Healthcare decision-making represents one of the most challenging domains for Artificial Intelligence (AI), requiring the integration of diverse knowledge sources, complex reasoning, and various external analytical tools. Current AI systems often rely on either task-specific models, which offer limited adaptability, or general language models without grounding with specialized external knowledge and tools. We introduce MedOrch, a novel framework that orchestrates multiple specialized tools and reasoning agents to provide comprehensive medical decision support. MedOrch employs a modular, agent-based architecture that facilitates the flexible integration of domain-specific tools without altering the core system. Furthermore, it ensures transparent and traceable reasoning processes, enabling clinicians to meticulously verify each intermediate step underlying the system's recommendations. We evaluate MedOrch across three distinct medical applications: Alzheimer's disease diagnosis, chest X-ray interpretation, and medical visual question answering, using authentic clinical datasets. The results demonstrate MedOrch's competitive performance across these diverse medical tasks. Notably, in Alzheimer's disease diagnosis, MedOrch achieves an accuracy of 93.26%, surpassing the state-of-the-art baseline by over four percentage points. For predicting Alzheimer's disease progression, it attains a 50.35% accuracy, marking a significant improvement. In chest X-ray analysis, MedOrch exhibits superior performance with a Macro AUC of 61.2% and a Macro F1-score of 25.5%. Moreover, in complex multimodal visual question answering (Image+Table), MedOrch achieves an accuracy of 54.47%. These findings underscore MedOrch's potential to advance healthcare AI by enabling reasoning-driven tool utilization for multimodal medical data processing and supporting intricate cognitive tasks in clinical decision-making.

We introduce Sorrel (https://github.com/social-ai-uoft/sorrel), a simple Python interface for generating and testing new multi-agent reinforcement learning environments. This interface places a high degree of emphasis on simplicity and accessibility, and uses a more psychologically intuitive structure for the basic agent-environment loop, making it a useful tool for social scientists to investigate how learning and social interaction leads to the development and change of group dynamics. In this short paper, we outline the basic design philosophy and features of Sorrel.

Multi-agent large language models (MA-LLMs) are a rapidly growing research area that leverages multiple interacting language agents to tackle complex tasks, outperforming single-agent large language models. This literature review synthesizes the latest research on agent profiles, communication structures, and decision-making processes, drawing insights from both traditional multi-agent systems and state-of-the-art MA-LLM studies. In doing so, it aims to address the lack of direct comparisons in the field, illustrating how factors like scalability, communication structure, and decision-making processes influence MA-LLM performance. By examining frequent practices and outlining current challenges, the review reveals that multi-agent approaches can yield superior results but also face elevated computational costs and under-explored challenges unique to MA-LLM. Overall, these findings provide researchers and practitioners with a roadmap for developing robust and efficient multi-agent AI solutions.

Cognitive science and artificial intelligence (AI) have grown up together as fields. The computational models of human minds developed in cognitive science have long served as benchmarks to articulate what it means for a system to be flexibly intelligent. Recent advances in AI, particularly around large language models (LLMs), are creating new opportunities to reciprocate this influence. Already, LLMs are being offered as scalable "cognitive models" of human behavior [Binz et al., 2024], automatic analysts of unstructured psychological text [Rathje et al., 2024], and components in neurosymbolic cognitive architectures (e.g., [W ong et al., 2023]). However, most applications of LLMs to cognitive science have so far focused on individual cognition.

Human motion generative modeling or synthesis aims to characterize complicated human motions of daily activities in diverse real-world environments. However, current research predominantly focuses on either low-level, short-period motions or high-level action planning, without taking into account the hierarchical goal-oriented nature of human activities. In this work, we take a step forward from human motion generation to human behavior modeling, which is inspired by cognitive science. We present a unified framework, dubbed Generative Behavior Control (GBC), to model diverse human motions driven by various high-level intentions by aligning motions with hierarchical behavior plans generated by large language models (LLMs). Our insight is that human motions can be jointly controlled by task and motion planning in robotics, but guided by LLMs to achieve improved motion diversity and physical fidelity. Meanwhile, to overcome the limitations of existing benchmarks, i.e., lack of behavioral plans, we propose GBC-100K dataset annotated with a hierarchical granularity of semantic and motion plans driven by target goals. Our experiments demonstrate that GBC can generate more diverse and purposeful high-quality human motions with 10* longer horizons compared with existing methods when trained on GBC-100K, laying a foundation for future research on behavioral modeling of human motions. Our dataset and source code will be made publicly available.

Reviewers found this paper to be an original and useful addition to the field of multi-agent games. While some of the presentation could be clarified (see specific reviewer comments, e.g. about the revelation game), there was a consensus that the paper is generally well-written and clear enough for publication, with the proposed corrections.

My two biggest complaints center on 1) the illustrative single-step matrix game of section 4.1 and figure 3 and 2) the practical applications of MACKRL. 1) Since the primary role of the single-step matrix game in section 4.1 is illustrative, it should be much clearer what is going on. How are all 3 policies parameterized? What information does each have access to? What is the training data? First, let's focus on the JAL policy. As presented up until this point in the paper, JAL means centralized training *and* execution.

All reviewers agreed this paper is well written presenting some interesting novel ideas. Reviewers believe that integrating common knowledge directly into Multi-agent RL training is a nice idea, and suggests some interesting future directions of research. Initially, there were shared concerns and confusion though about a number of issues, most prominently about the matrix game example. After reading and discussing the authors rebuttal though it seems the authors adequately addressed some of the primary concerns, and the general sense is that this paper is solid and of interest to be presented at NeurIPS.

The Starcraft results also seem fine, but not so strong as it make it obvious that committed exploration is a crucial empirical improvement for QMIX - while MAVEN agents learn faster in 3s5z, the final performance looks the same; MAVEN agents seem to have less variability in final win rate on 5m_vs_6m; and QMIX actually seems to have better final performance on 10m_vs_11m. The results in figure 2 and 4 do however suggest that there may be scenarios where the advantage of MAVEN is higher. Minor comments: 1) line 64 and others: the subscript "qmix" should probably be wrapped in a "\text{}" 2) first eqn in section 3: inconsistency between using subscripts and superscripts, i.e. u_i and u i 3) line 81: perhaps better phrased as: "the *best* action of agent i..." 4) line 86: u_n i - u_ U i? 5) line 87: I was confused by what "the set of all possible such orderings over the action-values" means. Besides a degeneracy when some of the Q values are identical, isn't there only one valid ordering? Or are you just trying to cover that degeneracy? 6) Definition 1: perhaps add an intuitive explanation, e.g. "Intuitively, a Q-function is non-monotonic if the ordering of best actions for agent i can be affected by the other agents action choices at that time step."

The paper presents a new exploration strategy for decentralized MARL that is based on a joint latent variable that is shared between the agent. This paper is a difficult case. While the theoretical insights concerning the difficulty of the exploration problem in decentralized MARL are insightful, the experimental results were not good enough in the original submission to convince the reviewers. The algorithm was only in one case considerably better than the competitor QMix and other baseline comparison were missing. However, in the rebuttal the authors provided much better results as well as additional comparison to Qtrans.