches that stop their execution and allow them to be remote-controlled for the safety of humans or to prevent damage to the system. It is theoretically possible for an autonomous system with sufficient sensor and effector capability that learn online using reinforcement learning to discover that the kill switch deprives it of long-term reward and thus learn to disable the switch or otherwise prevent a human operator from using the switch. This is referred to as the big red button problem. We present a technique that prevents a reinforcement learning agent from learning to disable the kill switch. We introduce an interruption process in which the agent's sensors and effectors are redirected to a virtual simulation where it continues to believe it is receiving reward. We illustrate our technique in a simple grid world environment.

Motivated by Shannon's model and recent rehabilitation of self-supervised artificial intelligence having a "World Model", this paper propose an unified intelligence-communication (UIC) model for describing a single agent and any multi-agent system. Firstly, the environment is modelled as the generic communication channel between agents. Secondly, the UIC model adopts a learning-agent model for unifying several well-adopted agent architecture, e.g. rule-based agent model in complex adaptive systems, layered model for describing human-level intelligence, world-model based agent model. The model may also provide an unified approach to investigate a multi-agent system (MAS) having multiple action-perception modalities, e.g. explicitly information transfer and implicit information transfer. This treatise would be divided into three parts, and this first part provides an overview of the UIC model without introducing cumbersome mathematical analysis and optimizations. In the second part of this treatise, case studies with quantitative analysis driven by the UIC model would be provided, exemplifying the adoption of the UIC model in multi-agent system. Specifically, two representative cases would be studied, namely the analysis of a natural multi-agent system, as well as the co-design of communication, perception and action in an artificial multi-agent system. In the third part of this treatise, the paper provides further insights and future research directions motivated by the UIC model, such as unification of single intelligence and collective intelligence, a possible explanation of intelligence emergence and a dual model for agent-environment intelligence hypothesis. Notes: This paper is a Previewed Version, the extended full-version would be released after being accepted.

We construct a complexity-based morphospace to study systems-level properties of conscious & intelligent systems. The axes of this space label 3 complexity types: autonomous, cognitive & social. Given recent proposals to synthesize consciousness, a generic complexity-based conceptualization provides a useful framework for identifying defining features of conscious & synthetic systems. Based on current clinical scales of consciousness that measure cognitive awareness and wakefulness, we take a perspective on how contemporary artificially intelligent machines & synthetically engineered life forms measure on these scales. It turns out that awareness & wakefulness can be associated to computational & autonomous complexity respectively. Subsequently, building on insights from cognitive robotics, we examine the function that consciousness serves, & argue the role of consciousness as an evolutionary game-theoretic strategy. This makes the case for a third type of complexity for describing consciousness: social complexity. Having identified these complexity types, allows for a representation of both, biological & synthetic systems in a common morphospace. A consequence of this classification is a taxonomy of possible conscious machines. We identify four types of consciousness, based on embodiment: (i) biological consciousness, (ii) synthetic consciousness, (iii) group consciousness (resulting from group interactions), & (iv) simulated consciousness (embodied by virtual agents within a simulated reality). This taxonomy helps in the investigation of comparative signatures of consciousness across domains, in order to highlight design principles necessary to engineer conscious machines. This is particularly relevant in the light of recent developments at the crossroads of cognitive neuroscience, biomedical engineering, artificial intelligence & biomimetics.

Nowadays, several crowdsourcing projects exploit social choice methods for computing an aggregate ranking of alternatives given individual rankings provided by workers. Motivated by such systems, we consider a setting where each worker is asked to rank a fixed (small) number of alternatives and, then, a positional scoring rule is used to compute the aggregate ranking. Among the apparently infinite such rules, what is the best one to use? To answer this question, we assume that we have partial access to an underlying true ranking. Then, the important optimization problem to be solved is to compute the positional scoring rule whose outcome, when applied to the profile of individual rankings, is as close as possible to the part of the underlying true ranking we know. We study this fundamental problem from a theoretical viewpoint and present positive and negative complexity results and, furthermore, complement our theoretical findings with experiments on real-world and synthetic data.

This work explores a sequential decision making problem with agents having diverse expertise and mismatched beliefs. We consider an $N$-agent sequential binary hypothesis test in which each agent sequentially makes a decision based not only on a private observation, but also on previous agents' decisions. In addition, the agents have their own beliefs instead of the true prior, and have varying expertise in terms of the noise variance in the private signal. We focus on the risk of the last-acting agent, where precedent agents are selfish. Thus, we call this advisor(s)-advisee sequential decision making. We first derive the optimal decision rule by recursive belief update and conclude, counterintuitively, that beliefs deviating from the true prior could be optimal in this setting. The impact of diverse noise levels (which means diverse expertise levels) in the two-agent case is also considered and the analytical properties of the optimal belief curves are given. These curves, for certain cases, resemble probability weighting functions from cumulative prospect theory, and so we also discuss the choice of Prelec weighting functions as an approximation for the optimal beliefs, and the possible psychophysical optimality of human beliefs. Next, we consider an advisor selection problem wherein the advisee of a certain belief chooses an advisor from a set of candidates with varying beliefs. We characterize the decision region for choosing such an advisor and argue that an advisee with beliefs varying from the true prior often ends up selecting a suboptimal advisor, indicating the need for a social planner. We close with a discussion on the implications of the study toward designing artificial intelligence systems for augmenting human intelligence.

There has been significant interest of late in generating behavior of agents that is interpretable to the human (observer) in the loop. However, the work in this area has typically lacked coherence on the topic, with proposed solutions for "explicable", "legible", "predictable" and "transparent" planning with overlapping, and sometimes conflicting, semantics all aimed at some notion of understanding what intentions the observer will ascribe to an agent by observing its behavior. This is also true for the recent works on "security" and "privacy" of plans which are also trying to answer the same question, but from the opposite point of view -- i.e. when the agent is trying to hide instead of revealing its intentions. This paper attempts to provide a workable taxonomy of relevant concepts in this exciting and emerging field of inquiry.

This paper reviews the reasons that Human-in-the-Loop is both critical for preventing widely-understood failure modes for machine learning, and not a practical solution. Following this, we review two current heuristic methods for addressing this. The first is provable safety envelopes, which are possible only when the dynamics of the system are fully known, but can be useful safety guarantees when optimal behavior is based on machine learning with poorly-understood safety characteristics. The second is the simpler circuit breaker model, which can forestall or prevent catastrophic outcomes by stopping the system, without any specific model of the system. This paper proposes using heuristic, dynamic safety envelopes, which are a plausible halfway point between these approaches that allows human oversight without some of the more difficult problems faced by Human-in-the-Loop systems. Finally, the paper concludes with how this approach can be used for governance of systems where otherwise unsafe systems are deployed.

Today, people increasingly rely on computer agents in their lives, from searching for information, to chatting with a bot, to performing everyday tasks. These agent-based systems are our first forays into a world in which machines will assist, teach, counsel, care for, and entertain us. While one could imagine purely rational agents in these roles, this prospect is not attractive for several reasons, which we will outline in this article. The field of affective computing concerns the design and development of computer systems that sense, interpret, adapt, and potentially respond appropriately to human emotions. Here, we specifically focus on the design of affective agents and assistants.

Inverse design is an outstanding challenge in disordered systems with multiple length scales such as polymers, particularly when designing polymers with desired phase behavior. We demonstrate high-accuracy tuning of poly(2-oxazoline) cloud point via machine learning. With a design space of four repeating units and a range of molecular masses, we achieve an accuracy of 4 {\deg}C root mean squared error (RMSE) in a temperature range of 24-90 {\deg}C, employing gradient boosting with decision trees. The RMSE is >3x better than linear and polynomial regression. We perform inverse design via particle-swarm optimization, predicting and synthesizing 17 polymers with constrained design at 4 target cloud points from 37 to 80 {\deg}C. Our approach challenges the status quo in polymer design with a machine learning algorithm, that is capable of fast and systematic discovery of new polymers.

Core-selection is a crucial property of rules in the literature of resource allocation. It is also desirable, from the perspective of mechanism design, to address the incentive of agents to cheat by misreporting their preferences. This paper investigates the exchange problem where (i) each agent is initially endowed with (possibly multiple) indivisible goods, (ii) agents' preferences are assumed to be conditionally lexicographic, and (iii) side payments are prohibited. We propose an exchange rule called augmented top-trading-cycles (ATTC), based on the original TTC procedure. We first show that ATTC is core-selecting and runs in polynomial time with respect to the number of goods. We then show that finding a beneficial misreport under ATTC is NP-hard. We finally clarify relationship of misreporting with splitting and hiding, two different types of manipulations, under ATTC.