An often neglected issue in multi-agent reinforcement learning (MARL) is the potential presence of unreliable agents in the environment whose deviations from expected behavior can prevent a system from accomplishing its intended tasks. In particular, consensus is a fundamental underpinning problem of cooperative distributed multi-agent systems. Consensus requires different agents, situated in a decentralized communication network, to reach an agreement out of a set of initial proposals that they put forward. Learning-based agents should adopt a protocol that allows them to reach consensus despite having one or more unreliable agents in the system. This paper investigates the problem of unreliable agents in MARL, considering consensus as a case study. Echoing established results in the distributed systems literature, our experiments show that even a moderate fraction of such agents can greatly impact the ability of reaching consensus in a networked environment. We propose Reinforcement Learning-based Trusted Consensus (RLTC), a decentralized trust mechanism, in which agents can independently decide which neighbors to communicate with. We empirically demonstrate that our trust mechanism is able to handle unreliable agents effectively, as evidenced by higher consensus success rates.

The 2024 Index is our most comprehensive to date and arrives at an important moment when AI's influence on society has never been more pronounced. This year, we have broadened our scope to more extensively cover essential trends such as technical advancements in AI, public perceptions of the technology, and the geopolitical dynamics surrounding its development. Featuring more original data than ever before, this edition introduces new estimates on AI training costs, detailed analyses of the responsible AI landscape, and an entirely new chapter dedicated to AI's impact on science and medicine. The AI Index report tracks, collates, distills, and visualizes data related to artificial intelligence (AI). Our mission is to provide unbiased, rigorously vetted, broadly sourced data in order for policymakers, researchers, executives, journalists, and the general public to develop a more thorough and nuanced understanding of the complex field of AI. The AI Index is recognized globally as one of the most credible and authoritative sources for data and insights on artificial intelligence. Previous editions have been cited in major newspapers, including the The New York Times, Bloomberg, and The Guardian, have amassed hundreds of academic citations, and been referenced by high-level policymakers in the United States, the United Kingdom, and the European Union, among other places. This year's edition surpasses all previous ones in size, scale, and scope, reflecting the growing significance that AI is coming to hold in all of our lives.

Autonomous agents operating around human actors must consider how their behaviors might affect those humans, even when not directly interacting with them. To this end, it is often beneficial to be predictable and appear naturalistic. Existing methods to address this problem use human actor intent modeling or imitation learning techniques, but these approaches rarely capture all possible motivations for human behavior or require significant amounts of data. In contrast, we propose a technique for modeling naturalistic behavior as a set of convex hulls computed over a relatively small dataset of human behavior. Given this set, we design an optimization-based filter which projects arbitrary trajectories into it to make them more naturalistic for autonomous agents to execute while also satisfying dynamics constraints. We demonstrate our methods on real-world human driving data from the inD intersection dataset (Bock et al., 2020).

Evaluating startups in their early stages is a complex task that requires detailed analysis by experts. While automating this process on a large scale can significantly impact businesses, the inherent complexity poses challenges. This paper addresses this challenge by introducing the Startup Success Forecasting Framework (SSFF), a new automated system that combines traditional machine learning with advanced language models. This intelligent agent-based architecture is designed to reason, act, synthesize, and decide like a venture capitalist to perform the analysis end-to-end. The SSFF is made up of three main parts: - Prediction Block: Uses random forests and neural networks to make predictions. - Analyst Block: Simulates VC analysis scenario and uses SOTA prompting techniques - External Knowledge Block: Gathers real-time information from external sources. This framework requires minimal input data about the founder and startup description, enhances it with additional data from external resources, and performs a detailed analysis with high accuracy, all in an automated manner

The execution of Belief-Desire-Intention (BDI) agents in a Multi-Agent System (MAS) can be practically implemented on top of low-level concurrency mechanisms that impact on efficiency, determinism, and reproducibility. We argue that developers should specify the MAS behaviour independently of the execution model, and choose or configure the concurrency model later on, according to the specific needs of their target domain, leaving the MAS specification unaffected. We identify patterns for mapping the agent execution over the underlying concurrency abstractions, and investigate which concurrency models are supported by some of the most commonly used BDI platforms. Although most frameworks support multiple concurrency models, we find that they mostly hide them under the hood, making them opaque to the developer, and actually limiting the possibility of fine-tuning the MAS.

We consider inverse reinforcement learning problems with concave utilities. Concave Utility Reinforcement Learning (CURL) is a generalisation of the standard RL objective, which employs a concave function of the state occupancy measure, rather than a linear function. CURL has garnered recent attention for its ability to represent instances of many important applications including the standard RL such as imitation learning, pure exploration, constrained MDPs, offline RL, human-regularized RL, and others. Inverse reinforcement learning is a powerful paradigm that focuses on recovering an unknown reward function that can rationalize the observed behaviour of an agent. There has been recent theoretical advances in inverse RL where the problem is formulated as identifying the set of feasible reward functions. However, inverse RL for CURL problems has not been considered previously. In this paper we show that most of the standard IRL results do not apply to CURL in general, since CURL invalidates the classical Bellman equations. This calls for a new theoretical framework for the inverse CURL problem. Using a recent equivalence result between CURL and Mean-field Games, we propose a new definition for the feasible rewards for I-CURL by proving that this problem is equivalent to an inverse game theory problem in a subclass of mean-field games. We present initial query and sample complexity results for the I-CURL problem under assumptions such as Lipschitz-continuity. Finally, we outline future directions and applications in human--AI collaboration enabled by our results.

We study the problem of resilient average consensus in multi-agent systems where some of the agents are subject to failures or attacks. The objective of resilient average consensus is for non-faulty/normal agents to converge to the average of their initial values despite the erroneous effects from malicious agents. To this end, we propose a successful distributed iterative resilient average consensus algorithm for the multi-agent networks with general directed topologies. The proposed algorithm has two parts at each iteration: detection and averaging. For the detection part, we propose two distributed algorithms and one of them can detect malicious agents with only the information from direct in-neighbors. For the averaging part, we extend the applicability of an existing averaging algorithm where normal agents can remove the effects from malicious agents so far, after they are detected. Another important feature of our method is that it can handle the case where malicious agents are neighboring and collaborating with each other to mislead the normal ones from averaging. This case cannot be solved by existing detection approaches in related literature. Moreover, our algorithm is efficient in storage usage especially for large-scale networks as each agent only requires the values of neighbors within two hops. Lastly, numerical examples are given to verify the efficacy of the proposed algorithms.

Designing Reinforcement Learning (RL) solutions for real-life problems remains a significant challenge. A major area of concern is safety. "Shielding" is a popular technique to enforce safety in RL by turning user-defined safety specifications into safe agent behavior. However, these methods either suffer from extreme learning delays, demand extensive human effort in designing models and safe domains in the problem, or require pre-computation. In this paper, we propose a new permissibility-based framework to deal with safety and shield construction. Permissibility was originally designed for eliminating (non-permissible) actions that will not lead to an optimal solution to improve RL training efficiency. This paper shows that safety can be naturally incorporated into this framework, i.e. extending permissibility to include safety, and thereby we can achieve both safety and improved efficiency. Experimental evaluation using three standard RL applications shows the effectiveness of the approach.

Generative agents, which implement behaviors using a large language model (LLM) to interpret and evaluate an environment, has demonstrated the capacity to solve complex tasks across many social and technological domains. However, when these agents interact with other agents and humans in presence of social structures such as existing norms, fostering cooperation between them emerges as a fundamental challenge. In this paper, we develop the framework of a Normative Module: an architecture for generative agents designed to enhance cooperation by enabling agents to recognize and adapt to the normative infrastructure of a given environment, in the form of institutions that define acceptable behaviors within a group of agents. We focus on the equilibrium selection aspect of the cooperation problem and inform our agent design based on the existence of classification institutions that implement correlated equilibrium to provide effective resolution of the equilibrium selection problem. Specifically, the normative module enables agents to learn through peer interactions which of multiple candidate institutions in the environment, does a group treat as authoritative. By enabling normative competence in this sense, agents gain ability to coordinate their sanctioning behaviour; coordinated sanctioning behaviour in turn shapes primary behaviour within a social environment, leading to higher average welfare We design a new environment that supports institutions and evaluate the proposed framework based on two key criteria derived from agent interactions with peers and institutions: (i) the agent's ability to disregard non-authoritative institutions and (ii) the agent's ability to identify authoritative institutions among several options. Crucially, we show that these capabilities allow the agent to achieve more stable cooperative outcomes compared to baseline agents without the normative module, paving the way for future research in a new avenue of designing environments and agents that account for normative infrastructure.

We explore a Federated Reinforcement Learning (FRL) problem where $N$ agents collaboratively learn a common policy without sharing their trajectory data. To date, existing FRL work has primarily focused on agents operating in the same or ``similar" environments. In contrast, our problem setup allows for arbitrarily large levels of environment heterogeneity. To obtain the optimal policy which maximizes the average performance across all potentially completely different environments, we propose two algorithms: FedSVRPG-M and FedHAPG-M. In contrast to existing results, we demonstrate that both FedSVRPG-M and FedHAPG-M, both of which leverage momentum mechanisms, can exactly converge to a stationary point of the average performance function, regardless of the magnitude of environment heterogeneity. Furthermore, by incorporating the benefits of variance-reduction techniques or Hessian approximation, both algorithms achieve state-of-the-art convergence results, characterized by a sample complexity of $\mathcal{O}\left(\epsilon^{-\frac{3}{2}}/N\right)$. Notably, our algorithms enjoy linear convergence speedups with respect to the number of agents, highlighting the benefit of collaboration among agents in finding a common policy.